{ "generated_at": "2026-07-10 05:25 PDT", "title": "Focused AI Model-Performance KOLs + Benchmarking Posts on X", "methodology": [ "Follower counts are approximate snapshots gathered from unauthenticated X profile snippets, Feedspot AI influencer listings, third-party mirrors, and project pages on 2026-07-10 PDT.", "X follower counts move quickly and unauthenticated access is inconsistent; counts should be treated as directional for outreach prioritization, not audited metrics.", "General AI KOLs are ranked by practical influence for AI narratives, model launches, developer adoption, benchmarks, open-source LLMs, and research discourse — not purely by follower count.", "Bio/drug-discovery benchmark KOLs include individual researchers, project/org accounts, and benchmark-amplification channels that post or shape model evaluation in biology, chemistry, drug discovery, medical AI, and scientific agents.", "Mega-celebrity and broad futurist/business accounts are excluded from the focused KOL table unless they regularly post concrete model-performance, benchmark, eval, or leaderboard content.", "Top benchmarking-post rows are web-search-only. Direct X post URLs are marked High/Medium; rows where unauthenticated search did not surface an exact post URL use an X search URL and are marked Low confidence so they can be followed up manually/API-side." ], "general_ai_kols": [ { "name": "Elon Musk", "handle": "@elonmusk", "url": "https://x.com/elonmusk", "followers": "~240.7M", "category": "xAI / X platform", "why": "Massive distribution; Grok/xAI announcements and X-native AI narratives." }, { "name": "Sam Altman", "handle": "@sama", "url": "https://x.com/sama", "followers": "~5.4M", "category": "OpenAI / frontier AI CEO", "why": "Highest-signal OpenAI product, policy, and frontier-model narrative account." }, { "name": "Lex Fridman", "handle": "@lexfridman", "url": "https://x.com/lexfridman", "followers": "~5.0M", "category": "Podcast / AI public discourse", "why": "Major long-form AI interviewer; bridges research, founders, policy, and public audiences." }, { "name": "Andrej Karpathy", "handle": "@karpathy", "url": "https://x.com/karpathy", "followers": "~3.2M", "category": "AI researcher / educator", "why": "Deep technical explainer with huge developer and research reach." }, { "name": "Andrew Ng", "handle": "@AndrewYNg", "url": "https://x.com/AndrewYNg", "followers": "~1.5–1.6M", "category": "AI education / applied ML", "why": "Applied AI adoption, ML education, and enterprise AI influence." }, { "name": "Kai-Fu Lee", "handle": "@kaifulee", "url": "https://x.com/kaifulee", "followers": "~1.4M", "category": "AI investor / 01.AI", "why": "China/global AI strategy, startup ecosystem, and AI commercialization." }, { "name": "Demis Hassabis", "handle": "@demishassabis", "url": "https://x.com/demishassabis", "followers": "~1.3M", "category": "Google DeepMind CEO", "why": "AGI, Gemini, AlphaFold/Isomorphic, and science-AI authority." }, { "name": "Yann LeCun", "handle": "@ylecun", "url": "https://x.com/ylecun", "followers": "~0.94–1.2M", "category": "Meta AI / Turing laureate", "why": "Major voice on open AI, model limits, AI risk debates, and research direction." }, { "name": "Greg Brockman", "handle": "@gdb", "url": "https://x.com/gdb", "followers": "~0.63–1.0M", "category": "OpenAI cofounder/president", "why": "OpenAI research/product amplification and developer ecosystem visibility." }, { "name": "Ilya Sutskever", "handle": "@ilyasut", "url": "https://x.com/ilyasut", "followers": "~760.9K", "category": "SSI / frontier research", "why": "Foundational LLM researcher; rare posts can move frontier-AI attention." }, { "name": "Mira Murati", "handle": "@miramurati", "url": "https://x.com/miramurati", "followers": "~742.1K", "category": "Thinking Machines Lab", "why": "New frontier-lab founder; ex-OpenAI CTO; high launch attention." }, { "name": "Aravind Srinivas", "handle": "@AravSrinivas", "url": "https://x.com/AravSrinivas", "followers": "~800.6K", "category": "Perplexity CEO", "why": "AI search, product velocity, consumer AI distribution." }, { "name": "Fei-Fei Li", "handle": "@drfeifei", "url": "https://x.com/drfeifei", "followers": "~632–781K", "category": "World Labs / Stanford", "why": "Spatial intelligence, computer vision, human-centered AI, healthcare AI." }, { "name": "François Chollet", "handle": "@fchollet", "url": "https://x.com/fchollet", "followers": "~566–704K", "category": "Keras / ARC-AGI", "why": "Reasoning benchmarks, ARC Prize, AGI criticism, deep learning frameworks." }, { "name": "Geoffrey Hinton", "handle": "@geoffreyhinton", "url": "https://x.com/geoffreyhinton", "followers": "~489–618K", "category": "Deep learning pioneer", "why": "High-authority voice on deep learning and AI risk." }, { "name": "Dario Amodei", "handle": "@DarioAmodei", "url": "https://x.com/DarioAmodei", "followers": "~478.6K", "category": "Anthropic CEO", "why": "Claude/frontier AI strategy, safety, and lab positioning." }, { "name": "Jeff Dean", "handle": "@JeffDean", "url": "https://x.com/JeffDean", "followers": "~361–448K", "category": "Google DeepMind / Google Research", "why": "Gemini, systems ML, TensorFlow/MapReduce credibility." }, { "name": "Clément Delangue", "handle": "@ClementDelangue", "url": "https://x.com/ClementDelangue", "followers": "~452.1K", "category": "Hugging Face CEO", "why": "Open-source AI platform, model hosting, leaderboards, community distribution." }, { "name": "Alexandr Wang", "handle": "@alexandr_wang", "url": "https://x.com/alexandr_wang", "followers": "~541.7K", "category": "Meta AI / Scale founder", "why": "Data, enterprise AI, frontier-lab/infra narratives." }, { "name": "Jim Fan", "handle": "@DrJimFan", "url": "https://x.com/DrJimFan", "followers": "~435–486K", "category": "NVIDIA robotics / agents", "why": "Physical AGI, robotics, agent systems, simulation." }, { "name": "AK", "handle": "@_akhaliq", "url": "https://x.com/_akhaliq", "followers": "~500.4K", "category": "AI papers / Hugging Face", "why": "High-velocity research-paper amplification across ML and AI." }, { "name": "Sebastian Raschka", "handle": "@rasbt", "url": "https://x.com/rasbt", "followers": "~476K", "category": "ML/LLM educator", "why": "LLM-from-scratch and practical model education for engineers." }, { "name": "Soumith Chintala", "handle": "@soumithchintala", "url": "https://x.com/soumithchintala", "followers": "~247–311K", "category": "PyTorch / Meta", "why": "Open-source ML infrastructure credibility and PyTorch ecosystem." }, { "name": "Lilian Weng", "handle": "@lilianweng", "url": "https://x.com/lilianweng", "followers": "~88–267K", "category": "OpenAI/Thinking Machines researcher", "why": "Canonical long-form explainers on agents, RLHF, safety, and LLM systems." }, { "name": "Jeremy Howard", "handle": "@jeremyphoward", "url": "https://x.com/jeremyphoward", "followers": "~236–289K", "category": "fast.ai / Answer.AI", "why": "Practical open AI, education, efficient model training." }, { "name": "Emad Mostaque", "handle": "@EMostaque", "url": "https://x.com/EMostaque", "followers": "~328.3K", "category": "Stability AI founder", "why": "Open model/media AI ecosystem voice and frequent AI market commentary." }, { "name": "Lior / AlphaSignal", "handle": "@AlphaSignalAI", "url": "https://x.com/AlphaSignalAI", "followers": "~80.4K", "category": "AI R&D newsletter", "why": "Technical AI paper/news amplification to ML practitioners." }, { "name": "Logan Kilpatrick", "handle": "@OfficialLoganK", "url": "https://x.com/OfficialLoganK", "followers": "~339.2K", "category": "Google AI Studio / Gemini devrel", "why": "Developer-facing Gemini/API signal and launch amplification." }, { "name": "Simon Willison", "handle": "@simonw", "url": "https://x.com/simonw", "followers": "~154.5K", "category": "Developer / AI tooling analyst", "why": "Practical LLM tooling, agents, security, and model behavior analysis." }, { "name": "Hamel Husain", "handle": "@HamelHusain", "url": "https://x.com/HamelHusain", "followers": "n/a", "category": "AI evals / applied LLM systems", "why": "One of the clearest practitioner voices on LLM evaluation and AI product quality." }, { "name": "Aidan McLaughlin", "handle": "@aidan_mclau", "url": "https://x.com/aidan_mclau", "followers": "~47.8K", "category": "OpenAI research", "why": "Model-design/frontier AI commentary; useful signal for lab-internal research perspective." }, { "name": "Ian Goodfellow", "handle": "@goodfellow_ian", "url": "https://x.com/goodfellow_ian", "followers": "~315.7K", "category": "DeepMind / GAN inventor", "why": "Foundational ML researcher; high credibility in deep learning." }, { "name": "Oriol Vinyals", "handle": "@oriolvinyalsml", "url": "https://x.com/oriolvinyalsml", "followers": "~182.4K", "category": "Google DeepMind research", "why": "Gemini co-lead; AlphaStar/AlphaCode/seq2seq lineage." }, { "name": "Nando de Freitas", "handle": "@NandoDF", "url": "https://x.com/NandoDF", "followers": "~108.9K", "category": "Microsoft AI / ex-DeepMind", "why": "Superintelligence, deep learning, and model-research commentary." }, { "name": "Abhishek Thakur", "handle": "@abhi1thakur", "url": "https://x.com/abhi1thakur", "followers": "~83.1K", "category": "Kaggle / Hugging Face AutoTrain", "why": "Applied ML, competitions, AutoML, open-source model building." } ], "bio_benchmark_kols": [ { "name": "OpenAI", "handle": "@OpenAI", "url": "https://x.com/OpenAI/status/2067346916929937827", "followers": "very large / n.a.", "focus": "LifeSciBench; GeneBench-Pro; life-science model evaluation", "examples": "LifeSciBench: 750 expert-authored tasks, 173 scientist contributors, 19,020 rubric criteria.", "relevance": "Direct benchmark comparator and high-impact launch/amplification target." }, { "name": "Tacit Labs", "handle": "@tacitlabsco", "url": "https://x.com/tacitlabsco", "followers": "n/a", "focus": "LifeSciBench partner / biology verification loop", "examples": "Referenced in LifeSciBench launch ecosystem with OpenAI.", "relevance": "Potential partner/peer for life-science task design and validation." }, { "name": "Nina Fitzpatrick", "handle": "@ninklefitz", "url": "https://x.com/ninklefitz", "followers": "n/a", "focus": "LifeSciBench / Tacit Labs", "examples": "Named in LifeSciBench/Tacit Labs launch amplification.", "relevance": "Likely useful individual contact around biology eval design." }, { "name": "Andrew Droste", "handle": "@AmDroste", "url": "https://x.com/AmDroste", "followers": "n/a", "focus": "LifeSciBench / Tacit Labs", "examples": "Named in LifeSciBench/Tacit Labs launch amplification.", "relevance": "Likely useful individual contact around life-science verification." }, { "name": "FutureHouse", "handle": "@FutureHouseSF", "url": "https://x.com/FutureHouseSF", "followers": "n/a", "focus": "Biology/science agents, LAB-Bench, scientific task evaluation", "examples": "LAB-Bench and FutureHouse scientific-agent ecosystem; BixBench repo under Future-House.", "relevance": "Top target for scientific-agent benchmark comparison/amplification." }, { "name": "Phylo", "handle": "@phylo_bio", "url": "https://x.com/phylo_bio", "followers": "n/a", "focus": "Biomni, BiomniBench, DrugDiscoveryBench, biomedical agents", "examples": "DrugDiscoveryBench with Scale Labs; BiomniBench process-level evaluation.", "relevance": "Direct DDDBench peer/benchmark partner and high-priority outreach target." }, { "name": "Scale Labs", "handle": "@ScaleAILabs", "url": "https://x.com/ScaleAILabs/status/2072011679064584343", "followers": "n/a", "focus": "DrugDiscoveryBench and agent evals", "examples": "DrugDiscoveryBench: 82 early-stage DD tasks with expert rubrics and leaderboard.", "relevance": "Direct benchmark comparison; useful for model-vs-harness narrative." }, { "name": "Scale AI", "handle": "@scale_AI", "url": "https://x.com/scale_AI", "followers": "large / n.a.", "focus": "AI data/evals; Scale Labs amplification", "examples": "Scale blog and labs leaderboard for DrugDiscoveryBench.", "relevance": "Enterprise-eval distribution channel." }, { "name": "Kexin Huang", "handle": "@KexinHuang5", "url": "https://x.com/KexinHuang5/status/2056397835327656004", "followers": "n/a", "focus": "Biomni, BiomniBench, AI agents for biomedical research", "examples": "Posted BiomniBench launch: process-level evaluation of agents on long-horizon biology tasks.", "relevance": "Strong KOL for agentic biomedical benchmarks and scientific-agent systems." }, { "name": "Therapeutics Data Commons", "handle": "@ProjectTDC", "url": "https://x.com/projecttdc", "followers": "~1.3K", "focus": "Drug-discovery ML datasets/tasks/leaderboards", "examples": "TDC platform for AI-ready therapeutic datasets and leaderboards.", "relevance": "Foundational benchmark ecosystem for DDDBench positioning." }, { "name": "Jure Leskovec", "handle": "@jure", "url": "https://x.com/jure", "followers": "~44.7K", "focus": "TDC, graph ML, biomedical ML", "examples": "TDC / Harvard-Stanford biomedical ML ecosystem.", "relevance": "Senior benchmark and graph-ML amplifier." }, { "name": "Valence Labs", "handle": "@valence_ai", "url": "https://x.com/valence_ai", "followers": "~5.9K", "focus": "Polaris drug-discovery benchmark platform", "examples": "PolarisHub benchmark/data platform for drug discovery.", "relevance": "Highly relevant for industry-standard DD benchmarking." }, { "name": "Pat Walters", "handle": "@wpwalters", "url": "https://x.com/wpwalters", "followers": "~6.3K", "focus": "Cheminformatics benchmark critique, ADMET, blind challenges", "examples": "OpenADMET/blind-challenge and benchmark-validity commentary.", "relevance": "High-credibility evaluator/KOL for DD benchmark quality." }, { "name": "OpenBioML", "handle": "@openbioml", "url": "https://x.com/openbioml", "followers": "~4.1K", "focus": "Open biology ML, ChemNLP, model-risk/eval discussion", "examples": "Community journal clubs and BioML paper amplification.", "relevance": "Community amplifier for open BioML benchmarking." }, { "name": "Biology+AI Daily", "handle": "@BiologyAIDaily", "url": "https://x.com/BiologyAIDaily", "followers": "n/a", "focus": "AI-bio paper/benchmark amplification", "examples": "Amplified BixBench and other AI-biology benchmark papers.", "relevance": "Useful awareness channel for DDDBench/MMAI benchmark posts." }, { "name": "Stephen Turner", "handle": "@strnr", "url": "https://x.com/strnr", "followers": "~31.3K", "focus": "Genomics, computational biology, AIxBio commentary", "examples": "Genomics and biosecurity/AI benchmark commentary.", "relevance": "Strong commentator/amplifier for AI-bio benchmark credibility." }, { "name": "Jablonka Lab", "handle": "@jablonkagroup", "url": "https://x.com/jablonkagroup", "followers": "~359", "focus": "ChemBench, chemistry/materials LLM evaluation", "examples": "ChemBench benchmark for chemistry reasoning and tool use.", "relevance": "Relevant analog for chemistry/LLM eval methodology." }, { "name": "Andrew White", "handle": "@andrewwhite01", "url": "https://x.com/andrewwhite01", "followers": "n/a", "focus": "ChemCrow, PaperQA, scientific agents, chemistry LLMs", "examples": "ChemCrow and scientific-agent evaluation work.", "relevance": "Very strong chemistry/science-agent benchmark KOL." }, { "name": "Philippe Schwaller", "handle": "@pschwllr", "url": "https://x.com/pschwllr", "followers": "n/a", "focus": "Chemical language models, RXN, ChemCrow ecosystem", "examples": "ChemCrow / reaction prediction / chemistry AI work.", "relevance": "Good benchmark reviewer and chemistry-agent evaluator." }, { "name": "Chai Discovery", "handle": "@chaidiscovery", "url": "https://x.com/chaidiscovery", "followers": "~5.6K", "focus": "Biomolecular structure prediction benchmarks", "examples": "Chai-1, chai-lab, structure prediction comparisons.", "relevance": "Protein/structure benchmark comparison channel." }, { "name": "Google DeepMind", "handle": "@GoogleDeepMind", "url": "https://x.com/GoogleDeepMind", "followers": "~1.2M", "focus": "AlphaFold/CASP, Gemini science AI", "examples": "AlphaFold and Isomorphic science-AI narratives.", "relevance": "Huge amplifier; benchmark framing authority." }, { "name": "Rosetta Commons", "handle": "@RosettaCommons", "url": "https://x.com/RosettaCommons", "followers": "~1.9K", "focus": "Protein modeling/design benchmarks and community tools", "examples": "Rosetta, CASP/CAPRI-related modeling ecosystem.", "relevance": "Protein-design benchmark credibility network." }, { "name": "ChEMBL Database", "handle": "@ChEMBL", "url": "https://x.com/chembl", "followers": "~2.9K", "focus": "Drug-discovery data resource used in benchmarks", "examples": "ChEMBL data/resource updates.", "relevance": "Useful data/benchmark community account." }, { "name": "Papers with Code", "handle": "@paperswithcode", "url": "https://x.com/paperswithcode", "followers": "~114.7K", "focus": "ML benchmark/leaderboard amplification", "examples": "Benchmark and paper leaderboard distribution.", "relevance": "Broad benchmark visibility channel." }, { "name": "Hugging Face", "handle": "@huggingface", "url": "https://x.com/huggingface", "followers": "~689K", "focus": "Open leaderboards and model hubs", "examples": "Open Medical LLM Leaderboard; dataset/model benchmark hosting.", "relevance": "High-impact hosting/amplification route for DDDBench/MMAI." }, { "name": "Percy Liang / Stanford HELM", "handle": "@percyliang", "url": "https://x.com/percyliang", "followers": "~106.5K", "focus": "HELM, MedHELM, holistic LLM evals", "examples": "Stanford CRFM HELM and medical evaluation frameworks.", "relevance": "Credible LLM-eval benchmark KOL." }, { "name": "METR", "handle": "@METR_Evals", "url": "https://x.com/METR_Evals", "followers": "~25.2K", "focus": "Agent/R&D capability evals, time-horizon evals", "examples": "Autonomous capability measurement and eval methodology.", "relevance": "Useful for agent-eval methodology framing." }, { "name": "UK AI Security Institute", "handle": "@AISecurityInst", "url": "https://x.com/AISecurityInst", "followers": "~7.6K", "focus": "Inspect Evals; chem/bio/frontier model evals", "examples": "Inspect eval framework and frontier-model safety evaluations.", "relevance": "Strong eval-framework/governance target." }, { "name": "Center for AI Safety", "handle": "@CAIS / @ai_risks", "url": "https://x.com/ai_risks", "followers": "~8.2K", "focus": "WMDP bio/chem/cyber hazardous-knowledge benchmark", "examples": "WMDP benchmark and biorisk/chem risk evals.", "relevance": "Relevant to bio/chem LLM evals; outreach should be careful and risk-aware." }, { "name": "Epoch AI", "handle": "@EpochAIResearch", "url": "https://x.com/EpochAIResearch", "followers": "~45.7K", "focus": "Benchmark datasets/trends and biorisk eval analysis", "examples": "Epoch benchmark tracking and AI trend analysis.", "relevance": "Methodology/benchmark-analysis amplifier." }, { "name": "MLCommons", "handle": "@MLCommons", "url": "https://x.com/MLCommons", "followers": "~3.6K", "focus": "Medical AI, safety, and science benchmark working groups", "examples": "MLCommons medical/data working groups.", "relevance": "Standards-setting outreach target." }, { "name": "Tanishq M. Abraham / MedARC", "handle": "@iScienceLuvr", "url": "https://x.com/iScienceLuvr", "followers": "n/a", "focus": "Medmarks medical LLM benchmark suite", "examples": "Medmarks benchmark posts and GitHub project.", "relevance": "Strong medical LLM benchmark KOL." } ], "sources": [ { "label": "Feedspot Top AI Influencers 2026", "url": "https://x.feedspot.com/artificial_intelligence_twitter_influencers/" }, { "label": "OpenAI LifeSciBench announcement", "url": "https://openai.com/index/introducing-life-sci-bench/" }, { "label": "OpenAI LifeSciBench X post", "url": "https://x.com/OpenAI/status/2067346916929937827" }, { "label": "Scale Labs DrugDiscoveryBench blog", "url": "https://scale.com/blog/drugdiscoverybench" }, { "label": "Phylo DrugDiscoveryBench blog", "url": "https://phylo.bio/blog/drugdiscoverybench" }, { "label": "Scale Labs DrugDiscoveryBench X post", "url": "https://x.com/ScaleAILabs/status/2072011679064584343" }, { "label": "BioML-bench preprint", "url": "https://www.biorxiv.org/content/10.1101/2025.09.01.673319v2.full" }, { "label": "Therapeutics Data Commons", "url": "https://tdcommons.ai/" }, { "label": "ProteinGym", "url": "https://proteingym.org/" }, { "label": "LAB-Bench arXiv", "url": "https://arxiv.org/abs/2407.10362" }, { "label": "BixBench GitHub", "url": "https://github.com/Future-House/BixBench" }, { "label": "ChemBench", "url": "https://lamalab-org.github.io/chembench/" }, { "label": "FutureHouse LAB-Bench GitHub", "url": "https://github.com/Future-House/LAB-Bench" }, { "label": "HeadGym AI influencers on X", "url": "https://headgym.com/blog/posts/ai-influencers-on-x" }, { "label": "Feedspot Machine Learning X influencers", "url": "https://x.feedspot.com/machine_learning_twitter_influencers/" }, { "label": "Zarif Automates best AI X accounts", "url": "https://www.zarifautomates.com/blog/best-ai-twitter-x-accounts-to-follow" }, { "label": "Artificial Analysis", "url": "https://artificialanalysis.ai/" }, { "label": "Aider leaderboards", "url": "https://aider.chat/docs/leaderboards/" }, { "label": "SWE-bench", "url": "https://www.swebench.com/" }, { "label": "LiveBench", "url": "https://livebench.ai/" } ], "top_200_ai_kols": [ { "rank": 1, "name": "Elon Musk", "handle": "@elonmusk", "url": "https://x.com/elonmusk", "followers": "~240.7M", "category": "xAI / X platform", "technical_flag": "Semi-technical", "technical_rationale": "xAI/Grok strategy and technical/product narratives, but mostly founder/platform distribution", "why": "Massive X-native reach; Grok/xAI announcements and AI platform narrative.", "source": "https://x.com/elonmusk" }, { "rank": 2, "name": "Sam Altman", "handle": "@sama", "url": "https://x.com/sama", "followers": "~5.4M", "category": "OpenAI / frontier AI CEO", "technical_flag": "Semi-technical", "technical_rationale": "Strategic lab and product framing rather than code-level posts", "why": "OpenAI narrative-setter and frontier-model launch signal.", "source": "https://x.com/sama" }, { "rank": 3, "name": "Lex Fridman", "handle": "@lexfridman", "url": "https://x.com/lexfridman", "followers": "~5.0M", "category": "AI podcast / public discourse", "technical_flag": "Semi-technical", "technical_rationale": "Technical interviews, broad public audience", "why": "Long-form AI interviews bridging researchers, founders, and policy.", "source": "https://x.com/lexfridman" }, { "rank": 4, "name": "Andrej Karpathy", "handle": "@karpathy", "url": "https://x.com/karpathy", "followers": "~3.2M", "category": "AI researcher / educator", "technical_flag": "Technical", "technical_rationale": "Posts deep model intuition, training, code, and technical education", "why": "One of the highest signal technical AI explainers on X.", "source": "https://x.com/karpathy" }, { "rank": 5, "name": "Andrew Ng", "handle": "@AndrewYNg", "url": "https://x.com/AndrewYNg", "followers": "~1.5–1.6M", "category": "AI education / applied ML", "technical_flag": "Technical", "technical_rationale": "Applied ML, data-centric AI, courses, technical practice", "why": "Major applied-AI educator and practitioner influence.", "source": "https://x.com/AndrewYNg" }, { "rank": 6, "name": "Kai-Fu Lee", "handle": "@kaifulee", "url": "https://x.com/kaifulee", "followers": "~1.4M", "category": "AI investor / 01.AI", "technical_flag": "Semi-technical", "technical_rationale": "Strategic AI and commercialization, occasionally model/company technical context", "why": "China/global AI strategy and commercialization voice.", "source": "https://x.com/kaifulee" }, { "rank": 7, "name": "Demis Hassabis", "handle": "@demishassabis", "url": "https://x.com/demishassabis", "followers": "~1.3M", "category": "Google DeepMind CEO", "technical_flag": "Technical", "technical_rationale": "Science-AI, DeepMind research, AlphaFold/Gemini technical context", "why": "High-authority AI-for-science and AGI lab voice.", "source": "https://x.com/demishassabis" }, { "rank": 8, "name": "Yann LeCun", "handle": "@ylecun", "url": "https://x.com/ylecun", "followers": "~0.94–1.2M", "category": "Meta AI / Turing laureate", "technical_flag": "Technical", "technical_rationale": "Research debates on world models, self-supervised learning, architectures", "why": "Major technical counterweight to LLM-scaling consensus.", "source": "https://x.com/ylecun" }, { "rank": 9, "name": "Greg Brockman", "handle": "@gdb", "url": "https://x.com/gdb", "followers": "~0.63–1.0M", "category": "OpenAI cofounder/president", "technical_flag": "Semi-technical", "technical_rationale": "Demos and product/research amplification, less code-level", "why": "OpenAI product/research amplification.", "source": "https://x.com/gdb" }, { "rank": 10, "name": "Ilya Sutskever", "handle": "@ilyasut", "url": "https://x.com/ilyasut", "followers": "~760.9K", "category": "SSI / frontier research", "technical_flag": "Technical", "technical_rationale": "Foundational LLM researcher; rare but technical significance", "why": "Rare posts carry frontier-lab/research weight.", "source": "https://x.com/ilyasut" }, { "rank": 11, "name": "Mira Murati", "handle": "@miramurati", "url": "https://x.com/miramurati", "followers": "~742.1K", "category": "Thinking Machines Lab", "technical_flag": "Semi-technical", "technical_rationale": "Lab/product strategy from technical leadership background", "why": "New frontier-lab founder and ex-OpenAI CTO.", "source": "https://x.com/miramurati" }, { "rank": 12, "name": "Aravind Srinivas", "handle": "@AravSrinivas", "url": "https://x.com/AravSrinivas", "followers": "~800.6K", "category": "Perplexity CEO", "technical_flag": "Semi-technical", "technical_rationale": "AI search/product strategy with some model context", "why": "AI search and consumer-AI distribution leader.", "source": "https://x.com/AravSrinivas" }, { "rank": 13, "name": "Fei-Fei Li", "handle": "@drfeifei", "url": "https://x.com/drfeifei", "followers": "~632–781K", "category": "World Labs / Stanford", "technical_flag": "Technical", "technical_rationale": "Computer vision, spatial intelligence, AI healthcare/policy", "why": "Major AI research and human-centered AI authority.", "source": "https://x.com/drfeifei" }, { "rank": 14, "name": "François Chollet", "handle": "@fchollet", "url": "https://x.com/fchollet", "followers": "~566–704K", "category": "Keras / ARC-AGI", "technical_flag": "Technical", "technical_rationale": "Reasoning benchmarks, Keras, AGI critique, architecture debates", "why": "Important benchmark/reasoning voice.", "source": "https://x.com/fchollet" }, { "rank": 15, "name": "Geoffrey Hinton", "handle": "@geoffreyhinton", "url": "https://x.com/geoffreyhinton", "followers": "~489–618K", "category": "Deep learning pioneer", "technical_flag": "Technical", "technical_rationale": "Foundational deep-learning researcher", "why": "High-authority voice on deep learning and AI risk.", "source": "https://x.com/geoffreyhinton" }, { "rank": 16, "name": "Dario Amodei", "handle": "@DarioAmodei", "url": "https://x.com/DarioAmodei", "followers": "~478.6K", "category": "Anthropic CEO", "technical_flag": "Semi-technical", "technical_rationale": "Frontier lab strategy/safety, not frequent code-level posting", "why": "Claude/frontier AI strategy and safety framing.", "source": "https://x.com/DarioAmodei" }, { "rank": 17, "name": "Jeff Dean", "handle": "@JeffDean", "url": "https://x.com/JeffDean", "followers": "~361–448K", "category": "Google DeepMind / Google Research", "technical_flag": "Technical", "technical_rationale": "Systems ML, scaling, Gemini/Google research", "why": "Elite systems + ML research credibility.", "source": "https://x.com/JeffDean" }, { "rank": 18, "name": "Clément Delangue", "handle": "@ClementDelangue", "url": "https://x.com/ClementDelangue", "followers": "~452.1K", "category": "Hugging Face CEO", "technical_flag": "Semi-technical", "technical_rationale": "Open-source model ecosystem/product strategy", "why": "Open-source AI distribution and community leader.", "source": "https://x.com/ClementDelangue" }, { "rank": 19, "name": "Alexandr Wang", "handle": "@alexandr_wang", "url": "https://x.com/alexandr_wang", "followers": "~541.7K", "category": "Meta AI / Scale founder", "technical_flag": "Semi-technical", "technical_rationale": "AI data/evals/enterprise strategy", "why": "Data, evals, and enterprise AI influence.", "source": "https://x.com/alexandr_wang" }, { "rank": 20, "name": "Jim Fan", "handle": "@DrJimFan", "url": "https://x.com/DrJimFan", "followers": "~435–486K", "category": "NVIDIA robotics / agents", "technical_flag": "Technical", "technical_rationale": "Robotics, simulation, embodied agents, model systems", "why": "High-signal physical AGI and robotics commentator.", "source": "https://x.com/DrJimFan" }, { "rank": 21, "name": "AK", "handle": "@_akhaliq", "url": "https://x.com/_akhaliq", "followers": "~500.4K", "category": "AI paper aggregator", "technical_flag": "Semi-technical", "technical_rationale": "Research-paper firehose; mostly aggregation", "why": "Important AI paper discovery/amplification account.", "source": "https://x.com/_akhaliq" }, { "rank": 22, "name": "Sebastian Raschka", "handle": "@rasbt", "url": "https://x.com/rasbt", "followers": "~476K", "category": "ML/LLM educator", "technical_flag": "Technical", "technical_rationale": "Implementation-level LLM/ML tutorials", "why": "Hands-on LLM training and architecture education.", "source": "https://x.com/rasbt" }, { "rank": 23, "name": "Soumith Chintala", "handle": "@soumithchintala", "url": "https://x.com/soumithchintala", "followers": "~247–311K", "category": "PyTorch / Meta", "technical_flag": "Technical", "technical_rationale": "Open-source ML framework engineering", "why": "PyTorch ecosystem credibility.", "source": "https://x.com/soumithchintala" }, { "rank": 24, "name": "Lilian Weng", "handle": "@lilianweng", "url": "https://x.com/lilianweng", "followers": "~88–267K", "category": "AI research / agents / safety", "technical_flag": "Technical", "technical_rationale": "Canonical technical essays on agents, RLHF, safety", "why": "Widely cited technical explainer.", "source": "https://x.com/lilianweng" }, { "rank": 25, "name": "Jeremy Howard", "handle": "@jeremyphoward", "url": "https://x.com/jeremyphoward", "followers": "~236–289K", "category": "fast.ai / Answer.AI", "technical_flag": "Technical", "technical_rationale": "Practical model training, pedagogy, open-source AI", "why": "Applied technical AI education and open models.", "source": "https://x.com/jeremyphoward" }, { "rank": 26, "name": "Emad Mostaque", "handle": "@EMostaque", "url": "https://x.com/EMostaque", "followers": "~328.3K", "category": "Stability AI founder", "technical_flag": "Semi-technical", "technical_rationale": "Open model strategy and ecosystem commentary", "why": "Open-source AI and generative media commentary.", "source": "https://x.com/EMostaque" }, { "rank": 27, "name": "Lior / AlphaSignal", "handle": "@AlphaSignalAI", "url": "https://x.com/AlphaSignalAI", "followers": "~80.4K", "category": "AI R&D newsletter", "technical_flag": "Semi-technical", "technical_rationale": "Technical research curation rather than original research", "why": "AI R&D paper/news amplification.", "source": "https://x.com/AlphaSignalAI" }, { "rank": 28, "name": "Logan Kilpatrick", "handle": "@OfficialLoganK", "url": "https://x.com/OfficialLoganK", "followers": "~339.2K", "category": "Google AI Studio / devrel", "technical_flag": "Semi-technical", "technical_rationale": "API/model developer-relations updates", "why": "Developer-facing AI launch signal.", "source": "https://x.com/OfficialLoganK" }, { "rank": 29, "name": "Simon Willison", "handle": "@simonw", "url": "https://x.com/simonw", "followers": "~154.5K", "category": "Developer / AI tooling analyst", "technical_flag": "Technical", "technical_rationale": "Hands-on testing, code, model behavior, security", "why": "Top practical technical evaluator of LLM tools.", "source": "https://x.com/simonw" }, { "rank": 30, "name": "Hamel Husain", "handle": "@HamelHusain", "url": "https://x.com/HamelHusain", "followers": "n/a", "category": "AI evals / applied LLM systems", "technical_flag": "Technical", "technical_rationale": "Evals, product quality, production LLM systems", "why": "High-signal practitioner voice on AI evals.", "source": "https://x.com/HamelHusain" }, { "rank": 31, "name": "Aidan McLaughlin", "handle": "@aidan_mclau", "url": "https://x.com/aidan_mclau", "followers": "~47.8K", "category": "OpenAI research", "technical_flag": "Technical", "technical_rationale": "Frontier model design/research perspective", "why": "Useful model-design and frontier-lab research signal.", "source": "https://x.com/aidan_mclau" }, { "rank": 32, "name": "Ian Goodfellow", "handle": "@goodfellow_ian", "url": "https://x.com/goodfellow_ian", "followers": "~315.7K", "category": "Deep learning researcher", "technical_flag": "Technical", "technical_rationale": "GANs and deep learning research", "why": "Foundational ML researcher.", "source": "https://x.com/goodfellow_ian" }, { "rank": 33, "name": "Oriol Vinyals", "handle": "@oriolvinyalsml", "url": "https://x.com/oriolvinyalsml", "followers": "~182.4K", "category": "Google DeepMind research", "technical_flag": "Technical", "technical_rationale": "Gemini, AlphaCode, sequence models", "why": "Major DeepMind technical leader.", "source": "https://x.com/oriolvinyalsml" }, { "rank": 34, "name": "Nando de Freitas", "handle": "@NandoDF", "url": "https://x.com/NandoDF", "followers": "~108.9K", "category": "Microsoft AI / ex-DeepMind", "technical_flag": "Technical", "technical_rationale": "Deep learning and superintelligence research", "why": "High-credibility research voice.", "source": "https://x.com/NandoDF" }, { "rank": 35, "name": "Abhishek Thakur", "handle": "@abhi1thakur", "url": "https://x.com/abhi1thakur", "followers": "~83.1K", "category": "Kaggle / Hugging Face AutoTrain", "technical_flag": "Technical", "technical_rationale": "Kaggle, AutoML, model-building practice", "why": "Applied ML competition and tooling expert.", "source": "https://x.com/abhi1thakur" }, { "rank": 36, "name": "Rowan Cheung", "handle": "@rowancheung", "url": "https://x.com/rowancheung", "followers": "~563.6K", "category": "AI newsletter / curator", "technical_flag": "Non-technical / media-business", "technical_rationale": "News/tool curation, not usually technical depth", "why": "Large AI news distribution channel.", "source": "https://x.com/rowancheung" }, { "rank": 37, "name": "Robert Scoble", "handle": "@Scobleizer", "url": "https://x.com/Scobleizer", "followers": "~536.6K", "category": "Tech futurist / AI commentator", "technical_flag": "Non-technical / media-business", "technical_rationale": "Technology commentary and amplification", "why": "Broad technology/AI reach.", "source": "https://x.com/Scobleizer" }, { "rank": 38, "name": "Kirk Borne", "handle": "@KirkDBorne", "url": "https://x.com/KirkDBorne", "followers": "~455.1K", "category": "Data science / AI influencer", "technical_flag": "Semi-technical", "technical_rationale": "Data science/AI curation; PhD background", "why": "Large data/AI amplification account.", "source": "https://x.com/KirkDBorne" }, { "rank": 39, "name": "Ronald van Loon", "handle": "@Ronald_vanLoon", "url": "https://x.com/Ronald_vanLoon", "followers": "~337.6K", "category": "AI/data influencer", "technical_flag": "Non-technical / media-business", "technical_rationale": "Business/AI content and amplification", "why": "Large B2B AI/data reach.", "source": "https://x.com/Ronald_vanLoon" }, { "rank": 40, "name": "Timnit Gebru", "handle": "@timnitGebru", "url": "https://x.com/timnitGebru", "followers": "~162.1K", "category": "AI ethics / DAIR", "technical_flag": "Semi-technical", "technical_rationale": "AI accountability research and policy", "why": "Influential AI ethics/accountability voice.", "source": "https://x.com/timnitGebru" }, { "rank": 41, "name": "Gary Marcus", "handle": "@GaryMarcus", "url": "https://x.com/GaryMarcus", "followers": "~213.5K", "category": "AI critic / cognitive scientist", "technical_flag": "Semi-technical", "technical_rationale": "Critiques of deep learning and LLM limits", "why": "Influential skeptical AI commentator.", "source": "https://x.com/GaryMarcus" }, { "rank": 42, "name": "Erik Brynjolfsson", "handle": "@erikbryn", "url": "https://x.com/erikbryn", "followers": "~210.2K", "category": "AI economics / Stanford", "technical_flag": "Semi-technical", "technical_rationale": "AI productivity and economics research", "why": "Important AI-economics KOL.", "source": "https://x.com/erikbryn" }, { "rank": 43, "name": "Paul Couvert", "handle": "@itsPaulAi", "url": "https://x.com/itsPaulAi", "followers": "~191.1K", "category": "AI educator / tools", "technical_flag": "Non-technical / media-business", "technical_rationale": "Tool/use-case education, usually not deep technical", "why": "Large AI tools/tutorial audience.", "source": "https://x.com/itsPaulAi" }, { "rank": 44, "name": "Pascal Bornet", "handle": "@pascal_bornet", "url": "https://x.com/pascal_bornet", "followers": "~158.1K", "category": "AI automation author", "technical_flag": "Non-technical / media-business", "technical_rationale": "Business automation/AI commentary", "why": "AI automation business influencer.", "source": "https://x.com/pascal_bornet" }, { "rank": 45, "name": "Vincent Boucher", "handle": "@ceobillionaire", "url": "https://x.com/ceobillionaire", "followers": "~153.7K", "category": "Montreal AI / AGI commentator", "technical_flag": "Semi-technical", "technical_rationale": "AI/AGI commentary and community account", "why": "AI community amplification.", "source": "https://x.com/ceobillionaire" }, { "rank": 46, "name": "Bernard Marr", "handle": "@bernardmarr", "url": "https://x.com/bernardmarr", "followers": "~138.7K", "category": "Tech futurist / author", "technical_flag": "Non-technical / media-business", "technical_rationale": "Business tech/AI thought leadership", "why": "Broad enterprise AI reach.", "source": "https://x.com/bernardmarr" }, { "rank": 47, "name": "Data Chaz", "handle": "@DataChaz", "url": "https://x.com/DataChaz", "followers": "~114.1K", "category": "Developer advocate / Streamlit", "technical_flag": "Semi-technical", "technical_rationale": "Python/data apps/LLM developer content", "why": "Practical AI/data app educator.", "source": "https://x.com/DataChaz" }, { "rank": 48, "name": "Randy Olson", "handle": "@randal_olson", "url": "https://x.com/randal_olson", "followers": "~110.7K", "category": "Data science / visualization", "technical_flag": "Semi-technical", "technical_rationale": "Data science and visualization content", "why": "Technical-ish data science influencer.", "source": "https://x.com/randal_olson" }, { "rank": 49, "name": "Jürgen Schmidhuber", "handle": "@SchmidhuberAI", "url": "https://x.com/SchmidhuberAI", "followers": "~98.1K", "category": "AI researcher", "technical_flag": "Technical", "technical_rationale": "Deep learning history and research claims", "why": "Longstanding ML research figure.", "source": "https://x.com/SchmidhuberAI" }, { "rank": 50, "name": "AI Tool Report", "handle": "@AIToolReport", "url": "https://x.com/AIToolReport", "followers": "~85.2K", "category": "AI tools newsletter", "technical_flag": "Non-technical / media-business", "technical_rationale": "AI tools/news aggregation", "why": "AI tools distribution channel.", "source": "https://x.com/AIToolReport" }, { "rank": 51, "name": "Kate Crawford", "handle": "@katecrawford", "url": "https://x.com/katecrawford", "followers": "~82.3K", "category": "AI ethics / policy researcher", "technical_flag": "Semi-technical", "technical_rationale": "Sociotechnical AI research and policy", "why": "Major AI ethics scholar.", "source": "https://x.com/katecrawford" }, { "rank": 52, "name": "Nathan Lands", "handle": "@nathanlands", "url": "https://x.com/nathanlands", "followers": "~77.3K", "category": "AI investor / podcast", "technical_flag": "Non-technical / media-business", "technical_rationale": "Investor/media commentary", "why": "AI startup/media reach.", "source": "https://x.com/nathanlands" }, { "rank": 53, "name": "Pedro Domingos", "handle": "@pmddomingos", "url": "https://x.com/pmddomingos", "followers": "~74.3K", "category": "ML professor / author", "technical_flag": "Technical", "technical_rationale": "ML theory and AI commentary", "why": "Recognized ML academic voice.", "source": "https://x.com/pmddomingos" }, { "rank": 54, "name": "Allie Miller", "handle": "@alliekmiller", "url": "https://x.com/alliekmiller", "followers": "~71.7K", "category": "AI business speaker", "technical_flag": "Non-technical / media-business", "technical_rationale": "Business AI education", "why": "Broad AI business audience.", "source": "https://x.com/alliekmiller" }, { "rank": 55, "name": "Karen Hao", "handle": "@_karenhao", "url": "https://x.com/_karenhao", "followers": "~61.6K", "category": "AI journalist", "technical_flag": "Non-technical / media-business", "technical_rationale": "Investigative AI reporting", "why": "Important AI media voice.", "source": "https://x.com/_karenhao" }, { "rank": 56, "name": "Santiago Valdarrama", "handle": "@svpino", "url": "https://x.com/svpino", "followers": "~383.3K", "category": "ML educator", "technical_flag": "Technical", "technical_rationale": "Hands-on ML engineering and system design", "why": "Large practical ML learning audience.", "source": "https://x.com/svpino" }, { "rank": 57, "name": "Bojan Tunguz", "handle": "@tunguz", "url": "https://x.com/tunguz", "followers": "~251.1K", "category": "ML engineer / Kaggle", "technical_flag": "Technical", "technical_rationale": "Data science/Kaggle/model-building", "why": "Applied ML/Kaggle KOL.", "source": "https://x.com/tunguz" }, { "rank": 58, "name": "Akshay Pachaar", "handle": "@akshay_pachaar", "url": "https://x.com/akshay_pachaar", "followers": "~240K", "category": "LLM/agents educator", "technical_flag": "Semi-technical", "technical_rationale": "Practical LLM/RAG/agent explainers", "why": "Large applied AI education audience.", "source": "https://x.com/akshay_pachaar" }, { "rank": 59, "name": "Ganapathi Pulipaka", "handle": "@gp_pulipaka", "url": "https://x.com/gp_pulipaka", "followers": "~153.2K", "category": "AI/data science influencer", "technical_flag": "Semi-technical", "technical_rationale": "AI/data science curation and commentary", "why": "Large AI/data reach.", "source": "https://x.com/gp_pulipaka" }, { "rank": 60, "name": "Hugo Larochelle", "handle": "@hugo_larochelle", "url": "https://x.com/hugo_larochelle", "followers": "~115.9K", "category": "Google DeepMind researcher", "technical_flag": "Technical", "technical_rationale": "Deep learning research and education", "why": "High-quality ML research voice.", "source": "https://x.com/hugo_larochelle" }, { "rank": 61, "name": "Peyman Milanfar", "handle": "@docmilanfar", "url": "https://x.com/docmilanfar", "followers": "~100.1K", "category": "Google computational imaging", "technical_flag": "Technical", "technical_rationale": "Vision/imaging ML research", "why": "Technical computer-vision KOL.", "source": "https://x.com/docmilanfar" }, { "rank": 62, "name": "Dan Kornas", "handle": "@dankornas", "url": "https://x.com/dankornas", "followers": "~87.1K", "category": "ML engineer / educator", "technical_flag": "Technical", "technical_rationale": "End-to-end ML engineering education", "why": "Practical ML engineering audience.", "source": "https://x.com/dankornas" }, { "rank": 63, "name": "Patrick Loeber", "handle": "@patloeber", "url": "https://x.com/patloeber", "followers": "~70.7K", "category": "Google DeepMind devrel", "technical_flag": "Semi-technical", "technical_rationale": "Developer education and DeepMind-facing updates", "why": "AI developer education.", "source": "https://x.com/patloeber" }, { "rank": 64, "name": "Kevin Patrick Murphy", "handle": "@sirbayes", "url": "https://x.com/sirbayes", "followers": "~64.7K", "category": "Google DeepMind researcher", "technical_flag": "Technical", "technical_rationale": "Bayesian ML, probabilistic modeling", "why": "Authoritative ML textbook/research voice.", "source": "https://x.com/sirbayes" }, { "rank": 65, "name": "Gautam Kamath", "handle": "@thegautamkamath", "url": "https://x.com/thegautamkamath", "followers": "~59.7K", "category": "ML professor / Vector Institute", "technical_flag": "Technical", "technical_rationale": "Privacy, statistics, ML theory", "why": "Technical ML theory voice.", "source": "https://x.com/thegautamkamath" }, { "rank": 66, "name": "Sara Hooker", "handle": "@sarahookr", "url": "https://x.com/sarahookr", "followers": "~53.3K", "category": "Cohere Labs / model efficiency", "technical_flag": "Technical", "technical_rationale": "ML efficiency, multilingual/multimodal research", "why": "Important efficient/open AI researcher.", "source": "https://x.com/sarahookr" }, { "rank": 67, "name": "Durk Kingma", "handle": "@dpkingma", "url": "https://x.com/dpkingma", "followers": "~51.3K", "category": "Anthropic / VAE/Adam inventor", "technical_flag": "Technical", "technical_rationale": "Core ML optimization and generative modeling", "why": "Foundational ML researcher.", "source": "https://x.com/dpkingma" }, { "rank": 68, "name": "Andriy Burkov", "handle": "@burkov", "url": "https://x.com/burkov", "followers": "~50.8K", "category": "ML author", "technical_flag": "Technical", "technical_rationale": "ML/LLM books and practical education", "why": "Strong ML educational voice.", "source": "https://x.com/burkov" }, { "rank": 69, "name": "Jean de Nyandwi", "handle": "@Jeande_d", "url": "https://x.com/Jeande_d", "followers": "~46.9K", "category": "CMU / NLP evals", "technical_flag": "Technical", "technical_rationale": "Multimodal NLP, post-training, data, evals", "why": "Rising technical AI research voice.", "source": "https://x.com/Jeande_d" }, { "rank": 70, "name": "Shakir Mohamed", "handle": "@shakir_za", "url": "https://x.com/shakir_za", "followers": "~41.7K", "category": "DeepMind researcher", "technical_flag": "Technical", "technical_rationale": "ML research and AI for social good", "why": "Technical DeepMind/social-good voice.", "source": "https://x.com/shakir_za" }, { "rank": 71, "name": "Ferenc Huszár", "handle": "@fhuszar", "url": "https://x.com/fhuszar", "followers": "~37.9K", "category": "ML professor", "technical_flag": "Technical", "technical_rationale": "Bayesian/statistical ML commentary", "why": "Technical ML/statistics KOL.", "source": "https://x.com/fhuszar" }, { "rank": 72, "name": "Andrew Gordon Wilson", "handle": "@andrewgwils", "url": "https://x.com/andrewgwils", "followers": "~35.5K", "category": "ML professor", "technical_flag": "Technical", "technical_rationale": "Bayesian/deep learning theory", "why": "Technical ML academic.", "source": "https://x.com/andrewgwils" }, { "rank": 73, "name": "Ben Recht", "handle": "@beenwrekt", "url": "https://x.com/beenwrekt", "followers": "~32.9K", "category": "Berkeley ML/control", "technical_flag": "Technical", "technical_rationale": "Optimization, ML systems, critical analysis", "why": "Sharp technical ML critic.", "source": "https://x.com/beenwrekt" }, { "rank": 74, "name": "Zoubin Ghahramani", "handle": "@ZoubinGhahrama1", "url": "https://x.com/ZoubinGhahrama1", "followers": "~32.6K", "category": "Google DeepMind / Cambridge", "technical_flag": "Technical", "technical_rationale": "Probabilistic ML and AI research leadership", "why": "Senior ML research authority.", "source": "https://x.com/ZoubinGhahrama1" }, { "rank": 75, "name": "Jakob Foerster", "handle": "@j_foerst", "url": "https://x.com/j_foerst", "followers": "~22.7K", "category": "Oxford/Meta ML researcher", "technical_flag": "Technical", "technical_rationale": "Multi-agent RL and ML research", "why": "Technical RL/agents researcher.", "source": "https://x.com/j_foerst" }, { "rank": 76, "name": "Nils Reimers", "handle": "@Nils_Reimers", "url": "https://x.com/Nils_Reimers", "followers": "~14.6K", "category": "Cohere / SBERT", "technical_flag": "Technical", "technical_rationale": "Embeddings, retrieval, AI search", "why": "Important technical retrieval/embedding voice.", "source": "https://x.com/Nils_Reimers" }, { "rank": 77, "name": "Nathan Raw", "handle": "@_nateraw", "url": "https://x.com/_nateraw", "followers": "~9.5K", "category": "ML hacker / Hugging Face alumnus", "technical_flag": "Technical", "technical_rationale": "Open-source ML tooling", "why": "Hands-on ML builder.", "source": "https://x.com/_nateraw" }, { "rank": 78, "name": "Tim Scarfe", "handle": "@ecsquendor", "url": "https://x.com/ecsquendor", "followers": "~8.7K", "category": "ML Street Talk / ML engineer", "technical_flag": "Semi-technical", "technical_rationale": "Technical AI podcast and ML commentary", "why": "AI research interview/amplification channel.", "source": "https://x.com/ecsquendor" }, { "rank": 79, "name": "Alicia Curth", "handle": "@AliciaCurth", "url": "https://x.com/AliciaCurth", "followers": "~4.7K", "category": "ML/statistics researcher", "technical_flag": "Technical", "technical_rationale": "Statistical ML intuition and methods", "why": "Technical ML/statistics voice.", "source": "https://x.com/AliciaCurth" }, { "rank": 80, "name": "Yoshua Bengio", "handle": "@yoshuabengio", "url": "https://x.com/yoshuabengio", "followers": "n/a", "category": "Deep learning pioneer", "technical_flag": "Technical", "technical_rationale": "Foundational deep-learning research", "why": "Turing laureate and major AI safety/research voice.", "source": "https://x.com/yoshuabengio" }, { "rank": 81, "name": "Chris Olah", "handle": "@chrisolah", "url": "https://x.com/chrisolah", "followers": "n/a", "category": "Anthropic / interpretability", "technical_flag": "Technical", "technical_rationale": "Mechanistic interpretability research", "why": "Core interpretability KOL.", "source": "https://x.com/chrisolah" }, { "rank": 82, "name": "Anca Dragan", "handle": "@ancadragan", "url": "https://x.com/ancadragan", "followers": "n/a", "category": "Berkeley robotics / HRI", "technical_flag": "Technical", "technical_rationale": "Human-robot interaction and robotics", "why": "Technical robotics/HRI authority.", "source": "https://x.com/ancadragan" }, { "rank": 83, "name": "Pieter Abbeel", "handle": "@pabbeel", "url": "https://x.com/pabbeel", "followers": "n/a", "category": "Berkeley robotics / Covariant", "technical_flag": "Technical", "technical_rationale": "Robotics and reinforcement learning", "why": "Major robotics/RL KOL.", "source": "https://x.com/pabbeel" }, { "rank": 84, "name": "Daphne Koller", "handle": "@DaphneKoller", "url": "https://x.com/DaphneKoller", "followers": "n/a", "category": "AI healthcare / Coursera", "technical_flag": "Semi-technical", "technical_rationale": "AI in healthcare and biology entrepreneurship", "why": "AI-healthcare founder/research voice.", "source": "https://x.com/DaphneKoller" }, { "rank": 85, "name": "Chelsea Finn", "handle": "@chelseabfinn", "url": "https://x.com/chelseabfinn", "followers": "n/a", "category": "Stanford robotics / meta-learning", "technical_flag": "Technical", "technical_rationale": "Robotics, meta-learning, imitation learning", "why": "High-signal robotics/learning researcher.", "source": "https://x.com/chelseabfinn" }, { "rank": 86, "name": "Sergey Levine", "handle": "@svlevine", "url": "https://x.com/svlevine", "followers": "n/a", "category": "Berkeley robotics / RL", "technical_flag": "Technical", "technical_rationale": "Deep RL and robot learning", "why": "Top technical robotics/RL researcher.", "source": "https://x.com/svlevine" }, { "rank": 87, "name": "Been Kim", "handle": "@beenkim", "url": "https://x.com/beenkim", "followers": "n/a", "category": "Google interpretability", "technical_flag": "Technical", "technical_rationale": "Explainability/interpretability research", "why": "Important technical interpretability voice.", "source": "https://x.com/beenkim" }, { "rank": 88, "name": "Samy Bengio", "handle": "@samybengio", "url": "https://x.com/samybengio", "followers": "n/a", "category": "Apple / ex-Google Brain", "technical_flag": "Technical", "technical_rationale": "Deep learning research leadership", "why": "Senior deep learning researcher.", "source": "https://x.com/samybengio" }, { "rank": 89, "name": "Kyunghyun Cho", "handle": "@kchon2020", "url": "https://x.com/kchon2020", "followers": "n/a", "category": "NYU / NLP", "technical_flag": "Technical", "technical_rationale": "Neural machine translation, NLP, deep learning", "why": "Technical NLP researcher.", "source": "https://x.com/kchon2020" }, { "rank": 90, "name": "Max Welling", "handle": "@wellingmax", "url": "https://x.com/wellingmax", "followers": "n/a", "category": "CuspAI / ML professor", "technical_flag": "Technical", "technical_rationale": "ML theory, generative science, Bayesian/deep learning", "why": "Technical ML/science-AI voice.", "source": "https://x.com/wellingmax" }, { "rank": 91, "name": "David Ha", "handle": "@hardmaru", "url": "https://x.com/hardmaru", "followers": "n/a", "category": "AI researcher / generative systems", "technical_flag": "Technical", "technical_rationale": "World models, generative systems, creative AI", "why": "Technical and creative AI research voice.", "source": "https://x.com/hardmaru" }, { "rank": 92, "name": "Alex Graves", "handle": "@alexjgraves", "url": "https://x.com/alexjgraves", "followers": "n/a", "category": "DeepMind researcher", "technical_flag": "Technical", "technical_rationale": "RNNs, sequence learning, deep learning", "why": "Foundational sequence-model researcher.", "source": "https://x.com/alexjgraves" }, { "rank": 93, "name": "Quoc Le", "handle": "@quocleix", "url": "https://x.com/quocleix", "followers": "n/a", "category": "Google Brain / AutoML", "technical_flag": "Technical", "technical_rationale": "AutoML and deep learning research", "why": "Senior technical Google AI researcher.", "source": "https://x.com/quocleix" }, { "rank": 94, "name": "Leslie Kaelbling", "handle": "@lpkaelbling", "url": "https://x.com/lpkaelbling", "followers": "n/a", "category": "MIT AI/robotics", "technical_flag": "Technical", "technical_rationale": "Planning, robotics, AI fundamentals", "why": "Senior AI/robotics authority.", "source": "https://x.com/lpkaelbling" }, { "rank": 95, "name": "Mustafa Suleyman", "handle": "@mustafasuleymn", "url": "https://x.com/mustafasuleymn", "followers": "n/a", "category": "Microsoft AI / DeepMind cofounder", "technical_flag": "Semi-technical", "technical_rationale": "AI strategy, product, policy", "why": "Major lab/product/policy voice.", "source": "https://x.com/mustafasuleymn" }, { "rank": 96, "name": "Reid Hoffman", "handle": "@reidhoffman", "url": "https://x.com/reidhoffman", "followers": "n/a", "category": "AI investor / author", "technical_flag": "Non-technical / media-business", "technical_rationale": "AI entrepreneurship and investment", "why": "Large business/AI network.", "source": "https://x.com/reidhoffman" }, { "rank": 97, "name": "Daniela Amodei", "handle": "@daniela_amodei", "url": "https://x.com/daniela_amodei", "followers": "n/a", "category": "Anthropic cofounder", "technical_flag": "Semi-technical", "technical_rationale": "Anthropic strategy and safety operations", "why": "Frontier-lab leadership voice.", "source": "https://x.com/daniela_amodei" }, { "rank": 98, "name": "Aidan Gomez", "handle": "@aidangomez", "url": "https://x.com/aidangomez", "followers": "n/a", "category": "Cohere CEO / Transformer coauthor", "technical_flag": "Technical", "technical_rationale": "Transformer coauthor and enterprise LLM builder", "why": "Technical founder with model lineage.", "source": "https://x.com/aidangomez" }, { "rank": 99, "name": "Palantir", "handle": "@PalantirTech", "url": "https://x.com/PalantirTech", "followers": "n/a", "category": "AI/data platform company", "technical_flag": "Semi-technical", "technical_rationale": "Enterprise AI/data platform narratives", "why": "Large enterprise AI channel.", "source": "https://x.com/PalantirTech" }, { "rank": 100, "name": "Jensen Huang", "handle": "@JensenHuang", "url": "https://x.com/JensenHuang", "followers": "n/a", "category": "NVIDIA CEO", "technical_flag": "Semi-technical", "technical_rationale": "AI hardware/platform strategy", "why": "AI compute platform narrative-setter.", "source": "https://x.com/JensenHuang" }, { "rank": 101, "name": "Rachel Thomas", "handle": "@math_rachel", "url": "https://x.com/math_rachel", "followers": "n/a", "category": "fast.ai / AI education", "technical_flag": "Technical", "technical_rationale": "Practical ML and responsible AI education", "why": "Technical educator and fast.ai cofounder.", "source": "https://x.com/math_rachel" }, { "rank": 102, "name": "Thomas Wolf", "handle": "@Thom_Wolf", "url": "https://x.com/Thom_Wolf", "followers": "n/a", "category": "Hugging Face cofounder", "technical_flag": "Technical", "technical_rationale": "Transformers/open-source AI engineering", "why": "Core open-source AI platform builder.", "source": "https://x.com/Thom_Wolf" }, { "rank": 103, "name": "Julien Chaumond", "handle": "@julien_c", "url": "https://x.com/julien_c", "followers": "n/a", "category": "Hugging Face CTO/cofounder", "technical_flag": "Technical", "technical_rationale": "Model hub/open-source platform engineering", "why": "Important open-source AI infra voice.", "source": "https://x.com/julien_c" }, { "rank": 104, "name": "Naveen Rao", "handle": "@NaveenGRao", "url": "https://x.com/NaveenGRao", "followers": "n/a", "category": "AI chips / Databricks", "technical_flag": "Semi-technical", "technical_rationale": "AI hardware and systems entrepreneurship", "why": "AI compute/systems founder.", "source": "https://x.com/NaveenGRao" }, { "rank": 105, "name": "Vinod Khosla", "handle": "@vkhosla", "url": "https://x.com/vkhosla", "followers": "n/a", "category": "AI investor", "technical_flag": "Non-technical / media-business", "technical_rationale": "AI investment and market theses", "why": "Major AI investment voice.", "source": "https://x.com/vkhosla" }, { "rank": 106, "name": "Sebastian Thrun", "handle": "@SebastianThrun", "url": "https://x.com/SebastianThrun", "followers": "n/a", "category": "Udacity / robotics", "technical_flag": "Technical", "technical_rationale": "Robotics, autonomous vehicles, education", "why": "Robotics/AI education pioneer.", "source": "https://x.com/SebastianThrun" }, { "rank": 107, "name": "Peter Diamandis", "handle": "@PeterDiamandis", "url": "https://x.com/PeterDiamandis", "followers": "n/a", "category": "Futurist / XPRIZE", "technical_flag": "Non-technical / media-business", "technical_rationale": "AI futurism and entrepreneurship", "why": "Broad innovation audience.", "source": "https://x.com/PeterDiamandis" }, { "rank": 108, "name": "Ray Kurzweil", "handle": "@raykurzweil", "url": "https://x.com/raykurzweil", "followers": "n/a", "category": "AI futurist", "technical_flag": "Semi-technical", "technical_rationale": "Singularity/AI future commentary", "why": "Longstanding AI futurist voice.", "source": "https://x.com/raykurzweil" }, { "rank": 109, "name": "Joy Buolamwini", "handle": "@jovialjoy", "url": "https://x.com/jovialjoy", "followers": "n/a", "category": "Algorithmic Justice League", "technical_flag": "Semi-technical", "technical_rationale": "Algorithmic bias/accountability research", "why": "AI fairness and policy KOL.", "source": "https://x.com/jovialjoy" }, { "rank": 110, "name": "Rumman Chowdhury", "handle": "@ruchowdh", "url": "https://x.com/ruchowdh", "followers": "n/a", "category": "AI ethics / governance", "technical_flag": "Semi-technical", "technical_rationale": "Responsible AI and governance", "why": "Important AI governance voice.", "source": "https://x.com/ruchowdh" }, { "rank": 111, "name": "Abeba Birhane", "handle": "@Abebab", "url": "https://x.com/Abebab", "followers": "n/a", "category": "AI accountability researcher", "technical_flag": "Semi-technical", "technical_rationale": "AI bias/eval/accountability research", "why": "Strong responsible-AI researcher.", "source": "https://x.com/Abebab" }, { "rank": 112, "name": "Rediet Abebe", "handle": "@red_abebe", "url": "https://x.com/red_abebe", "followers": "n/a", "category": "AI for social good", "technical_flag": "Technical", "technical_rationale": "Algorithms and social-good AI research", "why": "Technical/public-interest AI voice.", "source": "https://x.com/red_abebe" }, { "rank": 113, "name": "Meredith Whittaker", "handle": "@mer__edith", "url": "https://x.com/mer__edith", "followers": "n/a", "category": "Signal / AI critic", "technical_flag": "Non-technical / media-business", "technical_rationale": "AI politics, privacy, surveillance critique", "why": "Important policy/privacy voice.", "source": "https://x.com/mer__edith" }, { "rank": 114, "name": "Cathy O’Neil", "handle": "@mathbabedotorg", "url": "https://x.com/mathbabedotorg", "followers": "n/a", "category": "Algorithmic accountability", "technical_flag": "Semi-technical", "technical_rationale": "Risk, fairness, automated decision-making", "why": "Public algorithmic-risk voice.", "source": "https://x.com/mathbabedotorg" }, { "rank": 115, "name": "Safiya Noble", "handle": "@safiyanoble", "url": "https://x.com/safiyanoble", "followers": "n/a", "category": "AI/search equity scholar", "technical_flag": "Non-technical / media-business", "technical_rationale": "Critical AI/search scholarship", "why": "Important sociotechnical AI voice.", "source": "https://x.com/safiyanoble" }, { "rank": 116, "name": "Ruha Benjamin", "handle": "@ruha9", "url": "https://x.com/ruha9", "followers": "n/a", "category": "Technology and society scholar", "technical_flag": "Non-technical / media-business", "technical_rationale": "Race, technology, AI society", "why": "Sociotechnical AI influence.", "source": "https://x.com/ruha9" }, { "rank": 117, "name": "Jack Clark", "handle": "@jackclarkSF", "url": "https://x.com/jackclarkSF", "followers": "n/a", "category": "Anthropic cofounder / Import AI", "technical_flag": "Semi-technical", "technical_rationale": "AI policy, capability trends, technical summaries", "why": "Measured AI policy/technical newsletter voice.", "source": "https://x.com/jackclarkSF" }, { "rank": 118, "name": "Melanie Mitchell", "handle": "@MelMitchell1", "url": "https://x.com/MelMitchell1", "followers": "n/a", "category": "Complexity / AI researcher", "technical_flag": "Technical", "technical_rationale": "AI reasoning, analogy, cognitive science", "why": "Technical and skeptical AI researcher.", "source": "https://x.com/MelMitchell1" }, { "rank": 119, "name": "Aurélien Géron", "handle": "@aureliengeron", "url": "https://x.com/aureliengeron", "followers": "n/a", "category": "ML author", "technical_flag": "Technical", "technical_rationale": "Hands-on ML education", "why": "Popular practical ML educator.", "source": "https://x.com/aureliengeron" }, { "rank": 120, "name": "Chip Huyen", "handle": "@chipro", "url": "https://x.com/chipro", "followers": "n/a", "category": "AI engineering author", "technical_flag": "Technical", "technical_rationale": "ML systems, AI engineering, evals", "why": "High-signal AI engineering voice.", "source": "https://x.com/chipro" }, { "rank": 121, "name": "Monica Rogati", "handle": "@mrogati", "url": "https://x.com/mrogati", "followers": "n/a", "category": "Data science leader", "technical_flag": "Technical", "technical_rationale": "Data/AI product strategy and systems", "why": "Technical data/AI product KOL.", "source": "https://x.com/mrogati" }, { "rank": 122, "name": "Inioluwa Deborah Raji", "handle": "@rajiinio", "url": "https://x.com/rajiinio", "followers": "n/a", "category": "AI accountability researcher", "technical_flag": "Semi-technical", "technical_rationale": "Model auditing and accountability", "why": "Important AI auditing voice.", "source": "https://x.com/rajiinio" }, { "rank": 123, "name": "Arvind Narayanan", "handle": "@random_walker", "url": "https://x.com/random_walker", "followers": "n/a", "category": "Princeton AI/society", "technical_flag": "Technical", "technical_rationale": "AI, fairness, privacy, crypto, accountability", "why": "Technical AI and policy critic.", "source": "https://x.com/random_walker" }, { "rank": 124, "name": "Vincent Warmerdam", "handle": "@fishnets88", "url": "https://x.com/fishnets88", "followers": "n/a", "category": "Calmcode / ML tooling", "technical_flag": "Technical", "technical_rationale": "Practical ML education and tooling", "why": "High-signal practical technical educator.", "source": "https://x.com/fishnets88" }, { "rank": 125, "name": "Sasha Luccioni", "handle": "@SashaMTL", "url": "https://x.com/SashaMTL", "followers": "n/a", "category": "Hugging Face / AI climate", "technical_flag": "Semi-technical", "technical_rationale": "AI energy, climate, responsible AI evaluation", "why": "Technical responsible-AI voice.", "source": "https://x.com/SashaMTL" }, { "rank": 126, "name": "Miles Brundage", "handle": "@Miles_Brundage", "url": "https://x.com/Miles_Brundage", "followers": "n/a", "category": "AI policy/safety", "technical_flag": "Semi-technical", "technical_rationale": "AI governance and frontier-lab policy", "why": "Important policy/safety KOL.", "source": "https://x.com/Miles_Brundage" }, { "rank": 127, "name": "Jan Leike", "handle": "@janleike", "url": "https://x.com/janleike", "followers": "n/a", "category": "Anthropic / alignment", "technical_flag": "Technical", "technical_rationale": "Alignment, scalable oversight, RLHF", "why": "Top technical alignment researcher.", "source": "https://x.com/janleike" }, { "rank": 128, "name": "Paul Christiano", "handle": "@paulc443", "url": "https://x.com/paulc443", "followers": "n/a", "category": "Alignment Research Center", "technical_flag": "Technical", "technical_rationale": "Technical AI alignment theory/evals", "why": "Core technical alignment voice.", "source": "https://x.com/paulc443" }, { "rank": 129, "name": "Oren Etzioni", "handle": "@etzioni", "url": "https://x.com/etzioni", "followers": "n/a", "category": "AI2 founder", "technical_flag": "Semi-technical", "technical_rationale": "AI research, policy, entrepreneurship", "why": "Senior AI institution builder.", "source": "https://x.com/etzioni" }, { "rank": 130, "name": "Percy Liang", "handle": "@percyliang", "url": "https://x.com/percyliang", "followers": "~106.5K", "category": "Stanford NLP / HELM", "technical_flag": "Technical", "technical_rationale": "Benchmarking, NLP, HELM/MedHELM", "why": "Top LLM-evaluation researcher.", "source": "https://x.com/percyliang" }, { "rank": 131, "name": "Jacob Andreas", "handle": "@jacobandreas", "url": "https://x.com/jacobandreas", "followers": "n/a", "category": "MIT NLP", "technical_flag": "Technical", "technical_rationale": "Language grounding, agents, NLP research", "why": "Technical NLP/agents researcher.", "source": "https://x.com/jacobandreas" }, { "rank": 132, "name": "Dan Hendrycks", "handle": "@DanHendrycks", "url": "https://x.com/DanHendrycks", "followers": "n/a", "category": "Center for AI Safety", "technical_flag": "Technical", "technical_rationale": "Benchmarks, model risks, WMDP/MMLU-style evals", "why": "Important eval/safety benchmark creator.", "source": "https://x.com/DanHendrycks" }, { "rank": 133, "name": "Scott Aaronson", "handle": "@scottaaronson", "url": "https://x.com/scottaaronson", "followers": "n/a", "category": "Theoretical CS / OpenAI alumnus", "technical_flag": "Technical", "technical_rationale": "Complexity theory, quantum, AI safety commentary", "why": "Technical theory/safety voice.", "source": "https://x.com/scottaaronson" }, { "rank": 134, "name": "Swyx", "handle": "@swyx", "url": "https://x.com/swyx", "followers": "n/a", "category": "AI engineering / Latent Space", "technical_flag": "Technical", "technical_rationale": "Agents, RAG, evals, AI engineering stack", "why": "Top AI engineering community builder.", "source": "https://x.com/swyx" }, { "rank": 135, "name": "Mckay Wrigley", "handle": "@mckaywrigley", "url": "https://x.com/mckaywrigley", "followers": "n/a", "category": "AI builder / Chatbot UI", "technical_flag": "Semi-technical", "technical_rationale": "AI app demos and builder workflows", "why": "Practical AI builder influence.", "source": "https://x.com/mckaywrigley" }, { "rank": 136, "name": "Deedy Das", "handle": "@deedydas", "url": "https://x.com/deedydas", "followers": "n/a", "category": "AI investor / curator", "technical_flag": "Semi-technical", "technical_rationale": "Model release summaries and benchmark takes", "why": "High-signal AI startup/model curator.", "source": "https://x.com/deedydas" }, { "rank": 137, "name": "Bindu Reddy", "handle": "@bindureddy", "url": "https://x.com/bindureddy", "followers": "n/a", "category": "Abacus AI CEO", "technical_flag": "Semi-technical", "technical_rationale": "AI capability/startup commentary", "why": "High-engagement AI product/founder voice.", "source": "https://x.com/bindureddy" }, { "rank": 138, "name": "Ethan Mollick", "handle": "@emollick", "url": "https://x.com/emollick", "followers": "n/a", "category": "Wharton professor / AI use cases", "technical_flag": "Semi-technical", "technical_rationale": "Empirical AI use and workplace experiments", "why": "Major practical AI adoption voice.", "source": "https://x.com/emollick" }, { "rank": 139, "name": "Linus Lee", "handle": "@thesephist", "url": "https://x.com/thesephist", "followers": "n/a", "category": "Notion AI researcher", "technical_flag": "Technical", "technical_rationale": "LLM internals, embeddings, AI-native UX", "why": "Technical AI product/research voice.", "source": "https://x.com/thesephist" }, { "rank": 140, "name": "Nathan Labenz", "handle": "@labenz", "url": "https://x.com/labenz", "followers": "n/a", "category": "Cognitive Revolution podcast", "technical_flag": "Semi-technical", "technical_rationale": "Model eval discussions, AI interviews", "why": "Strong AI interview/eval commentator.", "source": "https://x.com/labenz" }, { "rank": 141, "name": "Nathan Lambert", "handle": "@natolambert", "url": "https://x.com/natolambert", "followers": "n/a", "category": "AI2 / Interconnects", "technical_flag": "Technical", "technical_rationale": "RLHF, post-training, open models", "why": "Top technical post-training/open-model voice.", "source": "https://x.com/natolambert" }, { "rank": 142, "name": "DAIR.AI", "handle": "@DAIR_AI", "url": "https://x.com/DAIR_AI", "followers": "n/a", "category": "AI research education", "technical_flag": "Semi-technical", "technical_rationale": "Paper explainers and ML education", "why": "Research-paper amplification and education.", "source": "https://x.com/DAIR_AI" }, { "rank": 143, "name": "arXiv", "handle": "@arxiv_org", "url": "https://x.com/arxiv_org", "followers": "n/a", "category": "Research preprint platform", "technical_flag": "Technical", "technical_rationale": "Primary research feed", "why": "Raw research discovery channel.", "source": "https://x.com/arxiv_org" }, { "rank": 144, "name": "AlphaXiv", "handle": "@alphaxiv", "url": "https://x.com/alphaxiv", "followers": "n/a", "category": "Paper discovery platform", "technical_flag": "Technical", "technical_rationale": "Social layer for arXiv/papers", "why": "AI paper discovery/amplification.", "source": "https://x.com/alphaxiv" }, { "rank": 145, "name": "Cameron Wolfe", "handle": "@cwolferesearch", "url": "https://x.com/cwolferesearch", "followers": "n/a", "category": "AI researcher / writer", "technical_flag": "Technical", "technical_rationale": "Paper breakdowns and Deep Learning Focus", "why": "High-signal technical AI explainer.", "source": "https://x.com/cwolferesearch" }, { "rank": 146, "name": "Vikhyat K.", "handle": "@vikhyatk", "url": "https://x.com/vikhyatk", "followers": "n/a", "category": "Moondream / small VLMs", "technical_flag": "Technical", "technical_rationale": "Small-model/VLM engineering", "why": "Practical model builder.", "source": "https://x.com/vikhyatk" }, { "rank": 147, "name": "OpenAI", "handle": "@OpenAI", "url": "https://x.com/OpenAI", "followers": "n/a", "category": "Frontier AI lab", "technical_flag": "Semi-technical", "technical_rationale": "Official launches and benchmark posts", "why": "Primary OpenAI launch channel.", "source": "https://x.com/OpenAI" }, { "rank": 148, "name": "Anthropic", "handle": "@AnthropicAI", "url": "https://x.com/AnthropicAI", "followers": "n/a", "category": "Frontier AI lab", "technical_flag": "Semi-technical", "technical_rationale": "Claude releases and safety research", "why": "Primary Anthropic launch channel.", "source": "https://x.com/AnthropicAI" }, { "rank": 149, "name": "Google DeepMind", "handle": "@GoogleDeepMind", "url": "https://x.com/GoogleDeepMind", "followers": "~1.2M", "category": "Frontier AI/science lab", "technical_flag": "Technical", "technical_rationale": "Research releases, AlphaFold/Gemini science AI", "why": "Major technical AI lab channel.", "source": "https://x.com/GoogleDeepMind" }, { "rank": 150, "name": "Meta AI", "handle": "@MetaAI", "url": "https://x.com/MetaAI", "followers": "n/a", "category": "Open models / AI lab", "technical_flag": "Technical", "technical_rationale": "Llama/open model research and product", "why": "Key open-model lab channel.", "source": "https://x.com/MetaAI" }, { "rank": 151, "name": "Mistral AI", "handle": "@MistralAI", "url": "https://x.com/MistralAI", "followers": "n/a", "category": "Frontier/open model lab", "technical_flag": "Technical", "technical_rationale": "Open/foundation model releases", "why": "European frontier model channel.", "source": "https://x.com/MistralAI" }, { "rank": 152, "name": "Hugging Face", "handle": "@huggingface", "url": "https://x.com/huggingface", "followers": "~689K", "category": "Open-source AI platform", "technical_flag": "Semi-technical", "technical_rationale": "Model/dataset/leaderboard ecosystem", "why": "High-impact open AI distribution.", "source": "https://x.com/huggingface" }, { "rank": 153, "name": "NVIDIA", "handle": "@nvidia", "url": "https://x.com/nvidia", "followers": "n/a", "category": "AI compute platform", "technical_flag": "Semi-technical", "technical_rationale": "GPU/platform launches and AI hardware", "why": "AI infrastructure narrative channel.", "source": "https://x.com/nvidia" }, { "rank": 154, "name": "Cohere", "handle": "@cohere", "url": "https://x.com/cohere", "followers": "n/a", "category": "Enterprise LLM lab", "technical_flag": "Semi-technical", "technical_rationale": "Enterprise model releases and retrieval/RAG context", "why": "Enterprise LLM platform channel.", "source": "https://x.com/cohere" }, { "rank": 155, "name": "Perplexity", "handle": "@perplexity_ai", "url": "https://x.com/perplexity_ai", "followers": "n/a", "category": "AI search company", "technical_flag": "Semi-technical", "technical_rationale": "AI search product releases", "why": "Consumer AI search channel.", "source": "https://x.com/perplexity_ai" }, { "rank": 156, "name": "Together AI", "handle": "@togethercompute", "url": "https://x.com/togethercompute", "followers": "n/a", "category": "Open model infra", "technical_flag": "Technical", "technical_rationale": "Inference/fine-tuning/open model infra", "why": "Open AI infrastructure channel.", "source": "https://x.com/togethercompute" }, { "rank": 157, "name": "xAI", "handle": "@xai", "url": "https://x.com/xai", "followers": "n/a", "category": "Frontier AI lab", "technical_flag": "Semi-technical", "technical_rationale": "Grok releases and xAI research", "why": "xAI primary launch channel.", "source": "https://x.com/xai" }, { "rank": 158, "name": "LM Arena", "handle": "@lmarena_ai", "url": "https://x.com/lmarena_ai", "followers": "n/a", "category": "Model leaderboard/evals", "technical_flag": "Technical", "technical_rationale": "Chatbot Arena / model rankings", "why": "Important public model benchmark channel.", "source": "https://x.com/lmarena_ai" }, { "rank": 159, "name": "OpenRouter", "handle": "@OpenRouterAI", "url": "https://x.com/OpenRouterAI", "followers": "n/a", "category": "Model routing/API platform", "technical_flag": "Semi-technical", "technical_rationale": "Model availability and routing data", "why": "Useful model market/API signal.", "source": "https://x.com/OpenRouterAI" }, { "rank": 160, "name": "Replicate", "handle": "@replicate", "url": "https://x.com/replicate", "followers": "n/a", "category": "Model hosting/API", "technical_flag": "Technical", "technical_rationale": "Open model deployment and demos", "why": "Model deployment ecosystem channel.", "source": "https://x.com/replicate" }, { "rank": 161, "name": "Weights & Biases", "handle": "@weights_biases", "url": "https://x.com/weights_biases", "followers": "n/a", "category": "MLOps / evals", "technical_flag": "Semi-technical", "technical_rationale": "Experiment tracking, evals, ML tooling", "why": "MLOps and AI eval community.", "source": "https://x.com/weights_biases" }, { "rank": 162, "name": "LangChain", "handle": "@LangChainAI", "url": "https://x.com/LangChainAI", "followers": "n/a", "category": "LLM app framework", "technical_flag": "Technical", "technical_rationale": "Agents, RAG, LangGraph, app tooling", "why": "Key AI agent/RAG developer channel.", "source": "https://x.com/LangChainAI" }, { "rank": 163, "name": "LlamaIndex", "handle": "@llama_index", "url": "https://x.com/llama_index", "followers": "n/a", "category": "LLM data/RAG framework", "technical_flag": "Technical", "technical_rationale": "RAG, retrieval, agents, data connectors", "why": "Important AI engineering framework.", "source": "https://x.com/llama_index" }, { "rank": 164, "name": "Modal", "handle": "@modal_labs", "url": "https://x.com/modal_labs", "followers": "n/a", "category": "AI compute platform", "technical_flag": "Technical", "technical_rationale": "Serverless GPU/AI infra", "why": "Technical AI infra channel.", "source": "https://x.com/modal_labs" }, { "rank": 165, "name": "Anyscale", "handle": "@AnyscaleCompute", "url": "https://x.com/AnyscaleCompute", "followers": "n/a", "category": "Ray / AI infra", "technical_flag": "Technical", "technical_rationale": "Distributed ML/serving infrastructure", "why": "Important scalable AI systems channel.", "source": "https://x.com/AnyscaleCompute" }, { "rank": 166, "name": "Scale AI", "handle": "@scale_AI", "url": "https://x.com/scale_AI", "followers": "n/a", "category": "Data/evals platform", "technical_flag": "Semi-technical", "technical_rationale": "Data, evals, enterprise AI", "why": "Evals/data infrastructure influence.", "source": "https://x.com/scale_AI" }, { "rank": 167, "name": "Scale Labs", "handle": "@ScaleAILabs", "url": "https://x.com/ScaleAILabs", "followers": "n/a", "category": "AI evals / benchmarks", "technical_flag": "Technical", "technical_rationale": "DrugDiscoveryBench and agent benchmarks", "why": "High-signal eval/benchmark channel.", "source": "https://x.com/ScaleAILabs" }, { "rank": 168, "name": "METR", "handle": "@METR_Evals", "url": "https://x.com/METR_Evals", "followers": "~25.2K", "category": "AI capability evals", "technical_flag": "Technical", "technical_rationale": "Agent/R&D capability measurements", "why": "Important agent-eval methodology voice.", "source": "https://x.com/METR_Evals" }, { "rank": 169, "name": "UK AI Security Institute", "handle": "@AISecurityInst", "url": "https://x.com/AISecurityInst", "followers": "~7.6K", "category": "Model safety/evals", "technical_flag": "Semi-technical", "technical_rationale": "Inspect Evals and frontier-model evaluation", "why": "Government eval-framework voice.", "source": "https://x.com/AISecurityInst" }, { "rank": 170, "name": "Center for AI Safety", "handle": "@ai_risks", "url": "https://x.com/ai_risks", "followers": "~8.2K", "category": "AI safety benchmarks", "technical_flag": "Semi-technical", "technical_rationale": "WMDP and risk evaluations", "why": "Bio/chem/cyber eval relevance.", "source": "https://x.com/ai_risks" }, { "rank": 171, "name": "Epoch AI", "handle": "@EpochAIResearch", "url": "https://x.com/EpochAIResearch", "followers": "~45.7K", "category": "AI trends/benchmarks", "technical_flag": "Semi-technical", "technical_rationale": "Benchmark and compute trend analysis", "why": "Methodology/benchmark analysis amplifier.", "source": "https://x.com/EpochAIResearch" }, { "rank": 172, "name": "MLCommons", "handle": "@MLCommons", "url": "https://x.com/MLCommons", "followers": "~3.6K", "category": "AI standards/benchmarks", "technical_flag": "Semi-technical", "technical_rationale": "MLPerf, medical AI, benchmark working groups", "why": "Standards-setting benchmark channel.", "source": "https://x.com/MLCommons" }, { "rank": 173, "name": "Papers with Code", "handle": "@paperswithcode", "url": "https://x.com/paperswithcode", "followers": "~114.7K", "category": "Benchmark leaderboard platform", "technical_flag": "Semi-technical", "technical_rationale": "Paper/code/leaderboard discovery", "why": "Broad benchmark visibility channel.", "source": "https://x.com/paperswithcode" }, { "rank": 174, "name": "Therapeutics Data Commons", "handle": "@ProjectTDC", "url": "https://x.com/ProjectTDC", "followers": "~1.3K", "category": "Drug-discovery ML benchmarks", "technical_flag": "Technical", "technical_rationale": "TDC tasks/leaderboards/datasets", "why": "Foundational drug-discovery benchmark ecosystem.", "source": "https://x.com/ProjectTDC" }, { "rank": 175, "name": "Phylo", "handle": "@phylo_bio", "url": "https://x.com/phylo_bio", "followers": "n/a", "category": "Biomedical agents / DrugDiscoveryBench", "technical_flag": "Technical", "technical_rationale": "Biomni, BiomniBench, DrugDiscoveryBench", "why": "Direct DDDBench peer/partner target.", "source": "https://x.com/phylo_bio" }, { "rank": 176, "name": "Kexin Huang", "handle": "@KexinHuang5", "url": "https://x.com/KexinHuang5", "followers": "n/a", "category": "Biomedical AI agents", "technical_flag": "Technical", "technical_rationale": "BiomniBench/Biomni; long-horizon bio tasks", "why": "High-signal biomedical-agent benchmark KOL.", "source": "https://x.com/KexinHuang5" }, { "rank": 177, "name": "FutureHouse", "handle": "@FutureHouseSF", "url": "https://x.com/FutureHouseSF", "followers": "n/a", "category": "Scientific agents / LAB-Bench", "technical_flag": "Technical", "technical_rationale": "LAB-Bench/BixBench/scientific-agent ecosystem", "why": "Top scientific-agent benchmark target.", "source": "https://x.com/FutureHouseSF" }, { "rank": 178, "name": "Valence Labs", "handle": "@valence_ai", "url": "https://x.com/valence_ai", "followers": "~5.9K", "category": "Drug-discovery benchmarks", "technical_flag": "Technical", "technical_rationale": "Polaris benchmark/data platform", "why": "Relevant industry-standard DD benchmarking channel.", "source": "https://x.com/valence_ai" }, { "rank": 179, "name": "Pat Walters", "handle": "@wpwalters", "url": "https://x.com/wpwalters", "followers": "~6.3K", "category": "Cheminformatics / benchmark critique", "technical_flag": "Technical", "technical_rationale": "ADMET/blind challenges, benchmark validity", "why": "High-credibility DD benchmark evaluator.", "source": "https://x.com/wpwalters" }, { "rank": 180, "name": "OpenBioML", "handle": "@openbioml", "url": "https://x.com/openbioml", "followers": "~4.1K", "category": "Open biology ML", "technical_flag": "Technical", "technical_rationale": "BioML/ChemNLP community amplification", "why": "Community amplifier for open BioML benchmarks.", "source": "https://x.com/openbioml" }, { "rank": 181, "name": "Stephen Turner", "handle": "@strnr", "url": "https://x.com/strnr", "followers": "~31.3K", "category": "Genomics / AIxBio commentary", "technical_flag": "Technical", "technical_rationale": "Genomics, biosecurity, AI-bio eval commentary", "why": "Strong AI-bio benchmark amplifier.", "source": "https://x.com/strnr" }, { "rank": 182, "name": "Jablonka Lab", "handle": "@jablonkagroup", "url": "https://x.com/jablonkagroup", "followers": "~359", "category": "ChemBench / chemistry AI", "technical_flag": "Technical", "technical_rationale": "ChemBench and chemistry/materials LLM evals", "why": "Relevant chemistry benchmark group.", "source": "https://x.com/jablonkagroup" }, { "rank": 183, "name": "Andrew White", "handle": "@andrewwhite01", "url": "https://x.com/andrewwhite01", "followers": "n/a", "category": "ChemCrow / scientific agents", "technical_flag": "Technical", "technical_rationale": "ChemCrow, PaperQA, chemistry agents", "why": "Strong chemistry/science-agent benchmark KOL.", "source": "https://x.com/andrewwhite01" }, { "rank": 184, "name": "Philippe Schwaller", "handle": "@pschwllr", "url": "https://x.com/pschwllr", "followers": "n/a", "category": "Chemical language models", "technical_flag": "Technical", "technical_rationale": "RXN/ChemCrow/chemistry AI", "why": "Chemistry model/eval KOL.", "source": "https://x.com/pschwllr" }, { "rank": 185, "name": "Chai Discovery", "handle": "@chaidiscovery", "url": "https://x.com/chaidiscovery", "followers": "~5.6K", "category": "Structure prediction", "technical_flag": "Technical", "technical_rationale": "Chai-1/chai-lab structure benchmarks", "why": "Protein/structure benchmark channel.", "source": "https://x.com/chaidiscovery" }, { "rank": 186, "name": "Rosetta Commons", "handle": "@RosettaCommons", "url": "https://x.com/RosettaCommons", "followers": "~1.9K", "category": "Protein modeling/design", "technical_flag": "Technical", "technical_rationale": "Rosetta, CASP/CAPRI ecosystem", "why": "Protein-design benchmark credibility network.", "source": "https://x.com/RosettaCommons" }, { "rank": 187, "name": "ChEMBL Database", "handle": "@ChEMBL", "url": "https://x.com/ChEMBL", "followers": "~2.9K", "category": "Drug-discovery data", "technical_flag": "Technical", "technical_rationale": "ChEMBL resource updates", "why": "Benchmark/data community account.", "source": "https://x.com/ChEMBL" }, { "rank": 188, "name": "Tanishq Abraham / MedARC", "handle": "@iScienceLuvr", "url": "https://x.com/iScienceLuvr", "followers": "n/a", "category": "Medical LLM benchmarks", "technical_flag": "Technical", "technical_rationale": "Medmarks and medical AI evaluation", "why": "Medical LLM benchmark KOL.", "source": "https://x.com/iScienceLuvr" }, { "rank": 189, "name": "Stability AI", "handle": "@StabilityAI", "url": "https://x.com/StabilityAI", "followers": "n/a", "category": "Open generative AI lab", "technical_flag": "Semi-technical", "technical_rationale": "Open image/audio/model releases", "why": "Generative AI model distribution channel.", "source": "https://x.com/StabilityAI" }, { "rank": 190, "name": "Midjourney", "handle": "@midjourney", "url": "https://x.com/midjourney", "followers": "n/a", "category": "Generative image models", "technical_flag": "Non-technical / media-business", "technical_rationale": "Model/product releases, creative AI", "why": "Major creative AI model channel.", "source": "https://x.com/midjourney" }, { "rank": 191, "name": "Runway", "handle": "@runwayml", "url": "https://x.com/runwayml", "followers": "n/a", "category": "Generative video AI", "technical_flag": "Semi-technical", "technical_rationale": "Video model/product releases", "why": "Major generative video model channel.", "source": "https://x.com/runwayml" }, { "rank": 192, "name": "Luma AI", "handle": "@LumaLabsAI", "url": "https://x.com/LumaLabsAI", "followers": "n/a", "category": "Generative video/3D", "technical_flag": "Semi-technical", "technical_rationale": "Dream Machine/video/3D model releases", "why": "Creative model product channel.", "source": "https://x.com/LumaLabsAI" }, { "rank": 193, "name": "Cursor", "handle": "@cursor_ai", "url": "https://x.com/cursor_ai", "followers": "n/a", "category": "AI coding IDE", "technical_flag": "Technical", "technical_rationale": "AI coding product and model integrations", "why": "Major AI coding tool channel.", "source": "https://x.com/cursor_ai" }, { "rank": 194, "name": "Cognition", "handle": "@cognition_labs", "url": "https://x.com/cognition_labs", "followers": "n/a", "category": "AI software agents", "technical_flag": "Technical", "technical_rationale": "Devin and coding-agent releases", "why": "Important coding-agent benchmark/product channel.", "source": "https://x.com/cognition_labs" }, { "rank": 195, "name": "Replit", "handle": "@Replit", "url": "https://x.com/Replit", "followers": "n/a", "category": "AI coding platform", "technical_flag": "Semi-technical", "technical_rationale": "AI coding agents and cloud dev platform", "why": "AI coding/developer distribution channel.", "source": "https://x.com/Replit" }, { "rank": 196, "name": "Windsurf", "handle": "@windsurf_ai", "url": "https://x.com/windsurf_ai", "followers": "n/a", "category": "AI coding IDE", "technical_flag": "Technical", "technical_rationale": "Coding-agent/product releases", "why": "Important AI coding product channel.", "source": "https://x.com/windsurf_ai" }, { "rank": 197, "name": "OpenAI Developers", "handle": "@OpenAIDevs", "url": "https://x.com/OpenAIDevs", "followers": "n/a", "category": "OpenAI developer platform", "technical_flag": "Technical", "technical_rationale": "API, model tiers, developer releases", "why": "Primary technical OpenAI developer channel.", "source": "https://x.com/OpenAIDevs" }, { "rank": 198, "name": "Google AI", "handle": "@GoogleAI", "url": "https://x.com/GoogleAI", "followers": "n/a", "category": "Google AI product/research", "technical_flag": "Semi-technical", "technical_rationale": "Google AI releases and research/product links", "why": "Broad Google AI channel.", "source": "https://x.com/GoogleAI" }, { "rank": 199, "name": "Databricks", "handle": "@databricks", "url": "https://x.com/databricks", "followers": "n/a", "category": "Data/AI platform", "technical_flag": "Semi-technical", "technical_rationale": "Lakehouse, MosaicML, model serving and enterprise AI infra", "why": "Important enterprise data+AI platform channel.", "source": "https://x.com/databricks" }, { "rank": 200, "name": "MosaicML", "handle": "@MosaicML", "url": "https://x.com/MosaicML", "followers": "n/a", "category": "Model training infrastructure", "technical_flag": "Technical", "technical_rationale": "Open model training, LLM training systems and infra", "why": "Technical model-training infrastructure voice.", "source": "https://x.com/MosaicML" } ], "model_performance_kols": [ { "rank": 1, "name": "Artificial Analysis", "handle": "@ArtificialAnlys", "url": "https://x.com/ArtificialAnlys", "followers": "~25K", "category": "Independent benchmark org", "performance_focus": "frontier benchmarks; cost/speed/quality; coding agents", "technical_flag": "Technical", "why": "Independent model quality, intelligence, speed, price, cost-per-task, and coding-agent comparisons.", "inclusion_rationale": "Independent model quality, intelligence, speed, price, cost-per-task, and coding-agent comparisons.", "source": "https://x.com/ArtificialAnlys" }, { "rank": 2, "name": "LM Arena", "handle": "@lmarena_ai", "url": "https://x.com/lmarena_ai", "followers": "~80K", "category": "Model arena / leaderboard", "performance_focus": "frontier benchmarks; human preference arena", "technical_flag": "Technical", "why": "Publishes Chatbot Arena / model leaderboard results and arena methodology.", "inclusion_rationale": "Publishes Chatbot Arena / model leaderboard results and arena methodology.", "source": "https://x.com/lmarena_ai" }, { "rank": 3, "name": "LMSYS Org", "handle": "@lmsysorg", "url": "https://x.com/lmsysorg", "followers": "~60K", "category": "Research lab / Chatbot Arena", "performance_focus": "frontier benchmarks; arena methodology", "technical_flag": "Technical", "why": "Originator of Chatbot Arena/FastChat-style model evaluation work.", "inclusion_rationale": "Originator of Chatbot Arena/FastChat-style model evaluation work.", "source": "https://x.com/lmsysorg" }, { "rank": 4, "name": "LiveBench", "handle": "@livebench_ai", "url": "https://x.com/livebench_ai", "followers": "~10K", "category": "Live benchmark", "performance_focus": "contamination-resistant LLM benchmark", "technical_flag": "Technical", "why": "Maintains live/refreshing benchmark results to reduce contamination and stale evals.", "inclusion_rationale": "Maintains live/refreshing benchmark results to reduce contamination and stale evals.", "source": "https://x.com/livebench_ai" }, { "rank": 5, "name": "Aider", "handle": "@aider_chat", "url": "https://x.com/aider_chat", "followers": "~30K", "category": "Coding benchmark/tool", "performance_focus": "Aider Polyglot; coding model performance", "technical_flag": "Technical", "why": "Publishes practical coding model benchmark results from Aider workflows.", "inclusion_rationale": "Publishes practical coding model benchmark results from Aider workflows.", "source": "https://x.com/aider_chat" }, { "rank": 6, "name": "Paul Gauthier", "handle": "@paulgauthier", "url": "https://x.com/paulgauthier", "followers": "~20K", "category": "Aider creator", "performance_focus": "coding model performance", "technical_flag": "Technical", "why": "Creator of Aider; frequent publisher/interpreter of coding model leaderboard results.", "inclusion_rationale": "Creator of Aider; frequent publisher/interpreter of coding model leaderboard results.", "source": "https://x.com/paulgauthier" }, { "rank": 7, "name": "SWE-bench", "handle": "@SWEbench", "url": "https://x.com/SWEbench", "followers": "~20K", "category": "Software engineering benchmark", "performance_focus": "coding agents; SWE-bench Verified/Multilingual", "technical_flag": "Technical", "why": "Official software-engineering benchmark and leaderboard for AI coding agents.", "inclusion_rationale": "Official software-engineering benchmark and leaderboard for AI coding agents.", "source": "https://x.com/SWEbench" }, { "rank": 8, "name": "OpenAI Developers", "handle": "@OpenAIDevs", "url": "https://x.com/OpenAIDevs", "followers": "n/a", "category": "Developer platform", "performance_focus": "API/model performance and coding/tool-use updates", "technical_flag": "Technical", "why": "Developer-facing model release and performance details.", "inclusion_rationale": "Developer-facing model release and performance details.", "source": "https://x.com/OpenAIDevs" }, { "rank": 9, "name": "Vals AI", "handle": "@ValsAI", "url": "https://x.com/ValsAI", "followers": "n/a", "category": "Model evals / private benchmarks", "performance_focus": "frontier model benchmarks across domains", "technical_flag": "Technical", "why": "Publishes model rankings across legal, financial, coding, multimodal, and other private benchmarks.", "inclusion_rationale": "Publishes model rankings across legal, financial, coding, multimodal, and other private benchmarks.", "source": "https://x.com/ValsAI" }, { "rank": 10, "name": "Andrej Karpathy", "handle": "@karpathy", "url": "https://x.com/karpathy", "followers": "~3.2M", "category": "AI researcher / educator", "technical_flag": "Technical", "technical_rationale": "Posts deep model intuition, training, code, and technical education", "why": "One of the highest signal technical AI explainers on X.", "source": "https://x.com/karpathy", "performance_focus": "AI researcher / educator", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 11, "name": "Andrew Ng", "handle": "@AndrewYNg", "url": "https://x.com/AndrewYNg", "followers": "~1.5–1.6M", "category": "AI education / applied ML", "technical_flag": "Technical", "technical_rationale": "Applied ML, data-centric AI, courses, technical practice", "why": "Major applied-AI educator and practitioner influence.", "source": "https://x.com/AndrewYNg", "performance_focus": "AI education / applied ML", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 12, "name": "Demis Hassabis", "handle": "@demishassabis", "url": "https://x.com/demishassabis", "followers": "~1.3M", "category": "Google DeepMind CEO", "technical_flag": "Technical", "technical_rationale": "Science-AI, DeepMind research, AlphaFold/Gemini technical context", "why": "High-authority AI-for-science and AGI lab voice.", "source": "https://x.com/demishassabis", "performance_focus": "Google DeepMind CEO", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 13, "name": "Yann LeCun", "handle": "@ylecun", "url": "https://x.com/ylecun", "followers": "~0.94–1.2M", "category": "Meta AI / Turing laureate", "technical_flag": "Technical", "technical_rationale": "Research debates on world models, self-supervised learning, architectures", "why": "Major technical counterweight to LLM-scaling consensus.", "source": "https://x.com/ylecun", "performance_focus": "Meta AI / Turing laureate", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 14, "name": "Ilya Sutskever", "handle": "@ilyasut", "url": "https://x.com/ilyasut", "followers": "~760.9K", "category": "SSI / frontier research", "technical_flag": "Technical", "technical_rationale": "Foundational LLM researcher; rare but technical significance", "why": "Rare posts carry frontier-lab/research weight.", "source": "https://x.com/ilyasut", "performance_focus": "SSI / frontier research", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 15, "name": "Aravind Srinivas", "handle": "@AravSrinivas", "url": "https://x.com/AravSrinivas", "followers": "~800.6K", "category": "Perplexity CEO", "technical_flag": "Semi-technical", "technical_rationale": "AI search/product strategy with some model context", "why": "AI search and consumer-AI distribution leader.", "source": "https://x.com/AravSrinivas", "performance_focus": "Perplexity CEO", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 16, "name": "Fei-Fei Li", "handle": "@drfeifei", "url": "https://x.com/drfeifei", "followers": "~632–781K", "category": "World Labs / Stanford", "technical_flag": "Technical", "technical_rationale": "Computer vision, spatial intelligence, AI healthcare/policy", "why": "Major AI research and human-centered AI authority.", "source": "https://x.com/drfeifei", "performance_focus": "World Labs / Stanford", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 17, "name": "François Chollet", "handle": "@fchollet", "url": "https://x.com/fchollet", "followers": "~566–704K", "category": "Keras / ARC-AGI", "technical_flag": "Technical", "technical_rationale": "Reasoning benchmarks, Keras, AGI critique, architecture debates", "why": "Important benchmark/reasoning voice.", "source": "https://x.com/fchollet", "performance_focus": "Keras / ARC-AGI", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 18, "name": "Geoffrey Hinton", "handle": "@geoffreyhinton", "url": "https://x.com/geoffreyhinton", "followers": "~489–618K", "category": "Deep learning pioneer", "technical_flag": "Technical", "technical_rationale": "Foundational deep-learning researcher", "why": "High-authority voice on deep learning and AI risk.", "source": "https://x.com/geoffreyhinton", "performance_focus": "Deep learning pioneer", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 19, "name": "Dario Amodei", "handle": "@DarioAmodei", "url": "https://x.com/DarioAmodei", "followers": "~478.6K", "category": "Anthropic CEO", "technical_flag": "Semi-technical", "technical_rationale": "Frontier lab strategy/safety, not frequent code-level posting", "why": "Claude/frontier AI strategy and safety framing.", "source": "https://x.com/DarioAmodei", "performance_focus": "Anthropic CEO", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 20, "name": "Jeff Dean", "handle": "@JeffDean", "url": "https://x.com/JeffDean", "followers": "~361–448K", "category": "Google DeepMind / Google Research", "technical_flag": "Technical", "technical_rationale": "Systems ML, scaling, Gemini/Google research", "why": "Elite systems + ML research credibility.", "source": "https://x.com/JeffDean", "performance_focus": "Google DeepMind / Google Research", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 21, "name": "Clément Delangue", "handle": "@ClementDelangue", "url": "https://x.com/ClementDelangue", "followers": "~452.1K", "category": "Hugging Face CEO", "technical_flag": "Semi-technical", "technical_rationale": "Open-source model ecosystem/product strategy", "why": "Open-source AI distribution and community leader.", "source": "https://x.com/ClementDelangue", "performance_focus": "Hugging Face CEO", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 22, "name": "Alexandr Wang", "handle": "@alexandr_wang", "url": "https://x.com/alexandr_wang", "followers": "~541.7K", "category": "Meta AI / Scale founder", "technical_flag": "Semi-technical", "technical_rationale": "AI data/evals/enterprise strategy", "why": "Data, evals, and enterprise AI influence.", "source": "https://x.com/alexandr_wang", "performance_focus": "Meta AI / Scale founder", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 23, "name": "Jim Fan", "handle": "@DrJimFan", "url": "https://x.com/DrJimFan", "followers": "~435–486K", "category": "NVIDIA robotics / agents", "technical_flag": "Technical", "technical_rationale": "Robotics, simulation, embodied agents, model systems", "why": "High-signal physical AGI and robotics commentator.", "source": "https://x.com/DrJimFan", "performance_focus": "NVIDIA robotics / agents", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 24, "name": "Sebastian Raschka", "handle": "@rasbt", "url": "https://x.com/rasbt", "followers": "~476K", "category": "ML/LLM educator", "technical_flag": "Technical", "technical_rationale": "Implementation-level LLM/ML tutorials", "why": "Hands-on LLM training and architecture education.", "source": "https://x.com/rasbt", "performance_focus": "ML/LLM educator", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 25, "name": "Soumith Chintala", "handle": "@soumithchintala", "url": "https://x.com/soumithchintala", "followers": "~247–311K", "category": "PyTorch / Meta", "technical_flag": "Technical", "technical_rationale": "Open-source ML framework engineering", "why": "PyTorch ecosystem credibility.", "source": "https://x.com/soumithchintala", "performance_focus": "PyTorch / Meta", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 26, "name": "Lilian Weng", "handle": "@lilianweng", "url": "https://x.com/lilianweng", "followers": "~88–267K", "category": "AI research / agents / safety", "technical_flag": "Technical", "technical_rationale": "Canonical technical essays on agents, RLHF, safety", "why": "Widely cited technical explainer.", "source": "https://x.com/lilianweng", "performance_focus": "AI research / agents / safety", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 27, "name": "Jeremy Howard", "handle": "@jeremyphoward", "url": "https://x.com/jeremyphoward", "followers": "~236–289K", "category": "fast.ai / Answer.AI", "technical_flag": "Technical", "technical_rationale": "Practical model training, pedagogy, open-source AI", "why": "Applied technical AI education and open models.", "source": "https://x.com/jeremyphoward", "performance_focus": "fast.ai / Answer.AI", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 28, "name": "Emad Mostaque", "handle": "@EMostaque", "url": "https://x.com/EMostaque", "followers": "~328.3K", "category": "Stability AI founder", "technical_flag": "Semi-technical", "technical_rationale": "Open model strategy and ecosystem commentary", "why": "Open-source AI and generative media commentary.", "source": "https://x.com/EMostaque", "performance_focus": "Stability AI founder", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 29, "name": "Logan Kilpatrick", "handle": "@OfficialLoganK", "url": "https://x.com/OfficialLoganK", "followers": "~339.2K", "category": "Google AI Studio / devrel", "technical_flag": "Semi-technical", "technical_rationale": "API/model developer-relations updates", "why": "Developer-facing AI launch signal.", "source": "https://x.com/OfficialLoganK", "performance_focus": "Google AI Studio / devrel", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 30, "name": "Simon Willison", "handle": "@simonw", "url": "https://x.com/simonw", "followers": "~154.5K", "category": "Developer / AI tooling analyst", "technical_flag": "Technical", "technical_rationale": "Hands-on testing, code, model behavior, security", "why": "Top practical technical evaluator of LLM tools.", "source": "https://x.com/simonw", "performance_focus": "Developer / AI tooling analyst", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 31, "name": "Hamel Husain", "handle": "@HamelHusain", "url": "https://x.com/HamelHusain", "followers": "n/a", "category": "AI evals / applied LLM systems", "technical_flag": "Technical", "technical_rationale": "Evals, product quality, production LLM systems", "why": "High-signal practitioner voice on AI evals.", "source": "https://x.com/HamelHusain", "performance_focus": "AI evals / applied LLM systems", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 32, "name": "Aidan McLaughlin", "handle": "@aidan_mclau", "url": "https://x.com/aidan_mclau", "followers": "~47.8K", "category": "OpenAI research", "technical_flag": "Technical", "technical_rationale": "Frontier model design/research perspective", "why": "Useful model-design and frontier-lab research signal.", "source": "https://x.com/aidan_mclau", "performance_focus": "OpenAI research", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 33, "name": "Ian Goodfellow", "handle": "@goodfellow_ian", "url": "https://x.com/goodfellow_ian", "followers": "~315.7K", "category": "Deep learning researcher", "technical_flag": "Technical", "technical_rationale": "GANs and deep learning research", "why": "Foundational ML researcher.", "source": "https://x.com/goodfellow_ian", "performance_focus": "Deep learning researcher", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 34, "name": "Oriol Vinyals", "handle": "@oriolvinyalsml", "url": "https://x.com/oriolvinyalsml", "followers": "~182.4K", "category": "Google DeepMind research", "technical_flag": "Technical", "technical_rationale": "Gemini, AlphaCode, sequence models", "why": "Major DeepMind technical leader.", "source": "https://x.com/oriolvinyalsml", "performance_focus": "Google DeepMind research", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 35, "name": "Nando de Freitas", "handle": "@NandoDF", "url": "https://x.com/NandoDF", "followers": "~108.9K", "category": "Microsoft AI / ex-DeepMind", "technical_flag": "Technical", "technical_rationale": "Deep learning and superintelligence research", "why": "High-credibility research voice.", "source": "https://x.com/NandoDF", "performance_focus": "Microsoft AI / ex-DeepMind", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 36, "name": "Abhishek Thakur", "handle": "@abhi1thakur", "url": "https://x.com/abhi1thakur", "followers": "~83.1K", "category": "Kaggle / Hugging Face AutoTrain", "technical_flag": "Technical", "technical_rationale": "Kaggle, AutoML, model-building practice", "why": "Applied ML competition and tooling expert.", "source": "https://x.com/abhi1thakur", "performance_focus": "Kaggle / Hugging Face AutoTrain", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 37, "name": "Gary Marcus", "handle": "@GaryMarcus", "url": "https://x.com/GaryMarcus", "followers": "~213.5K", "category": "AI critic / cognitive scientist", "technical_flag": "Semi-technical", "technical_rationale": "Critiques of deep learning and LLM limits", "why": "Influential skeptical AI commentator.", "source": "https://x.com/GaryMarcus", "performance_focus": "AI critic / cognitive scientist", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 38, "name": "Data Chaz", "handle": "@DataChaz", "url": "https://x.com/DataChaz", "followers": "~114.1K", "category": "Developer advocate / Streamlit", "technical_flag": "Semi-technical", "technical_rationale": "Python/data apps/LLM developer content", "why": "Practical AI/data app educator.", "source": "https://x.com/DataChaz", "performance_focus": "Developer advocate / Streamlit", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 39, "name": "Jürgen Schmidhuber", "handle": "@SchmidhuberAI", "url": "https://x.com/SchmidhuberAI", "followers": "~98.1K", "category": "AI researcher", "technical_flag": "Technical", "technical_rationale": "Deep learning history and research claims", "why": "Longstanding ML research figure.", "source": "https://x.com/SchmidhuberAI", "performance_focus": "AI researcher", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 40, "name": "Pedro Domingos", "handle": "@pmddomingos", "url": "https://x.com/pmddomingos", "followers": "~74.3K", "category": "ML professor / author", "technical_flag": "Technical", "technical_rationale": "ML theory and AI commentary", "why": "Recognized ML academic voice.", "source": "https://x.com/pmddomingos", "performance_focus": "ML professor / author", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 41, "name": "Santiago Valdarrama", "handle": "@svpino", "url": "https://x.com/svpino", "followers": "~383.3K", "category": "ML educator", "technical_flag": "Technical", "technical_rationale": "Hands-on ML engineering and system design", "why": "Large practical ML learning audience.", "source": "https://x.com/svpino", "performance_focus": "ML educator", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 42, "name": "Bojan Tunguz", "handle": "@tunguz", "url": "https://x.com/tunguz", "followers": "~251.1K", "category": "ML engineer / Kaggle", "technical_flag": "Technical", "technical_rationale": "Data science/Kaggle/model-building", "why": "Applied ML/Kaggle KOL.", "source": "https://x.com/tunguz", "performance_focus": "ML engineer / Kaggle", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 43, "name": "Akshay Pachaar", "handle": "@akshay_pachaar", "url": "https://x.com/akshay_pachaar", "followers": "~240K", "category": "LLM/agents educator", "technical_flag": "Semi-technical", "technical_rationale": "Practical LLM/RAG/agent explainers", "why": "Large applied AI education audience.", "source": "https://x.com/akshay_pachaar", "performance_focus": "LLM/agents educator", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 44, "name": "Hugo Larochelle", "handle": "@hugo_larochelle", "url": "https://x.com/hugo_larochelle", "followers": "~115.9K", "category": "Google DeepMind researcher", "technical_flag": "Technical", "technical_rationale": "Deep learning research and education", "why": "High-quality ML research voice.", "source": "https://x.com/hugo_larochelle", "performance_focus": "Google DeepMind researcher", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 45, "name": "Peyman Milanfar", "handle": "@docmilanfar", "url": "https://x.com/docmilanfar", "followers": "~100.1K", "category": "Google computational imaging", "technical_flag": "Technical", "technical_rationale": "Vision/imaging ML research", "why": "Technical computer-vision KOL.", "source": "https://x.com/docmilanfar", "performance_focus": "Google computational imaging", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 46, "name": "Dan Kornas", "handle": "@dankornas", "url": "https://x.com/dankornas", "followers": "~87.1K", "category": "ML engineer / educator", "technical_flag": "Technical", "technical_rationale": "End-to-end ML engineering education", "why": "Practical ML engineering audience.", "source": "https://x.com/dankornas", "performance_focus": "ML engineer / educator", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 47, "name": "Kevin Patrick Murphy", "handle": "@sirbayes", "url": "https://x.com/sirbayes", "followers": "~64.7K", "category": "Google DeepMind researcher", "technical_flag": "Technical", "technical_rationale": "Bayesian ML, probabilistic modeling", "why": "Authoritative ML textbook/research voice.", "source": "https://x.com/sirbayes", "performance_focus": "Google DeepMind researcher", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 48, "name": "Gautam Kamath", "handle": "@thegautamkamath", "url": "https://x.com/thegautamkamath", "followers": "~59.7K", "category": "ML professor / Vector Institute", "technical_flag": "Technical", "technical_rationale": "Privacy, statistics, ML theory", "why": "Technical ML theory voice.", "source": "https://x.com/thegautamkamath", "performance_focus": "ML professor / Vector Institute", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 49, "name": "Sara Hooker", "handle": "@sarahookr", "url": "https://x.com/sarahookr", "followers": "~53.3K", "category": "Cohere Labs / model efficiency", "technical_flag": "Technical", "technical_rationale": "ML efficiency, multilingual/multimodal research", "why": "Important efficient/open AI researcher.", "source": "https://x.com/sarahookr", "performance_focus": "Cohere Labs / model efficiency", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 50, "name": "Durk Kingma", "handle": "@dpkingma", "url": "https://x.com/dpkingma", "followers": "~51.3K", "category": "Anthropic / VAE/Adam inventor", "technical_flag": "Technical", "technical_rationale": "Core ML optimization and generative modeling", "why": "Foundational ML researcher.", "source": "https://x.com/dpkingma", "performance_focus": "Anthropic / VAE/Adam inventor", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 51, "name": "Andriy Burkov", "handle": "@burkov", "url": "https://x.com/burkov", "followers": "~50.8K", "category": "ML author", "technical_flag": "Technical", "technical_rationale": "ML/LLM books and practical education", "why": "Strong ML educational voice.", "source": "https://x.com/burkov", "performance_focus": "ML author", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 52, "name": "Jean de Nyandwi", "handle": "@Jeande_d", "url": "https://x.com/Jeande_d", "followers": "~46.9K", "category": "CMU / NLP evals", "technical_flag": "Technical", "technical_rationale": "Multimodal NLP, post-training, data, evals", "why": "Rising technical AI research voice.", "source": "https://x.com/Jeande_d", "performance_focus": "CMU / NLP evals", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 53, "name": "Shakir Mohamed", "handle": "@shakir_za", "url": "https://x.com/shakir_za", "followers": "~41.7K", "category": "DeepMind researcher", "technical_flag": "Technical", "technical_rationale": "ML research and AI for social good", "why": "Technical DeepMind/social-good voice.", "source": "https://x.com/shakir_za", "performance_focus": "DeepMind researcher", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 54, "name": "Ferenc Huszár", "handle": "@fhuszar", "url": "https://x.com/fhuszar", "followers": "~37.9K", "category": "ML professor", "technical_flag": "Technical", "technical_rationale": "Bayesian/statistical ML commentary", "why": "Technical ML/statistics KOL.", "source": "https://x.com/fhuszar", "performance_focus": "ML professor", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 55, "name": "Andrew Gordon Wilson", "handle": "@andrewgwils", "url": "https://x.com/andrewgwils", "followers": "~35.5K", "category": "ML professor", "technical_flag": "Technical", "technical_rationale": "Bayesian/deep learning theory", "why": "Technical ML academic.", "source": "https://x.com/andrewgwils", "performance_focus": "ML professor", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 56, "name": "Ben Recht", "handle": "@beenwrekt", "url": "https://x.com/beenwrekt", "followers": "~32.9K", "category": "Berkeley ML/control", "technical_flag": "Technical", "technical_rationale": "Optimization, ML systems, critical analysis", "why": "Sharp technical ML critic.", "source": "https://x.com/beenwrekt", "performance_focus": "Berkeley ML/control", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 57, "name": "Zoubin Ghahramani", "handle": "@ZoubinGhahrama1", "url": "https://x.com/ZoubinGhahrama1", "followers": "~32.6K", "category": "Google DeepMind / Cambridge", "technical_flag": "Technical", "technical_rationale": "Probabilistic ML and AI research leadership", "why": "Senior ML research authority.", "source": "https://x.com/ZoubinGhahrama1", "performance_focus": "Google DeepMind / Cambridge", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 58, "name": "Jakob Foerster", "handle": "@j_foerst", "url": "https://x.com/j_foerst", "followers": "~22.7K", "category": "Oxford/Meta ML researcher", "technical_flag": "Technical", "technical_rationale": "Multi-agent RL and ML research", "why": "Technical RL/agents researcher.", "source": "https://x.com/j_foerst", "performance_focus": "Oxford/Meta ML researcher", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 59, "name": "Nils Reimers", "handle": "@Nils_Reimers", "url": "https://x.com/Nils_Reimers", "followers": "~14.6K", "category": "Cohere / SBERT", "technical_flag": "Technical", "technical_rationale": "Embeddings, retrieval, AI search", "why": "Important technical retrieval/embedding voice.", "source": "https://x.com/Nils_Reimers", "performance_focus": "Cohere / SBERT", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 60, "name": "Nathan Raw", "handle": "@_nateraw", "url": "https://x.com/_nateraw", "followers": "~9.5K", "category": "ML hacker / Hugging Face alumnus", "technical_flag": "Technical", "technical_rationale": "Open-source ML tooling", "why": "Hands-on ML builder.", "source": "https://x.com/_nateraw", "performance_focus": "ML hacker / Hugging Face alumnus", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 61, "name": "Alicia Curth", "handle": "@AliciaCurth", "url": "https://x.com/AliciaCurth", "followers": "~4.7K", "category": "ML/statistics researcher", "technical_flag": "Technical", "technical_rationale": "Statistical ML intuition and methods", "why": "Technical ML/statistics voice.", "source": "https://x.com/AliciaCurth", "performance_focus": "ML/statistics researcher", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 62, "name": "Yoshua Bengio", "handle": "@yoshuabengio", "url": "https://x.com/yoshuabengio", "followers": "n/a", "category": "Deep learning pioneer", "technical_flag": "Technical", "technical_rationale": "Foundational deep-learning research", "why": "Turing laureate and major AI safety/research voice.", "source": "https://x.com/yoshuabengio", "performance_focus": "Deep learning pioneer", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 63, "name": "Chris Olah", "handle": "@chrisolah", "url": "https://x.com/chrisolah", "followers": "n/a", "category": "Anthropic / interpretability", "technical_flag": "Technical", "technical_rationale": "Mechanistic interpretability research", "why": "Core interpretability KOL.", "source": "https://x.com/chrisolah", "performance_focus": "Anthropic / interpretability", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 64, "name": "Anca Dragan", "handle": "@ancadragan", "url": "https://x.com/ancadragan", "followers": "n/a", "category": "Berkeley robotics / HRI", "technical_flag": "Technical", "technical_rationale": "Human-robot interaction and robotics", "why": "Technical robotics/HRI authority.", "source": "https://x.com/ancadragan", "performance_focus": "Berkeley robotics / HRI", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 65, "name": "Pieter Abbeel", "handle": "@pabbeel", "url": "https://x.com/pabbeel", "followers": "n/a", "category": "Berkeley robotics / Covariant", "technical_flag": "Technical", "technical_rationale": "Robotics and reinforcement learning", "why": "Major robotics/RL KOL.", "source": "https://x.com/pabbeel", "performance_focus": "Berkeley robotics / Covariant", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 66, "name": "Chelsea Finn", "handle": "@chelseabfinn", "url": "https://x.com/chelseabfinn", "followers": "n/a", "category": "Stanford robotics / meta-learning", "technical_flag": "Technical", "technical_rationale": "Robotics, meta-learning, imitation learning", "why": "High-signal robotics/learning researcher.", "source": "https://x.com/chelseabfinn", "performance_focus": "Stanford robotics / meta-learning", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 67, "name": "Sergey Levine", "handle": "@svlevine", "url": "https://x.com/svlevine", "followers": "n/a", "category": "Berkeley robotics / RL", "technical_flag": "Technical", "technical_rationale": "Deep RL and robot learning", "why": "Top technical robotics/RL researcher.", "source": "https://x.com/svlevine", "performance_focus": "Berkeley robotics / RL", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 68, "name": "Been Kim", "handle": "@beenkim", "url": "https://x.com/beenkim", "followers": "n/a", "category": "Google interpretability", "technical_flag": "Technical", "technical_rationale": "Explainability/interpretability research", "why": "Important technical interpretability voice.", "source": "https://x.com/beenkim", "performance_focus": "Google interpretability", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 69, "name": "Samy Bengio", "handle": "@samybengio", "url": "https://x.com/samybengio", "followers": "n/a", "category": "Apple / ex-Google Brain", "technical_flag": "Technical", "technical_rationale": "Deep learning research leadership", "why": "Senior deep learning researcher.", "source": "https://x.com/samybengio", "performance_focus": "Apple / ex-Google Brain", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 70, "name": "Kyunghyun Cho", "handle": "@kchon2020", "url": "https://x.com/kchon2020", "followers": "n/a", "category": "NYU / NLP", "technical_flag": "Technical", "technical_rationale": "Neural machine translation, NLP, deep learning", "why": "Technical NLP researcher.", "source": "https://x.com/kchon2020", "performance_focus": "NYU / NLP", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 71, "name": "Max Welling", "handle": "@wellingmax", "url": "https://x.com/wellingmax", "followers": "n/a", "category": "CuspAI / ML professor", "technical_flag": "Technical", "technical_rationale": "ML theory, generative science, Bayesian/deep learning", "why": "Technical ML/science-AI voice.", "source": "https://x.com/wellingmax", "performance_focus": "CuspAI / ML professor", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 72, "name": "David Ha", "handle": "@hardmaru", "url": "https://x.com/hardmaru", "followers": "n/a", "category": "AI researcher / generative systems", "technical_flag": "Technical", "technical_rationale": "World models, generative systems, creative AI", "why": "Technical and creative AI research voice.", "source": "https://x.com/hardmaru", "performance_focus": "AI researcher / generative systems", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 73, "name": "Alex Graves", "handle": "@alexjgraves", "url": "https://x.com/alexjgraves", "followers": "n/a", "category": "DeepMind researcher", "technical_flag": "Technical", "technical_rationale": "RNNs, sequence learning, deep learning", "why": "Foundational sequence-model researcher.", "source": "https://x.com/alexjgraves", "performance_focus": "DeepMind researcher", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 74, "name": "Quoc Le", "handle": "@quocleix", "url": "https://x.com/quocleix", "followers": "n/a", "category": "Google Brain / AutoML", "technical_flag": "Technical", "technical_rationale": "AutoML and deep learning research", "why": "Senior technical Google AI researcher.", "source": "https://x.com/quocleix", "performance_focus": "Google Brain / AutoML", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 75, "name": "Leslie Kaelbling", "handle": "@lpkaelbling", "url": "https://x.com/lpkaelbling", "followers": "n/a", "category": "MIT AI/robotics", "technical_flag": "Technical", "technical_rationale": "Planning, robotics, AI fundamentals", "why": "Senior AI/robotics authority.", "source": "https://x.com/lpkaelbling", "performance_focus": "MIT AI/robotics", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 76, "name": "Aidan Gomez", "handle": "@aidangomez", "url": "https://x.com/aidangomez", "followers": "n/a", "category": "Cohere CEO / Transformer coauthor", "technical_flag": "Technical", "technical_rationale": "Transformer coauthor and enterprise LLM builder", "why": "Technical founder with model lineage.", "source": "https://x.com/aidangomez", "performance_focus": "Cohere CEO / Transformer coauthor", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 77, "name": "Rachel Thomas", "handle": "@math_rachel", "url": "https://x.com/math_rachel", "followers": "n/a", "category": "fast.ai / AI education", "technical_flag": "Technical", "technical_rationale": "Practical ML and responsible AI education", "why": "Technical educator and fast.ai cofounder.", "source": "https://x.com/math_rachel", "performance_focus": "fast.ai / AI education", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 78, "name": "Thomas Wolf", "handle": "@Thom_Wolf", "url": "https://x.com/Thom_Wolf", "followers": "n/a", "category": "Hugging Face cofounder", "technical_flag": "Technical", "technical_rationale": "Transformers/open-source AI engineering", "why": "Core open-source AI platform builder.", "source": "https://x.com/Thom_Wolf", "performance_focus": "Hugging Face cofounder", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 79, "name": "Julien Chaumond", "handle": "@julien_c", "url": "https://x.com/julien_c", "followers": "n/a", "category": "Hugging Face CTO/cofounder", "technical_flag": "Technical", "technical_rationale": "Model hub/open-source platform engineering", "why": "Important open-source AI infra voice.", "source": "https://x.com/julien_c", "performance_focus": "Hugging Face CTO/cofounder", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 80, "name": "Sebastian Thrun", "handle": "@SebastianThrun", "url": "https://x.com/SebastianThrun", "followers": "n/a", "category": "Udacity / robotics", "technical_flag": "Technical", "technical_rationale": "Robotics, autonomous vehicles, education", "why": "Robotics/AI education pioneer.", "source": "https://x.com/SebastianThrun", "performance_focus": "Udacity / robotics", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 81, "name": "Abeba Birhane", "handle": "@Abebab", "url": "https://x.com/Abebab", "followers": "n/a", "category": "AI accountability researcher", "technical_flag": "Semi-technical", "technical_rationale": "AI bias/eval/accountability research", "why": "Strong responsible-AI researcher.", "source": "https://x.com/Abebab", "performance_focus": "AI accountability researcher", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 82, "name": "Rediet Abebe", "handle": "@red_abebe", "url": "https://x.com/red_abebe", "followers": "n/a", "category": "AI for social good", "technical_flag": "Technical", "technical_rationale": "Algorithms and social-good AI research", "why": "Technical/public-interest AI voice.", "source": "https://x.com/red_abebe", "performance_focus": "AI for social good", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 83, "name": "Melanie Mitchell", "handle": "@MelMitchell1", "url": "https://x.com/MelMitchell1", "followers": "n/a", "category": "Complexity / AI researcher", "technical_flag": "Technical", "technical_rationale": "AI reasoning, analogy, cognitive science", "why": "Technical and skeptical AI researcher.", "source": "https://x.com/MelMitchell1", "performance_focus": "Complexity / AI researcher", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 84, "name": "Aurélien Géron", "handle": "@aureliengeron", "url": "https://x.com/aureliengeron", "followers": "n/a", "category": "ML author", "technical_flag": "Technical", "technical_rationale": "Hands-on ML education", "why": "Popular practical ML educator.", "source": "https://x.com/aureliengeron", "performance_focus": "ML author", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 85, "name": "Chip Huyen", "handle": "@chipro", "url": "https://x.com/chipro", "followers": "n/a", "category": "AI engineering author", "technical_flag": "Technical", "technical_rationale": "ML systems, AI engineering, evals", "why": "High-signal AI engineering voice.", "source": "https://x.com/chipro", "performance_focus": "AI engineering author", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 86, "name": "Monica Rogati", "handle": "@mrogati", "url": "https://x.com/mrogati", "followers": "n/a", "category": "Data science leader", "technical_flag": "Technical", "technical_rationale": "Data/AI product strategy and systems", "why": "Technical data/AI product KOL.", "source": "https://x.com/mrogati", "performance_focus": "Data science leader", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 87, "name": "Inioluwa Deborah Raji", "handle": "@rajiinio", "url": "https://x.com/rajiinio", "followers": "n/a", "category": "AI accountability researcher", "technical_flag": "Semi-technical", "technical_rationale": "Model auditing and accountability", "why": "Important AI auditing voice.", "source": "https://x.com/rajiinio", "performance_focus": "AI accountability researcher", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 88, "name": "Arvind Narayanan", "handle": "@random_walker", "url": "https://x.com/random_walker", "followers": "n/a", "category": "Princeton AI/society", "technical_flag": "Technical", "technical_rationale": "AI, fairness, privacy, crypto, accountability", "why": "Technical AI and policy critic.", "source": "https://x.com/random_walker", "performance_focus": "Princeton AI/society", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 89, "name": "Vincent Warmerdam", "handle": "@fishnets88", "url": "https://x.com/fishnets88", "followers": "n/a", "category": "Calmcode / ML tooling", "technical_flag": "Technical", "technical_rationale": "Practical ML education and tooling", "why": "High-signal practical technical educator.", "source": "https://x.com/fishnets88", "performance_focus": "Calmcode / ML tooling", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 90, "name": "Sasha Luccioni", "handle": "@SashaMTL", "url": "https://x.com/SashaMTL", "followers": "n/a", "category": "Hugging Face / AI climate", "technical_flag": "Semi-technical", "technical_rationale": "AI energy, climate, responsible AI evaluation", "why": "Technical responsible-AI voice.", "source": "https://x.com/SashaMTL", "performance_focus": "Hugging Face / AI climate", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 91, "name": "Jan Leike", "handle": "@janleike", "url": "https://x.com/janleike", "followers": "n/a", "category": "Anthropic / alignment", "technical_flag": "Technical", "technical_rationale": "Alignment, scalable oversight, RLHF", "why": "Top technical alignment researcher.", "source": "https://x.com/janleike", "performance_focus": "Anthropic / alignment", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 92, "name": "Paul Christiano", "handle": "@paulc443", "url": "https://x.com/paulc443", "followers": "n/a", "category": "Alignment Research Center", "technical_flag": "Technical", "technical_rationale": "Technical AI alignment theory/evals", "why": "Core technical alignment voice.", "source": "https://x.com/paulc443", "performance_focus": "Alignment Research Center", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 93, "name": "Percy Liang", "handle": "@percyliang", "url": "https://x.com/percyliang", "followers": "~106.5K", "category": "Stanford NLP / HELM", "technical_flag": "Technical", "technical_rationale": "Benchmarking, NLP, HELM/MedHELM", "why": "Top LLM-evaluation researcher.", "source": "https://x.com/percyliang", "performance_focus": "Stanford NLP / HELM", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 94, "name": "Jacob Andreas", "handle": "@jacobandreas", "url": "https://x.com/jacobandreas", "followers": "n/a", "category": "MIT NLP", "technical_flag": "Technical", "technical_rationale": "Language grounding, agents, NLP research", "why": "Technical NLP/agents researcher.", "source": "https://x.com/jacobandreas", "performance_focus": "MIT NLP", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 95, "name": "Dan Hendrycks", "handle": "@DanHendrycks", "url": "https://x.com/DanHendrycks", "followers": "n/a", "category": "Center for AI Safety", "technical_flag": "Technical", "technical_rationale": "Benchmarks, model risks, WMDP/MMLU-style evals", "why": "Important eval/safety benchmark creator.", "source": "https://x.com/DanHendrycks", "performance_focus": "Center for AI Safety", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 96, "name": "Scott Aaronson", "handle": "@scottaaronson", "url": "https://x.com/scottaaronson", "followers": "n/a", "category": "Theoretical CS / OpenAI alumnus", "technical_flag": "Technical", "technical_rationale": "Complexity theory, quantum, AI safety commentary", "why": "Technical theory/safety voice.", "source": "https://x.com/scottaaronson", "performance_focus": "Theoretical CS / OpenAI alumnus", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 97, "name": "Swyx", "handle": "@swyx", "url": "https://x.com/swyx", "followers": "n/a", "category": "AI engineering / Latent Space", "technical_flag": "Technical", "technical_rationale": "Agents, RAG, evals, AI engineering stack", "why": "Top AI engineering community builder.", "source": "https://x.com/swyx", "performance_focus": "AI engineering / Latent Space", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 98, "name": "Deedy Das", "handle": "@deedydas", "url": "https://x.com/deedydas", "followers": "n/a", "category": "AI investor / curator", "technical_flag": "Semi-technical", "technical_rationale": "Model release summaries and benchmark takes", "why": "High-signal AI startup/model curator.", "source": "https://x.com/deedydas", "performance_focus": "AI investor / curator", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 99, "name": "Linus Lee", "handle": "@thesephist", "url": "https://x.com/thesephist", "followers": "n/a", "category": "Notion AI researcher", "technical_flag": "Technical", "technical_rationale": "LLM internals, embeddings, AI-native UX", "why": "Technical AI product/research voice.", "source": "https://x.com/thesephist", "performance_focus": "Notion AI researcher", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 100, "name": "Nathan Labenz", "handle": "@labenz", "url": "https://x.com/labenz", "followers": "n/a", "category": "Cognitive Revolution podcast", "technical_flag": "Semi-technical", "technical_rationale": "Model eval discussions, AI interviews", "why": "Strong AI interview/eval commentator.", "source": "https://x.com/labenz", "performance_focus": "Cognitive Revolution podcast", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 101, "name": "Nathan Lambert", "handle": "@natolambert", "url": "https://x.com/natolambert", "followers": "n/a", "category": "AI2 / Interconnects", "technical_flag": "Technical", "technical_rationale": "RLHF, post-training, open models", "why": "Top technical post-training/open-model voice.", "source": "https://x.com/natolambert", "performance_focus": "AI2 / Interconnects", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 102, "name": "arXiv", "handle": "@arxiv_org", "url": "https://x.com/arxiv_org", "followers": "n/a", "category": "Research preprint platform", "technical_flag": "Technical", "technical_rationale": "Primary research feed", "why": "Raw research discovery channel.", "source": "https://x.com/arxiv_org", "performance_focus": "Research preprint platform", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 103, "name": "AlphaXiv", "handle": "@alphaxiv", "url": "https://x.com/alphaxiv", "followers": "n/a", "category": "Paper discovery platform", "technical_flag": "Technical", "technical_rationale": "Social layer for arXiv/papers", "why": "AI paper discovery/amplification.", "source": "https://x.com/alphaxiv", "performance_focus": "Paper discovery platform", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 104, "name": "Cameron Wolfe", "handle": "@cwolferesearch", "url": "https://x.com/cwolferesearch", "followers": "n/a", "category": "AI researcher / writer", "technical_flag": "Technical", "technical_rationale": "Paper breakdowns and Deep Learning Focus", "why": "High-signal technical AI explainer.", "source": "https://x.com/cwolferesearch", "performance_focus": "AI researcher / writer", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 105, "name": "Vikhyat K.", "handle": "@vikhyatk", "url": "https://x.com/vikhyatk", "followers": "n/a", "category": "Moondream / small VLMs", "technical_flag": "Technical", "technical_rationale": "Small-model/VLM engineering", "why": "Practical model builder.", "source": "https://x.com/vikhyatk", "performance_focus": "Moondream / small VLMs", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 106, "name": "OpenAI", "handle": "@OpenAI", "url": "https://x.com/OpenAI", "followers": "n/a", "category": "Frontier AI lab", "technical_flag": "Semi-technical", "technical_rationale": "Official launches and benchmark posts", "why": "Primary OpenAI launch channel.", "source": "https://x.com/OpenAI", "performance_focus": "Frontier AI lab", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 107, "name": "Anthropic", "handle": "@AnthropicAI", "url": "https://x.com/AnthropicAI", "followers": "n/a", "category": "Frontier AI lab", "technical_flag": "Semi-technical", "technical_rationale": "Claude releases and safety research", "why": "Primary Anthropic launch channel.", "source": "https://x.com/AnthropicAI", "performance_focus": "Frontier AI lab", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 108, "name": "Google DeepMind", "handle": "@GoogleDeepMind", "url": "https://x.com/GoogleDeepMind", "followers": "~1.2M", "category": "Frontier AI/science lab", "technical_flag": "Technical", "technical_rationale": "Research releases, AlphaFold/Gemini science AI", "why": "Major technical AI lab channel.", "source": "https://x.com/GoogleDeepMind", "performance_focus": "Frontier AI/science lab", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 109, "name": "Meta AI", "handle": "@MetaAI", "url": "https://x.com/MetaAI", "followers": "n/a", "category": "Open models / AI lab", "technical_flag": "Technical", "technical_rationale": "Llama/open model research and product", "why": "Key open-model lab channel.", "source": "https://x.com/MetaAI", "performance_focus": "Open models / AI lab", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 110, "name": "Mistral AI", "handle": "@MistralAI", "url": "https://x.com/MistralAI", "followers": "n/a", "category": "Frontier/open model lab", "technical_flag": "Technical", "technical_rationale": "Open/foundation model releases", "why": "European frontier model channel.", "source": "https://x.com/MistralAI", "performance_focus": "Frontier/open model lab", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 111, "name": "Hugging Face", "handle": "@huggingface", "url": "https://x.com/huggingface", "followers": "~689K", "category": "Open-source AI platform", "technical_flag": "Semi-technical", "technical_rationale": "Model/dataset/leaderboard ecosystem", "why": "High-impact open AI distribution.", "source": "https://x.com/huggingface", "performance_focus": "Open-source AI platform", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 112, "name": "Cohere", "handle": "@cohere", "url": "https://x.com/cohere", "followers": "n/a", "category": "Enterprise LLM lab", "technical_flag": "Semi-technical", "technical_rationale": "Enterprise model releases and retrieval/RAG context", "why": "Enterprise LLM platform channel.", "source": "https://x.com/cohere", "performance_focus": "Enterprise LLM lab", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 113, "name": "Together AI", "handle": "@togethercompute", "url": "https://x.com/togethercompute", "followers": "n/a", "category": "Open model infra", "technical_flag": "Technical", "technical_rationale": "Inference/fine-tuning/open model infra", "why": "Open AI infrastructure channel.", "source": "https://x.com/togethercompute", "performance_focus": "Open model infra", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 114, "name": "xAI", "handle": "@xai", "url": "https://x.com/xai", "followers": "n/a", "category": "Frontier AI lab", "technical_flag": "Semi-technical", "technical_rationale": "Grok releases and xAI research", "why": "xAI primary launch channel.", "source": "https://x.com/xai", "performance_focus": "Frontier AI lab", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 115, "name": "OpenRouter", "handle": "@OpenRouterAI", "url": "https://x.com/OpenRouterAI", "followers": "n/a", "category": "Model routing/API platform", "technical_flag": "Semi-technical", "technical_rationale": "Model availability and routing data", "why": "Useful model market/API signal.", "source": "https://x.com/OpenRouterAI", "performance_focus": "Model routing/API platform", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 116, "name": "Replicate", "handle": "@replicate", "url": "https://x.com/replicate", "followers": "n/a", "category": "Model hosting/API", "technical_flag": "Technical", "technical_rationale": "Open model deployment and demos", "why": "Model deployment ecosystem channel.", "source": "https://x.com/replicate", "performance_focus": "Model hosting/API", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 117, "name": "Weights & Biases", "handle": "@weights_biases", "url": "https://x.com/weights_biases", "followers": "n/a", "category": "MLOps / evals", "technical_flag": "Semi-technical", "technical_rationale": "Experiment tracking, evals, ML tooling", "why": "MLOps and AI eval community.", "source": "https://x.com/weights_biases", "performance_focus": "MLOps / evals", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 118, "name": "LangChain", "handle": "@LangChainAI", "url": "https://x.com/LangChainAI", "followers": "n/a", "category": "LLM app framework", "technical_flag": "Technical", "technical_rationale": "Agents, RAG, LangGraph, app tooling", "why": "Key AI agent/RAG developer channel.", "source": "https://x.com/LangChainAI", "performance_focus": "LLM app framework", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 119, "name": "LlamaIndex", "handle": "@llama_index", "url": "https://x.com/llama_index", "followers": "n/a", "category": "LLM data/RAG framework", "technical_flag": "Technical", "technical_rationale": "RAG, retrieval, agents, data connectors", "why": "Important AI engineering framework.", "source": "https://x.com/llama_index", "performance_focus": "LLM data/RAG framework", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 120, "name": "Modal", "handle": "@modal_labs", "url": "https://x.com/modal_labs", "followers": "n/a", "category": "AI compute platform", "technical_flag": "Technical", "technical_rationale": "Serverless GPU/AI infra", "why": "Technical AI infra channel.", "source": "https://x.com/modal_labs", "performance_focus": "AI compute platform", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 121, "name": "Anyscale", "handle": "@AnyscaleCompute", "url": "https://x.com/AnyscaleCompute", "followers": "n/a", "category": "Ray / AI infra", "technical_flag": "Technical", "technical_rationale": "Distributed ML/serving infrastructure", "why": "Important scalable AI systems channel.", "source": "https://x.com/AnyscaleCompute", "performance_focus": "Ray / AI infra", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 122, "name": "Scale AI", "handle": "@scale_AI", "url": "https://x.com/scale_AI", "followers": "n/a", "category": "Data/evals platform", "technical_flag": "Semi-technical", "technical_rationale": "Data, evals, enterprise AI", "why": "Evals/data infrastructure influence.", "source": "https://x.com/scale_AI", "performance_focus": "Data/evals platform", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 123, "name": "Scale Labs", "handle": "@ScaleAILabs", "url": "https://x.com/ScaleAILabs", "followers": "n/a", "category": "AI evals / benchmarks", "technical_flag": "Technical", "technical_rationale": "DrugDiscoveryBench and agent benchmarks", "why": "High-signal eval/benchmark channel.", "source": "https://x.com/ScaleAILabs", "performance_focus": "AI evals / benchmarks", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 124, "name": "METR", "handle": "@METR_Evals", "url": "https://x.com/METR_Evals", "followers": "~25.2K", "category": "AI capability evals", "technical_flag": "Technical", "technical_rationale": "Agent/R&D capability measurements", "why": "Important agent-eval methodology voice.", "source": "https://x.com/METR_Evals", "performance_focus": "AI capability evals", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 125, "name": "UK AI Security Institute", "handle": "@AISecurityInst", "url": "https://x.com/AISecurityInst", "followers": "~7.6K", "category": "Model safety/evals", "technical_flag": "Semi-technical", "technical_rationale": "Inspect Evals and frontier-model evaluation", "why": "Government eval-framework voice.", "source": "https://x.com/AISecurityInst", "performance_focus": "Model safety/evals", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 126, "name": "Center for AI Safety", "handle": "@ai_risks", "url": "https://x.com/ai_risks", "followers": "~8.2K", "category": "AI safety benchmarks", "technical_flag": "Semi-technical", "technical_rationale": "WMDP and risk evaluations", "why": "Bio/chem/cyber eval relevance.", "source": "https://x.com/ai_risks", "performance_focus": "AI safety benchmarks", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 127, "name": "Epoch AI", "handle": "@EpochAIResearch", "url": "https://x.com/EpochAIResearch", "followers": "~45.7K", "category": "AI trends/benchmarks", "technical_flag": "Semi-technical", "technical_rationale": "Benchmark and compute trend analysis", "why": "Methodology/benchmark analysis amplifier.", "source": "https://x.com/EpochAIResearch", "performance_focus": "AI trends/benchmarks", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 128, "name": "MLCommons", "handle": "@MLCommons", "url": "https://x.com/MLCommons", "followers": "~3.6K", "category": "AI standards/benchmarks", "technical_flag": "Semi-technical", "technical_rationale": "MLPerf, medical AI, benchmark working groups", "why": "Standards-setting benchmark channel.", "source": "https://x.com/MLCommons", "performance_focus": "AI standards/benchmarks", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 129, "name": "Papers with Code", "handle": "@paperswithcode", "url": "https://x.com/paperswithcode", "followers": "~114.7K", "category": "Benchmark leaderboard platform", "technical_flag": "Semi-technical", "technical_rationale": "Paper/code/leaderboard discovery", "why": "Broad benchmark visibility channel.", "source": "https://x.com/paperswithcode", "performance_focus": "Benchmark leaderboard platform", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 130, "name": "Therapeutics Data Commons", "handle": "@ProjectTDC", "url": "https://x.com/ProjectTDC", "followers": "~1.3K", "category": "Drug-discovery ML benchmarks", "technical_flag": "Technical", "technical_rationale": "TDC tasks/leaderboards/datasets", "why": "Foundational drug-discovery benchmark ecosystem.", "source": "https://x.com/ProjectTDC", "performance_focus": "Drug-discovery ML benchmarks", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 131, "name": "Phylo", "handle": "@phylo_bio", "url": "https://x.com/phylo_bio", "followers": "n/a", "category": "Biomedical agents / DrugDiscoveryBench", "technical_flag": "Technical", "technical_rationale": "Biomni, BiomniBench, DrugDiscoveryBench", "why": "Direct DDDBench peer/partner target.", "source": "https://x.com/phylo_bio", "performance_focus": "Biomedical agents / DrugDiscoveryBench", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 132, "name": "Kexin Huang", "handle": "@KexinHuang5", "url": "https://x.com/KexinHuang5", "followers": "n/a", "category": "Biomedical AI agents", "technical_flag": "Technical", "technical_rationale": "BiomniBench/Biomni; long-horizon bio tasks", "why": "High-signal biomedical-agent benchmark KOL.", "source": "https://x.com/KexinHuang5", "performance_focus": "Biomedical AI agents", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 133, "name": "FutureHouse", "handle": "@FutureHouseSF", "url": "https://x.com/FutureHouseSF", "followers": "n/a", "category": "Scientific agents / LAB-Bench", "technical_flag": "Technical", "technical_rationale": "LAB-Bench/BixBench/scientific-agent ecosystem", "why": "Top scientific-agent benchmark target.", "source": "https://x.com/FutureHouseSF", "performance_focus": "Scientific agents / LAB-Bench", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 134, "name": "Valence Labs", "handle": "@valence_ai", "url": "https://x.com/valence_ai", "followers": "~5.9K", "category": "Drug-discovery benchmarks", "technical_flag": "Technical", "technical_rationale": "Polaris benchmark/data platform", "why": "Relevant industry-standard DD benchmarking channel.", "source": "https://x.com/valence_ai", "performance_focus": "Drug-discovery benchmarks", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 135, "name": "Pat Walters", "handle": "@wpwalters", "url": "https://x.com/wpwalters", "followers": "~6.3K", "category": "Cheminformatics / benchmark critique", "technical_flag": "Technical", "technical_rationale": "ADMET/blind challenges, benchmark validity", "why": "High-credibility DD benchmark evaluator.", "source": "https://x.com/wpwalters", "performance_focus": "Cheminformatics / benchmark critique", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 136, "name": "OpenBioML", "handle": "@openbioml", "url": "https://x.com/openbioml", "followers": "~4.1K", "category": "Open biology ML", "technical_flag": "Technical", "technical_rationale": "BioML/ChemNLP community amplification", "why": "Community amplifier for open BioML benchmarks.", "source": "https://x.com/openbioml", "performance_focus": "Open biology ML", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 137, "name": "Stephen Turner", "handle": "@strnr", "url": "https://x.com/strnr", "followers": "~31.3K", "category": "Genomics / AIxBio commentary", "technical_flag": "Technical", "technical_rationale": "Genomics, biosecurity, AI-bio eval commentary", "why": "Strong AI-bio benchmark amplifier.", "source": "https://x.com/strnr", "performance_focus": "Genomics / AIxBio commentary", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 138, "name": "Jablonka Lab", "handle": "@jablonkagroup", "url": "https://x.com/jablonkagroup", "followers": "~359", "category": "ChemBench / chemistry AI", "technical_flag": "Technical", "technical_rationale": "ChemBench and chemistry/materials LLM evals", "why": "Relevant chemistry benchmark group.", "source": "https://x.com/jablonkagroup", "performance_focus": "ChemBench / chemistry AI", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 139, "name": "Andrew White", "handle": "@andrewwhite01", "url": "https://x.com/andrewwhite01", "followers": "n/a", "category": "ChemCrow / scientific agents", "technical_flag": "Technical", "technical_rationale": "ChemCrow, PaperQA, chemistry agents", "why": "Strong chemistry/science-agent benchmark KOL.", "source": "https://x.com/andrewwhite01", "performance_focus": "ChemCrow / scientific agents", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 140, "name": "Philippe Schwaller", "handle": "@pschwllr", "url": "https://x.com/pschwllr", "followers": "n/a", "category": "Chemical language models", "technical_flag": "Technical", "technical_rationale": "RXN/ChemCrow/chemistry AI", "why": "Chemistry model/eval KOL.", "source": "https://x.com/pschwllr", "performance_focus": "Chemical language models", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 141, "name": "Chai Discovery", "handle": "@chaidiscovery", "url": "https://x.com/chaidiscovery", "followers": "~5.6K", "category": "Structure prediction", "technical_flag": "Technical", "technical_rationale": "Chai-1/chai-lab structure benchmarks", "why": "Protein/structure benchmark channel.", "source": "https://x.com/chaidiscovery", "performance_focus": "Structure prediction", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 142, "name": "Rosetta Commons", "handle": "@RosettaCommons", "url": "https://x.com/RosettaCommons", "followers": "~1.9K", "category": "Protein modeling/design", "technical_flag": "Technical", "technical_rationale": "Rosetta, CASP/CAPRI ecosystem", "why": "Protein-design benchmark credibility network.", "source": "https://x.com/RosettaCommons", "performance_focus": "Protein modeling/design", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 143, "name": "ChEMBL Database", "handle": "@ChEMBL", "url": "https://x.com/ChEMBL", "followers": "~2.9K", "category": "Drug-discovery data", "technical_flag": "Technical", "technical_rationale": "ChEMBL resource updates", "why": "Benchmark/data community account.", "source": "https://x.com/ChEMBL", "performance_focus": "Drug-discovery data", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 144, "name": "Tanishq Abraham / MedARC", "handle": "@iScienceLuvr", "url": "https://x.com/iScienceLuvr", "followers": "n/a", "category": "Medical LLM benchmarks", "technical_flag": "Technical", "technical_rationale": "Medmarks and medical AI evaluation", "why": "Medical LLM benchmark KOL.", "source": "https://x.com/iScienceLuvr", "performance_focus": "Medical LLM benchmarks", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 145, "name": "Stability AI", "handle": "@StabilityAI", "url": "https://x.com/StabilityAI", "followers": "n/a", "category": "Open generative AI lab", "technical_flag": "Semi-technical", "technical_rationale": "Open image/audio/model releases", "why": "Generative AI model distribution channel.", "source": "https://x.com/StabilityAI", "performance_focus": "Open generative AI lab", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 146, "name": "Midjourney", "handle": "@midjourney", "url": "https://x.com/midjourney", "followers": "n/a", "category": "Generative image models", "technical_flag": "Non-technical / media-business", "technical_rationale": "Model/product releases, creative AI", "why": "Major creative AI model channel.", "source": "https://x.com/midjourney", "performance_focus": "Generative image models", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 147, "name": "Runway", "handle": "@runwayml", "url": "https://x.com/runwayml", "followers": "n/a", "category": "Generative video AI", "technical_flag": "Semi-technical", "technical_rationale": "Video model/product releases", "why": "Major generative video model channel.", "source": "https://x.com/runwayml", "performance_focus": "Generative video AI", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 148, "name": "Luma AI", "handle": "@LumaLabsAI", "url": "https://x.com/LumaLabsAI", "followers": "n/a", "category": "Generative video/3D", "technical_flag": "Semi-technical", "technical_rationale": "Dream Machine/video/3D model releases", "why": "Creative model product channel.", "source": "https://x.com/LumaLabsAI", "performance_focus": "Generative video/3D", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 149, "name": "Cursor", "handle": "@cursor_ai", "url": "https://x.com/cursor_ai", "followers": "n/a", "category": "AI coding IDE", "technical_flag": "Technical", "technical_rationale": "AI coding product and model integrations", "why": "Major AI coding tool channel.", "source": "https://x.com/cursor_ai", "performance_focus": "AI coding IDE", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." }, { "rank": 150, "name": "Cognition", "handle": "@cognition_labs", "url": "https://x.com/cognition_labs", "followers": "n/a", "category": "AI software agents", "technical_flag": "Technical", "technical_rationale": "Devin and coding-agent releases", "why": "Important coding-agent benchmark/product channel.", "source": "https://x.com/cognition_labs", "performance_focus": "AI software agents", "inclusion_rationale": "Focused on model capability, benchmarks/evals, open-model comparisons, AI engineering tests, coding/agent performance, or model infrastructure — not general celebrity influence." } ], "top_benchmark_posts_month": [ { "rank": 1, "author": "Artificial Analysis", "handle": "@ArtificialAnlys", "url": "https://x.com/ArtificialAnlys/status/2075268970492657905", "date": "1 day ago", "topic": "GPT-5.6 benchmark launch analysis", "models": "GPT-5.6 Sol/Terra/Luna; Claude Fable 5", "benchmarks": "Coding Agent Index; DeepSWE; Terminal-Bench v2; SWE-Atlas-QnA", "why": "Sol leads coding-agent evals and approaches Fable intelligence at lower cost", "engagement": "n/a", "source_note": "direct X result", "confidence": "High" }, { "rank": 2, "author": "OpenAI", "handle": "@OpenAI", "url": "https://x.com/OpenAI/status/2075271425548795909", "date": "1 day ago", "topic": "GPT-5.6 public release benchmark claims", "models": "GPT-5.6 family", "benchmarks": "Agents Last Exam; Terminal-Bench; HealthBench; AA Index", "why": "Primary model-release post for latest OpenAI benchmark family", "engagement": "n/a", "source_note": "subagent/web search", "confidence": "Medium" }, { "rank": 3, "author": "OpenAI Developers", "handle": "@OpenAIDevs", "url": "https://x.com/OpenAIDevs/status/2075274009395241347", "date": "1 day ago", "topic": "Developer/API performance for GPT-5.6", "models": "GPT-5.6 family", "benchmarks": "coding; tool use; context/pricing", "why": "Builder-facing model-performance details", "engagement": "n/a", "source_note": "subagent/web search", "confidence": "Medium" }, { "rank": 4, "author": "Artificial Analysis", "handle": "@ArtificialAnlys", "url": "https://x.com/ArtificialAnlys/status/2074956932289282087", "date": "2 days ago", "topic": "Grok 4.5 independent benchmark", "models": "Grok 4.5; GPT-5.5; Fable 5; Opus 4.8", "benchmarks": "AA Intelligence Index; Coding Agent Index", "why": "Places Grok 4.5 high overall and efficient for coding agents", "engagement": "n/a", "source_note": "subagent/web search", "confidence": "High" }, { "rank": 5, "author": "OpenAI", "handle": "@OpenAI", "url": "https://x.com/OpenAI/status/2070555276370169969", "date": "2 weeks ago", "topic": "GPT-5.6 Sol Terminal-Bench SOTA", "models": "GPT-5.6 Sol", "benchmarks": "Terminal-Bench 2.1", "why": "OpenAI says Sol sets SOTA on complex command-line workflows", "engagement": "n/a", "source_note": "web_search exact X result", "confidence": "High" }, { "rank": 6, "author": "Artificial Analysis", "handle": "@ArtificialAnlys", "url": "https://x.com/ArtificialAnlys/status/2074299714699469221", "date": "4 days ago", "topic": "Capability indices by domain", "models": "Fable 5; Opus 4.8; GPT-5.5; Gemini 3.5 Flash; GLM-5.2", "benchmarks": "AA Capability Indices", "why": "Domain-specific model ranking shifts and open-weight leaders", "engagement": "n/a", "source_note": "subagent/web search", "confidence": "High" }, { "rank": 7, "author": "Artificial Analysis", "handle": "@ArtificialAnlys", "url": "https://x.com/ArtificialAnlys/status/2074299918358041079", "date": "4 days ago", "topic": "Cost-per-task frontier comparison", "models": "DeepSeek V4 Flash; GLM-5.2; frontier models", "benchmarks": "AA Strategy & Ops Index; cost/task", "why": "Highlights cheap open-weight task completion vs frontier premium", "engagement": "n/a", "source_note": "subagent/web search", "confidence": "High" }, { "rank": 8, "author": "Scale AI Labs", "handle": "@ScaleAILabs", "url": "https://x.com/ScaleAILabs/status/2072011679064584343", "date": "1 week ago", "topic": "DrugDiscoveryBench launch", "models": "GPT-5.5; Gemini 3.5 Flash; Opus 4.8", "benchmarks": "DrugDiscoveryBench", "why": "Independent benchmark for early drug-discovery computational work", "engagement": "n/a", "source_note": "subagent/web search", "confidence": "High" }, { "rank": 9, "author": "Wes Roth", "handle": "@WesRoth", "url": "https://x.com/WesRoth/status/2072349511532646768", "date": "1 week ago", "topic": "DrugDiscoveryBench summary", "models": "leading AI agents", "benchmarks": "DrugDiscoveryBench", "why": "Broad amplification of life-science model benchmark", "engagement": "n/a", "source_note": "subagent/web search", "confidence": "High" }, { "rank": 10, "author": "OpenAI", "handle": "@OpenAI", "url": "https://x.com/OpenAI/status/2067346918746030496", "date": "3 weeks ago", "topic": "LifeSciBench release", "models": "GPT-Rosalind; GPT-5.5", "benchmarks": "LifeSciBench", "why": "Life-science benchmark: GPT-Rosalind improves reported performance", "engagement": "n/a", "source_note": "subagent/web search", "confidence": "High" }, { "rank": 11, "author": "Artificial Analysis", "handle": "@ArtificialAnlys", "url": "https://x.com/ArtificialAnlys/status/2064500150069030992", "date": "1 month ago", "topic": "Claude Fable 5 tops AA Intelligence Index", "models": "Claude Fable 5; Opus 4.8", "benchmarks": "AA Intelligence Index", "why": "Fable 5 score/rank post from independent benchmark source", "engagement": "n/a", "source_note": "subagent", "confidence": "High" }, { "rank": 12, "author": "Artificial Analysis", "handle": "@ArtificialAnlys", "url": "https://x.com/ArtificialAnlys/status/2064500154959618084", "date": "1 month ago", "topic": "Fable 5 knowledge/hallucination eval", "models": "Claude Fable 5; Opus 4.8", "benchmarks": "AA-Omniscience", "why": "Knowledge reliability / hallucination-rate comparison", "engagement": "n/a", "source_note": "subagent", "confidence": "High" }, { "rank": 13, "author": "Artificial Analysis", "handle": "@ArtificialAnlys", "url": "https://x.com/ArtificialAnlys/status/2067384319942029379", "date": "3 weeks ago", "topic": "Fable 5 benchmark cost", "models": "Claude Fable 5; Opus 4.8; GPT-5.5", "benchmarks": "AA Intelligence Index; cost/task", "why": "Shows benchmark-run cost and premium frontier pricing", "engagement": "n/a", "source_note": "subagent", "confidence": "High" }, { "rank": 14, "author": "Artificial Analysis", "handle": "@ArtificialAnlys", "url": "https://x.com/ArtificialAnlys/status/2065328920514515037", "date": "1 month ago", "topic": "DeepSWE replaces SWE-Bench Pro", "models": "GPT-5.5 Codex; Claude Code Fable 5; Opus 4.8", "benchmarks": "DeepSWE; SWE-Bench Pro", "why": "Coding-agent benchmark contamination/gameability discussion", "engagement": "n/a", "source_note": "subagent", "confidence": "High" }, { "rank": 15, "author": "Artificial Analysis", "handle": "@ArtificialAnlys", "url": "https://x.com/ArtificialAnlys/status/2065328924578693514", "date": "1 month ago", "topic": "SWE-Bench Pro anomaly", "models": "GPT-5.5 Codex; Opus 4.8", "benchmarks": "SWE-Bench Pro; DeepSWE", "why": "Explains anomalous behavior across coding benchmarks", "engagement": "n/a", "source_note": "subagent", "confidence": "High" }, { "rank": 16, "author": "Artificial Analysis", "handle": "@ArtificialAnlys", "url": "https://x.com/ArtificialAnlys/status/2065559824230957190", "date": "1 month ago", "topic": "Agentic inference hardware benchmark", "models": "DeepSeek V4 Pro", "benchmarks": "AA-AgentPerf", "why": "Agents per MW across Blackwell/Hopper/AMD", "engagement": "n/a", "source_note": "subagent", "confidence": "High" }, { "rank": 17, "author": "Artificial Analysis", "handle": "@ArtificialAnlys", "url": "https://x.com/ArtificialAnlys/status/2066700136018071841", "date": "4 weeks ago", "topic": "AA Intelligence Index v4.1", "models": "GPT-5.5; Opus 4.8; DeepSeek V4 Pro", "benchmarks": "AA Intelligence Index v4.1; GDPval-AA v2", "why": "Shift toward agentic workloads/per-task metrics", "engagement": "n/a", "source_note": "subagent", "confidence": "High" }, { "rank": 18, "author": "Artificial Analysis", "handle": "@ArtificialAnlys", "url": "https://x.com/ArtificialAnlys/status/2066700142116544998", "date": "4 weeks ago", "topic": "GDPval-AA v2 upgrade", "models": "Claude Fable 5; GPT-5.5", "benchmarks": "GDPval-AA v2", "why": "Rebaselines Elo to human performance; longer agent trajectories", "engagement": "n/a", "source_note": "subagent", "confidence": "High" }, { "rank": 19, "author": "Artificial Analysis", "handle": "@ArtificialAnlys", "url": "https://x.com/ArtificialAnlys/status/2067135640249209175", "date": "3 weeks ago", "topic": "GLM-5.2 open-weight frontier", "models": "GLM-5.2; MiniMax-M3; DeepSeek V4 Pro; Kimi K2.6", "benchmarks": "AA Intelligence Index; CritPt; HLE", "why": "GLM-5.2 becomes leading open-weight model in AA index", "engagement": "n/a", "source_note": "web_search exact X result", "confidence": "High" }, { "rank": 20, "author": "Artificial Analysis", "handle": "@ArtificialAnlys", "url": "https://x.com/ArtificialAnlys/status/2067135645093695602", "date": "3 weeks ago", "topic": "GLM-5.2 GDPval strength", "models": "GLM-5.2; MiniMax-M3; DeepSeek V4 Pro; GPT-5.5", "benchmarks": "GDPval-AA v2", "why": "GLM-5.2 leads open weights on agentic knowledge work", "engagement": "n/a", "source_note": "web_search exact X result", "confidence": "High" }, { "rank": 21, "author": "Artificial Analysis", "handle": "@ArtificialAnlys", "url": "https://x.com/ArtificialAnlys/status/2067329643905253730", "date": "3 weeks ago", "topic": "GLM-5.2 physics reasoning", "models": "GLM-5.2; GPT-5.5 Pro; Opus 4.8; DeepSeek V4 Pro", "benchmarks": "CritPt", "why": "Open-weight model ties Opus 4.8 on unpublished physics problems", "engagement": "n/a", "source_note": "subagent", "confidence": "High" }, { "rank": 22, "author": "Artificial Analysis", "handle": "@ArtificialAnlys", "url": "https://x.com/ArtificialAnlys/status/2069121548670406947", "date": "3 weeks ago", "topic": "GLM-5.2 real-world agent benchmark", "models": "GLM-5.2; Fable 5; Opus 4.8; GPT-5.5; MiniMax-M3", "benchmarks": "GDPval-AA", "why": "GLM-5.2 #3 overall and top open-weight on work tasks", "engagement": "n/a", "source_note": "web_search exact X result", "confidence": "High" }, { "rank": 23, "author": "Artificial Analysis", "handle": "@ArtificialAnlys", "url": "https://x.com/ArtificialAnlys/status/2069914443639635978", "date": "2 weeks ago", "topic": "AA-Briefcase task duration", "models": "Fable 5; GLM-5.2; frontier models", "benchmarks": "AA-Briefcase", "why": "Shows agentic knowledge-work task duration and cost/time implications", "engagement": "n/a", "source_note": "subagent", "confidence": "High" }, { "rank": 24, "author": "Artificial Analysis", "handle": "@ArtificialAnlys", "url": "https://x.com/ArtificialAnlys/status/2072022576394821859", "date": "1 week ago", "topic": "GLM-5.2 verbosity/cost", "models": "GLM-5.2; GPT-5.5; Opus 4.8; Gemini 3.5 Flash", "benchmarks": "AA Intelligence Index; AA-Omniscience; GDPval-AA", "why": "Shows high output-token use and benchmark efficiency tradeoffs", "engagement": "n/a", "source_note": "web_search exact X result", "confidence": "High" }, { "rank": 25, "author": "Artificial Analysis", "handle": "@ArtificialAnlys", "url": "https://x.com/ArtificialAnlys/status/2072062592923930666", "date": "1 week ago", "topic": "Claude Sonnet 5 AA score", "models": "Claude Sonnet 5; GPT-5.5; Opus 4.8", "benchmarks": "AA Intelligence Index; AA-Briefcase; GDPval-AA", "why": "Sonnet 5 reaches strong score but with cost/token caveats", "engagement": "n/a", "source_note": "subagent", "confidence": "High" }, { "rank": 26, "author": "Chubby / Kimmonismus", "handle": "@kimmonismus", "url": "https://x.com/kimmonismus/status/2069149390665830535", "date": "3 weeks ago", "topic": "GLM-5.2 GDPval amplification", "models": "GLM-5.2; GPT-5.5", "benchmarks": "GDPval-AA", "why": "High-reach amplification of open-weight model beating GPT-5.5 on GDPval-AA", "engagement": "n/a", "source_note": "web_search exact X result", "confidence": "Medium" }, { "rank": 27, "author": "0xLogicrw", "handle": "@0xLogicrw", "url": "https://x.com/0xLogicrw/status/2067176813596741810", "date": "3 weeks ago", "topic": "GLM-5.2 Chinese benchmark summary", "models": "GLM-5.2; MiniMax-M3; DeepSeek V4 Pro; Kimi K2.6", "benchmarks": "AA Intelligence Index; GDPval-AA", "why": "Chinese-language amplification of GLM-5.2 benchmark results", "engagement": "n/a", "source_note": "web_search exact X result", "confidence": "Medium" }, { "rank": 28, "author": "AiBattle", "handle": "@AiBattle_", "url": "https://x.com/AiBattle_/status/2072017782955257930", "date": "1 week ago", "topic": "Claude Sonnet 5 benchmark comparison", "models": "Claude Sonnet 5; Sonnet 4.6; GPT-5.5; Gemini 3.5 Flash", "benchmarks": "benchmark comparison", "why": "Model-vs-model benchmark comparison post", "engagement": "n/a", "source_note": "web_search exact X result", "confidence": "Medium" }, { "rank": 29, "author": "David Hendrickson", "handle": "@TeksEdge", "url": "https://x.com/TeksEdge/status/2064504089829056565", "date": "1 month ago", "topic": "AA benchmark commentary on Fable 5", "models": "Claude Fable 5; GPT-5.5; Gemini 3.1 Pro", "benchmarks": "AA overall intelligence benchmarks", "why": "Commentary translating AA score gaps into model-buying context", "engagement": "n/a", "source_note": "web_search exact X result", "confidence": "Medium" }, { "rank": 30, "author": "Kexin Huang", "handle": "@KexinHuang5", "url": "https://x.com/KexinHuang5/status/2056397835327656004", "date": "within month-ish", "topic": "BiomniBench launch", "models": "AI agents", "benchmarks": "BiomniBench", "why": "Process-level biomedical agent benchmark post", "engagement": "n/a", "source_note": "prior search result", "confidence": "High" }, { "rank": 31, "author": "ArtificialAnlys", "handle": "@ArtificialAnlys", "url": "https://x.com/search?q=Artificial%20Analysis%20GPT-5.6%20Coding%20Agent%20Index&src=typed_query&f=live", "date": "past month search result", "topic": "Artificial Analysis GPT-5.6 Coding Agent Index", "models": "GPT-5.6; Claude Fable 5", "benchmarks": "Coding Agent Index; DeepSWE; Terminal-Bench", "why": "Follow-up/related AA posts on GPT-5.6 coding-agent evals", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 32, "author": "ArtificialAnlys", "handle": "@ArtificialAnlys", "url": "https://x.com/search?q=Artificial%20Analysis%20AutomationBench-AA%20Zapier&src=typed_query&f=live", "date": "past month search result", "topic": "Artificial Analysis AutomationBench-AA Zapier", "models": "Gemini; Claude; GPT", "benchmarks": "AutomationBench-AA", "why": "SaaS-workflow agent benchmark", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 33, "author": "ArtificialAnlys", "handle": "@ArtificialAnlys", "url": "https://x.com/search?q=Artificial%20Analysis%20Harvey%20LAB-AA%20legal%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Artificial Analysis Harvey LAB-AA legal benchmark", "models": "Claude; GPT; GLM", "benchmarks": "Harvey LAB-AA", "why": "Legal-agent benchmark and model cost spread", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 34, "author": "ArtificialAnlys", "handle": "@ArtificialAnlys", "url": "https://x.com/search?q=Artificial%20Analysis%20AA-Briefcase&src=typed_query&f=live", "date": "past month search result", "topic": "Artificial Analysis AA-Briefcase", "models": "Fable 5; GLM-5.2", "benchmarks": "AA-Briefcase", "why": "Agentic business-work benchmark", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 35, "author": "ArtificialAnlys", "handle": "@ArtificialAnlys", "url": "https://x.com/search?q=Artificial%20Analysis%20AA-Omniscience&src=typed_query&f=live", "date": "past month search result", "topic": "Artificial Analysis AA-Omniscience", "models": "Fable 5; GPT-5.5; Gemini", "benchmarks": "AA-Omniscience", "why": "Knowledge/hallucination benchmark", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 36, "author": "ArtificialAnlys", "handle": "@ArtificialAnlys", "url": "https://x.com/search?q=Artificial%20Analysis%20CritPt%20physics%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Artificial Analysis CritPt physics benchmark", "models": "GLM-5.2; Opus 4.8", "benchmarks": "CritPt", "why": "Scientific reasoning benchmark", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 37, "author": "ArtificialAnlys", "handle": "@ArtificialAnlys", "url": "https://x.com/search?q=Artificial%20Analysis%20HLE%20benchmark%20model&src=typed_query&f=live", "date": "past month search result", "topic": "Artificial Analysis HLE benchmark model", "models": "GLM-5.2; GPT-5.5; Claude", "benchmarks": "HLE", "why": "Hard reasoning benchmark commentary", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 38, "author": "lmarena_ai", "handle": "@lmarena_ai", "url": "https://x.com/search?q=LM%20Arena%20leaderboard%20GPT%20Claude%20Gemini&src=typed_query&f=live", "date": "past month search result", "topic": "LM Arena leaderboard GPT Claude Gemini", "models": "GPT; Claude; Gemini", "benchmarks": "Chatbot Arena; Arena leaderboard", "why": "Human preference/model-ranking updates", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 39, "author": "lmsysorg", "handle": "@lmsysorg", "url": "https://x.com/search?q=Chatbot%20Arena%20leaderboard%20new%20model&src=typed_query&f=live", "date": "past month search result", "topic": "Chatbot Arena leaderboard new model", "models": "frontier models", "benchmarks": "Chatbot Arena", "why": "Arena ranking update", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 40, "author": "livebench_ai", "handle": "@livebench_ai", "url": "https://x.com/search?q=LiveBench%20leaderboard%20GPT%20Claude%20Gemini&src=typed_query&f=live", "date": "past month search result", "topic": "LiveBench leaderboard GPT Claude Gemini", "models": "GPT; Claude; Gemini", "benchmarks": "LiveBench", "why": "Contamination-resistant benchmark updates", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 41, "author": "aider_chat", "handle": "@aider_chat", "url": "https://x.com/search?q=Aider%20Polyglot%20GPT%20Claude%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Aider Polyglot GPT Claude benchmark", "models": "GPT; Claude; Gemini", "benchmarks": "Aider Polyglot", "why": "Coding model benchmark updates", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 42, "author": "paulgauthier", "handle": "@paulgauthier", "url": "https://x.com/search?q=Paul%20Gauthier%20aider%20benchmark%20Claude%20GPT&src=typed_query&f=live", "date": "past month search result", "topic": "Paul Gauthier aider benchmark Claude GPT", "models": "GPT; Claude", "benchmarks": "Aider Polyglot", "why": "Creator commentary on coding model performance", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 43, "author": "SWEbench", "handle": "@SWEbench", "url": "https://x.com/search?q=SWE-bench%20GPT%20Claude%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "SWE-bench GPT Claude benchmark", "models": "GPT; Claude; Gemini", "benchmarks": "SWE-bench Verified; SWE-bench Multilingual", "why": "Coding-agent benchmark updates", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 44, "author": "terminal_bench", "handle": "@terminal_bench", "url": "https://x.com/search?q=Terminal-Bench%20GPT%20Claude%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Terminal-Bench GPT Claude benchmark", "models": "GPT; Claude", "benchmarks": "Terminal-Bench", "why": "Terminal-agent benchmark", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 45, "author": "METR_Evals", "handle": "@METR_Evals", "url": "https://x.com/search?q=METR%20GPT-5.6%20benchmark%20gaming&src=typed_query&f=live", "date": "past month search result", "topic": "METR GPT-5.6 benchmark gaming", "models": "GPT-5.6 Sol", "benchmarks": "METR agent evals", "why": "Agentic benchmark gaming/behavioral caveats", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 46, "author": "EpochAIResearch", "handle": "@EpochAIResearch", "url": "https://x.com/search?q=Epoch%20AI%20benchmark%20GPT%20Claude%20Gemini&src=typed_query&f=live", "date": "past month search result", "topic": "Epoch AI benchmark GPT Claude Gemini", "models": "frontier models", "benchmarks": "benchmark trends", "why": "Macro benchmark/progress analysis", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 47, "author": "StanfordCRFM", "handle": "@StanfordCRFM", "url": "https://x.com/search?q=HELM%20Stanford%20leaderboard%20GPT%20Claude&src=typed_query&f=live", "date": "past month search result", "topic": "HELM Stanford leaderboard GPT Claude", "models": "GPT; Claude; Gemini", "benchmarks": "HELM; MedHELM", "why": "Holistic LLM eval framework updates", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 48, "author": "huggingface", "handle": "@huggingface", "url": "https://x.com/search?q=Hugging%20Face%20Open%20LLM%20Leaderboard%20new%20model&src=typed_query&f=live", "date": "past month search result", "topic": "Hugging Face Open LLM Leaderboard new model", "models": "open models", "benchmarks": "Open LLM Leaderboard", "why": "Open model leaderboard updates", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 49, "author": "natolambert", "handle": "@natolambert", "url": "https://x.com/search?q=Nathan%20Lambert%20open%20model%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Nathan Lambert open model benchmark", "models": "Llama; Qwen; DeepSeek; GLM", "benchmarks": "open model benchmarks", "why": "Post-training/open-model evaluation commentary", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 50, "author": "swyx", "handle": "@swyx", "url": "https://x.com/search?q=swyx%20GPT-5.6%20benchmark%20eval&src=typed_query&f=live", "date": "past month search result", "topic": "swyx GPT-5.6 benchmark eval", "models": "GPT-5.6; Claude", "benchmarks": "benchmarks/evals", "why": "AI engineering interpretation of benchmark results", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 51, "author": "simonw", "handle": "@simonw", "url": "https://x.com/search?q=Simon%20Willison%20GPT-5.6%20benchmark%20eval&src=typed_query&f=live", "date": "past month search result", "topic": "Simon Willison GPT-5.6 benchmark eval", "models": "GPT-5.6; Claude; Gemini", "benchmarks": "hands-on evals", "why": "Practical testing of model behavior", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 52, "author": "HamelHusain", "handle": "@HamelHusain", "url": "https://x.com/search?q=Hamel%20Husain%20eval%20benchmark%20model&src=typed_query&f=live", "date": "past month search result", "topic": "Hamel Husain eval benchmark model", "models": "LLMs", "benchmarks": "evals", "why": "Product-eval methodology and model testing", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 53, "author": "OpenRouterAI", "handle": "@OpenRouterAI", "url": "https://x.com/search?q=OpenRouter%20model%20benchmark%20price%20performance&src=typed_query&f=live", "date": "past month search result", "topic": "OpenRouter model benchmark price performance", "models": "frontier/open models", "benchmarks": "price/performance", "why": "Routing/API market signals", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 54, "author": "paperswithcode", "handle": "@paperswithcode", "url": "https://x.com/search?q=Papers%20with%20Code%20leaderboard%20LLM%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Papers with Code leaderboard LLM benchmark", "models": "LLMs", "benchmarks": "leaderboards", "why": "Benchmark leaderboard amplification", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 55, "author": "ScaleAILabs", "handle": "@ScaleAILabs", "url": "https://x.com/search?q=Scale%20Labs%20leaderboard%20model%20benchmarks&src=typed_query&f=live", "date": "past month search result", "topic": "Scale Labs leaderboard model benchmarks", "models": "frontier models", "benchmarks": "Scale Labs leaderboards", "why": "Expert benchmark leaderboards", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 56, "author": "ScaleAILabs", "handle": "@ScaleAILabs", "url": "https://x.com/search?q=Scale%20Labs%20DrugDiscoveryBench&src=typed_query&f=live", "date": "past month search result", "topic": "Scale Labs DrugDiscoveryBench", "models": "GPT; Gemini; Claude", "benchmarks": "DrugDiscoveryBench", "why": "Life-science/drug discovery benchmark", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 57, "author": "phylo_bio", "handle": "@phylo_bio", "url": "https://x.com/search?q=Phylo%20DrugDiscoveryBench%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Phylo DrugDiscoveryBench benchmark", "models": "AI agents", "benchmarks": "DrugDiscoveryBench; BiomniBench", "why": "Biomedical agent benchmark", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 58, "author": "OpenAI", "handle": "@OpenAI", "url": "https://x.com/search?q=OpenAI%20LifeSciBench%20GPT-Rosalind&src=typed_query&f=live", "date": "past month search result", "topic": "OpenAI LifeSciBench GPT-Rosalind", "models": "GPT-Rosalind; GPT-5.5", "benchmarks": "LifeSciBench", "why": "Life-science model benchmark", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 59, "author": "OpenAI", "handle": "@OpenAI", "url": "https://x.com/search?q=OpenAI%20HealthBench%20GPT-5.6&src=typed_query&f=live", "date": "past month search result", "topic": "OpenAI HealthBench GPT-5.6", "models": "GPT-5.6", "benchmarks": "HealthBench", "why": "Medical model benchmark", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 60, "author": "OpenAI", "handle": "@OpenAI", "url": "https://x.com/search?q=OpenAI%20Terminal-Bench%20GPT-5.6%20Sol&src=typed_query&f=live", "date": "past month search result", "topic": "OpenAI Terminal-Bench GPT-5.6 Sol", "models": "GPT-5.6 Sol", "benchmarks": "Terminal-Bench", "why": "Coding/terminal benchmark", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 61, "author": "OpenAIDevs", "handle": "@OpenAIDevs", "url": "https://x.com/search?q=OpenAI%20Developers%20GPT-5.6%20evals&src=typed_query&f=live", "date": "past month search result", "topic": "OpenAI Developers GPT-5.6 evals", "models": "GPT-5.6", "benchmarks": "coding; tool use", "why": "Developer benchmark/performance details", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 62, "author": "AnthropicAI", "handle": "@AnthropicAI", "url": "https://x.com/search?q=Anthropic%20Claude%20Sonnet%205%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Anthropic Claude Sonnet 5 benchmark", "models": "Claude Sonnet 5", "benchmarks": "coding; reasoning", "why": "Official Claude benchmark claims", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 63, "author": "GoogleDeepMind", "handle": "@GoogleDeepMind", "url": "https://x.com/search?q=Google%20DeepMind%20Gemini%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Google DeepMind Gemini benchmark", "models": "Gemini", "benchmarks": "reasoning; coding; multimodal", "why": "Official Gemini performance claims", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 64, "author": "MistralAI", "handle": "@MistralAI", "url": "https://x.com/search?q=Mistral%20AI%20model%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Mistral AI model benchmark", "models": "Mistral models", "benchmarks": "open/frontier benchmarks", "why": "Mistral model release performance", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 65, "author": "AIatMeta", "handle": "@AIatMeta", "url": "https://x.com/search?q=Meta%20AI%20Llama%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Meta AI Llama benchmark", "models": "Llama", "benchmarks": "open model benchmarks", "why": "Llama benchmark/release comparisons", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 66, "author": "deepseek_ai", "handle": "@deepseek_ai", "url": "https://x.com/search?q=DeepSeek%20model%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "DeepSeek model benchmark", "models": "DeepSeek", "benchmarks": "coding/reasoning benchmarks", "why": "DeepSeek model-performance posts", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 67, "author": "Alibaba_Qwen", "handle": "@Alibaba_Qwen", "url": "https://x.com/search?q=Qwen%20benchmark%20model%20performance&src=typed_query&f=live", "date": "past month search result", "topic": "Qwen benchmark model performance", "models": "Qwen", "benchmarks": "open model benchmarks", "why": "Qwen model benchmark results", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 68, "author": "Zai_org", "handle": "@Zai_org", "url": "https://x.com/search?q=Z.ai%20GLM-5.2%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Z.ai GLM-5.2 benchmark", "models": "GLM-5.2", "benchmarks": "AA Index; GDPval-AA", "why": "Source account for GLM-5.2 benchmark claims", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 69, "author": "cursor_ai", "handle": "@cursor_ai", "url": "https://x.com/search?q=Cursor%20Grok%204.5%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Cursor Grok 4.5 benchmark", "models": "Grok 4.5", "benchmarks": "coding benchmarks", "why": "Coding-product benchmark context", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 70, "author": "cognition_labs", "handle": "@cognition_labs", "url": "https://x.com/search?q=Cognition%20Devin%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Cognition Devin benchmark", "models": "coding agents", "benchmarks": "SWE-bench", "why": "Coding agent performance", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 71, "author": "windsurf_ai", "handle": "@windsurf_ai", "url": "https://x.com/search?q=Windsurf%20model%20benchmark%20coding&src=typed_query&f=live", "date": "past month search result", "topic": "Windsurf model benchmark coding", "models": "coding models", "benchmarks": "coding benchmark", "why": "AI IDE model-performance posts", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 72, "author": "Replit", "handle": "@Replit", "url": "https://x.com/search?q=Replit%20Agent%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Replit Agent benchmark", "models": "coding agents", "benchmarks": "coding/task benchmarks", "why": "AI coding-agent performance", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 73, "author": "LangChainAI", "handle": "@LangChainAI", "url": "https://x.com/search?q=LangChain%20agent%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "LangChain agent benchmark", "models": "agent models", "benchmarks": "agent evals", "why": "Agent framework evaluation", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 74, "author": "llama_index", "handle": "@llama_index", "url": "https://x.com/search?q=LlamaIndex%20RAG%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "LlamaIndex RAG benchmark", "models": "LLMs", "benchmarks": "RAG evals", "why": "RAG/model evaluation", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 75, "author": "modal_labs", "handle": "@modal_labs", "url": "https://x.com/search?q=Modal%20GPU%20inference%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Modal GPU inference benchmark", "models": "inference stacks", "benchmarks": "latency/throughput", "why": "Inference-performance benchmarks", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 76, "author": "AnyscaleCompute", "handle": "@AnyscaleCompute", "url": "https://x.com/search?q=Anyscale%20LLM%20serving%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Anyscale LLM serving benchmark", "models": "LLMs", "benchmarks": "serving throughput", "why": "Infrastructure benchmarks", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 77, "author": "togethercompute", "handle": "@togethercompute", "url": "https://x.com/search?q=Together%20AI%20benchmark%20throughput&src=typed_query&f=live", "date": "past month search result", "topic": "Together AI benchmark throughput", "models": "open models", "benchmarks": "inference speed/cost", "why": "Model-serving benchmark posts", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 78, "author": "FireworksAI_HQ", "handle": "@FireworksAI_HQ", "url": "https://x.com/search?q=Fireworks%20AI%20inference%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Fireworks AI inference benchmark", "models": "open models", "benchmarks": "serving latency/throughput", "why": "Inference benchmark posts", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 79, "author": "GroqInc", "handle": "@GroqInc", "url": "https://x.com/search?q=Groq%20inference%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Groq inference benchmark", "models": "LLMs", "benchmarks": "tokens/sec latency", "why": "Inference speed benchmarks", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 80, "author": "CerebrasSystems", "handle": "@CerebrasSystems", "url": "https://x.com/search?q=Cerebras%20inference%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Cerebras inference benchmark", "models": "LLMs", "benchmarks": "tokens/sec", "why": "Inference speed benchmarks", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 81, "author": "perplexity_ai", "handle": "@perplexity_ai", "url": "https://x.com/search?q=Perplexity%20model%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Perplexity model benchmark", "models": "Perplexity models", "benchmarks": "search/model quality", "why": "AI search model performance", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 82, "author": "vellum_ai", "handle": "@vellum_ai", "url": "https://x.com/search?q=Vellum%20GPT-5.6%20benchmarks%20explained&src=typed_query&f=live", "date": "past month search result", "topic": "Vellum GPT-5.6 benchmarks explained", "models": "GPT-5.6", "benchmarks": "benchmark analysis", "why": "Practical model comparison", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 83, "author": "latentspacepod", "handle": "@latentspacepod", "url": "https://x.com/search?q=Latent%20Space%20AINews%20GPT-5.6%20benchmarks&src=typed_query&f=live", "date": "past month search result", "topic": "Latent Space AINews GPT-5.6 benchmarks", "models": "GPT-5.6", "benchmarks": "benchmark roundup", "why": "Benchmark news synthesis", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 84, "author": "deedydas", "handle": "@deedydas", "url": "https://x.com/search?q=Deedy%20Das%20model%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Deedy Das model benchmark", "models": "GPT; Claude; Gemini", "benchmarks": "benchmark summaries", "why": "High-signal benchmark commentary", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 85, "author": "AlphaSignalAI", "handle": "@AlphaSignalAI", "url": "https://x.com/search?q=AlphaSignal%20benchmark%20model&src=typed_query&f=live", "date": "past month search result", "topic": "AlphaSignal benchmark model", "models": "AI models", "benchmarks": "research benchmarks", "why": "AI R&D benchmark amplification", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 86, "author": "cwolferesearch", "handle": "@cwolferesearch", "url": "https://x.com/search?q=Cameron%20Wolfe%20model%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Cameron Wolfe model benchmark", "models": "AI models", "benchmarks": "benchmark analysis", "why": "Technical paper/model benchmark breakdowns", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 87, "author": "rasbt", "handle": "@rasbt", "url": "https://x.com/search?q=Sebastian%20Raschka%20model%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Sebastian Raschka model benchmark", "models": "open models", "benchmarks": "LLM benchmarks", "why": "Technical model comparison", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 88, "author": "_philschmid", "handle": "@_philschmid", "url": "https://x.com/search?q=Philipp%20Schmid%20model%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Philipp Schmid model benchmark", "models": "open models", "benchmarks": "leaderboards/inference", "why": "Hugging Face benchmark/inference testing", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 89, "author": "Thom_Wolf", "handle": "@Thom_Wolf", "url": "https://x.com/search?q=Thomas%20Wolf%20open%20model%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Thomas Wolf open model benchmark", "models": "open models", "benchmarks": "leaderboards", "why": "Open-source model performance", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 90, "author": "ClementDelangue", "handle": "@ClementDelangue", "url": "https://x.com/search?q=Clement%20Delangue%20open%20model%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Clement Delangue open model benchmark", "models": "open models", "benchmarks": "leaderboards", "why": "Open model release/benchmark amplification", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 91, "author": "abhi1thakur", "handle": "@abhi1thakur", "url": "https://x.com/search?q=Kaggle%20Abhishek%20model%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Kaggle Abhishek model benchmark", "models": "ML models", "benchmarks": "Kaggle/evals", "why": "Applied model-performance tests", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 92, "author": "valence_ai", "handle": "@valence_ai", "url": "https://x.com/search?q=Valence%20Labs%20Polaris%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Valence Labs Polaris benchmark", "models": "drug discovery models", "benchmarks": "Polaris", "why": "Drug-discovery benchmark platform", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 93, "author": "wpwalters", "handle": "@wpwalters", "url": "https://x.com/search?q=Pat%20Walters%20ADMET%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Pat Walters ADMET benchmark", "models": "cheminformatics models", "benchmarks": "ADMET/blind challenge", "why": "Drug-discovery benchmark critique", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 94, "author": "ProjectTDC", "handle": "@ProjectTDC", "url": "https://x.com/search?q=TDC%20benchmark%20model&src=typed_query&f=live", "date": "past month search result", "topic": "TDC benchmark model", "models": "therapeutic ML models", "benchmarks": "TDC", "why": "Therapeutics benchmark ecosystem", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 95, "author": "openbioml", "handle": "@openbioml", "url": "https://x.com/search?q=OpenBioML%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "OpenBioML benchmark", "models": "biology/chem models", "benchmarks": "BioML/Chem benchmarks", "why": "Open biology ML benchmark amplification", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 96, "author": "chaidiscovery", "handle": "@chaidiscovery", "url": "https://x.com/search?q=Chai%20Discovery%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Chai Discovery benchmark", "models": "structure models", "benchmarks": "protein/structure benchmarks", "why": "Biomolecular model performance", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 97, "author": "RosettaCommons", "handle": "@RosettaCommons", "url": "https://x.com/search?q=Rosetta%20Commons%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "Rosetta Commons benchmark", "models": "protein design models", "benchmarks": "protein modeling benchmarks", "why": "Protein modeling benchmark network", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 98, "author": "iScienceLuvr", "handle": "@iScienceLuvr", "url": "https://x.com/search?q=MedARC%20Medmarks%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "MedARC Medmarks benchmark", "models": "medical LLMs", "benchmarks": "Medmarks", "why": "Medical LLM benchmark", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 99, "author": "AISecurityInst", "handle": "@AISecurityInst", "url": "https://x.com/search?q=AI%20Security%20Institute%20Inspect%20Evals&src=typed_query&f=live", "date": "past month search result", "topic": "AI Security Institute Inspect Evals", "models": "frontier models", "benchmarks": "Inspect Evals", "why": "Model safety/eval framework", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" }, { "rank": 100, "author": "ai_risks", "handle": "@ai_risks", "url": "https://x.com/search?q=CAIS%20WMDP%20benchmark&src=typed_query&f=live", "date": "past month search result", "topic": "CAIS WMDP benchmark", "models": "LLMs", "benchmarks": "WMDP", "why": "Bio/chem/cyber risk benchmark", "engagement": "n/a", "source_note": "X search URL because exact post URL was not surfaced unauthenticated", "confidence": "Low" } ] }