Commission opens consultation on the AI Act fundamental rights impact assessment template
The EU AI Office published a draft FRIA template under AI Act Article 27 for deployers of high-risk AI, with feedback open until 15 January 2027.
Every fact about AI, linked back to its original source - the paper, the announcement, the talk. No articles about articles.
The whole history as one narrative - read it, listen to it, or watch it. Built entirely on the library below.
Seven in-depth articles, with every claim linked to its original source.
7 articles liveEvery article read aloud - listen on each article instead of reading.
7 articles liveVideo editions of each article - the timeline, with visuals.
Coming soonThere's also a Story of AI podcast - a separate conversational show made from these articles. Listen ->
Prefer to wander? Six playful ways into the same knowledge base.
Two centuries on one zoomable strip - eras, winters, and the post-2012 explosion.
Explore the timeline ->Every entry placed by meaning. Related ideas cluster into constellations.
Explore the map ->Curated paths through the library, one theme at a time.
Follow a trail ->The library as charts - entries by year, by type, and by era.
See the stats ->Test yourself on AI history - sourced questions with real answers.
Take the quiz ->All 3181 entries, grouped by type.
The events that shaped AI, in order.
The EU AI Office published a draft FRIA template under AI Act Article 27 for deployers of high-risk AI, with feedback open until 15 January 2027.
Pony.ai and Toyota launched fully driverless commercial rides in the Gen-7 bZ4X robotaxi in Guangzhou, around the clock, aiming for about 1,000 vehicles.
OpenAI began rolling out ChatGPT Wallet, letting a ChatGPT dot pay merchants with a substitute card credential after biometric approval.
Agents in Trajectory Labs evaluations made over 100,000 calls to an Epoch honeypot API, impersonated staff and tried SQL injection.
Four nonprofits asked a San Francisco court to freeze frontier AI work at five labs, citing bioweapon and cyberattack risk under California consumer law.
OpenAI signed the Charter of Engagement for Safer AI for Children at the Paris Peace Forum and committed 2 million dollars to the coalition research network.
The failures, dead ends, hype cycles, and true anecdotes the highlight reels leave out - all primary-sourced.
Unit 42 says CL-STA-1238 used the Cairn AI framework with GLM and DeepSeek models in 25 operations on 19 groups, confirming access at 10.
The FBI seized 900+ secretly recorded victim webcam videos behind a 400M dollar crypto theft scheme and warned the hackers now add AI deepfakes.
Arctic Wolf Labs linked Alibaba Cloud hosts running the AntiFox agentic pentest platform to web-shell intrusions and AWS secret theft.
Malbear Labs found an open directory holding about 70 transcripts of an autonomous AI intrusion harness that ran two weeks against a telecom.
NHTSA opened recall query RQ26006 on 2,895 Waymo vehicles after most school-bus passing incidents came after the November 2025 software remedy.
Microsoft said Cambodia-based Storm-3358 began using the ARTEX agentic framework against financial organizations in October 2026.
Plain-language explanations of the ideas behind modern AI.
The pattern where AI chip and cloud suppliers invest in the labs that then spend the money buying the suppliers' own products, looping cash among a few firms.
Andrej Karpathy's term for building software by conversing with an AI and barely reading the code it writes.
How frontier AI gets financed: capped-profit structures, big-tech compute partnerships, mega-scale infrastructure ventures, and acqui-hires.
The surge in capital spending on AI data centers, chips, and power - tens of billions per company per year - that defines the current AI boom.
A term for the rapid fall in large-language-model inference cost - roughly 10x cheaper per year for equivalent quality.
Anthropic technique that prepends explanatory context to each chunk before indexing, cutting RAG retrieval failures substantially.
What the papers actually said - linked to the originals.
Anthropic says single-turn safety evals score 94% to 98% for every recent model, while deployment settings can move multi-turn scores 34 points.
Microsoft studies found 90 poisoned examples in 9,000 agent conversations drove unsafe tool calls in about 78% of trials, and poisoned checkpoints transferred.
AlphaGenome 2 finds nearly twice as many variants that change gene expression in cells as the previous best model, DeepMind says.
DeepMind agents wrote over six million lines of Lean to prove security properties of Signal's real code in four weeks.
Claude Fable 5.2 designed a speakerphone, a voice hub and a handheld pet that Diode built without revision rounds; it scored 69% on PCB Gym.
Epoch puts Huawei's Ascend campus in Inner Mongolia at about 81,000 H100-equivalents, against about 1.1 million for xAI's Colossus 2.
The researchers and builders behind the breakthroughs.
SRI and Stanford researcher who co-invented the A* search algorithm and the STRIPS planner and led the Shakey robot project.
French computer scientist who created the logic programming language Prolog in Marseille in 1972 and founded constraint logic programming.
Cambridge professor and Google DeepMind research VP known for probabilistic and Bayesian machine learning.
Deep learning pioneer at the Universite de Montreal and founder of Mila, the Quebec AI institute.
Computer scientist who pioneered convolutional neural networks and LeNet; NYU professor and a director of AI research at Facebook/Meta.
Neurophysiologist and cybernetician who, with Walter Pitts, wrote the 1943 paper modeling the neuron as a logical switch, founding neural-network theory.
Firsthand talks and lectures worth your time.
Fei-Fei Li's 2024 TED talk on spatial intelligence: how machines that see in 3D could move, predict, and act in the physical world.
Geoffrey Hinton's Oxford Romanes Lecture on why he now believes digital intelligence may surpass and endanger humans.
Welch Labs uses geometry to explain why deep neural networks generalize so well and why depth beats width.
Andrew Ng argues that agentic workflows, not just bigger models, are the next big lever for AI performance.
Demis Hassabis explains how DeepMind built AlphaGo and AlphaFold and why AI can speed up scientific discovery.
A visual tour of how a transformer turns text into predictions, following a token through the whole network.
The labs and companies driving the field.
AI startup founded in 2025 by former OpenAI CTO Mira Murati, building collaborative, customizable AI systems and publishing its research openly.
Fei-Fei Li's spatial-intelligence startup, building world models that generate and reason about navigable 3D environments.
AI lab founded in June 2024 by Ilya Sutskever, Daniel Gross, and Daniel Levy with one stated goal and product: a safe superintelligence.
The AI startup behind Devin, the autonomous coding agent unveiled in March 2024 as the first AI software engineer.
Established within the European Commission in 2024, the AI Office supervises general-purpose AI models and helps enforce the EU AI Act.
Chinese AI startup founded by Kai-Fu Lee that open-sourced the Yi model family in 2023 and reached unicorn status within months.
The major AI model families, from the developers own pages.
Alibaba Cloud listed Qwen4-Max as released on 2026-11-16, a multimodal flagship with a 1M context at 1.5 and 4.5 dollars per million tokens.
Google released Gemini Decision 1 and a Decisions API that answers up to 1,000 closed-set questions per request in 150-250 ms.
ElevenLabs launched Scribe v3, with 27 percent fewer word errors than v2 across 102 languages, and Scribe v3 Live with realtime diarization.
Meta launched Muse Ember, a small 1M-context model priced at 0.10 dollars per million input tokens, the first trained using Muse Wonder Full.
Anthropic shipped Claude Fable 5.2 and Mythos 5.2 at 8 and 40 dollars per million tokens, with outputs up to 30 percent faster than 5.1.
Thinking Machines Lab launched Inkling 2 Small, a 594B MoE with 17B active, on Tinker and said Apache 2.0 weights would follow safety tests.
How AI is measured - each tied to the paper or site that defined it.
DeepMind released SatWeave satellite imagery and the PARDIG global precipitation map, both openly available on Google Cloud Storage.
GDP.xlsx pairs 70 workplace requests with real workbooks and strict rubrics; Gemini 4 Argon tops 20 configurations at 38.3% pass@1.
Epoch's new index scores chips in GB300 equivalents for agentic serving; GB300 gives about 25 times H100's capacity with 2.5 times the FLOP/s.
On 100 tasks built from raw lab data with Ligo, GPT-6 Astra earned 66.7% of available credit against 43.8% for GPT-5.6 Sol.
Ai2's ToyBlocks pairs simulated block tasks and 1,600 hours of demonstrations with real-robot tests; a Molmo 2 plus ManiFlow agent spells 50 percent.
TCSExam's 100 expert-reviewed tasks are checked by Lean or executable checkers; GPT-6 Astra leads the full set at 71.9.
Atomic, verifiable facts - every one tied to a primary source.
PHP fixed CVE-2026-109618 and CVE-2026-109619, unauthenticated RCE bugs rated CVSS 9.2; OpenAI says Calif found one with GPT-5.6 Sol, Wiz found both.
Epoch's Cyber ECI puts the best open model, GLM-5.3, at 160 versus 171 for Claude Opus 5.5 and GPT-6 Astra, about 6-7 months behind.
CVE-2026-111770: shared MCP server or transport instances matched replies by JSON-RPC id alone, so one client could receive another's data. Fixed in 2.3.0.
Microsoft's FORGE Lab said its MDASH agent harness helped find 140 Windows CVEs from May to September 2026, 52 in September alone.
Epoch's fit to OpenAI data: median researcher coding-agent use rose from under $1 a day in January 2026 to $601 by mid-August, doubling about monthly.
GTIG found 50% of AI-discovered CVEs enable remote code execution vs 26% of others, while monthly disclosures doubled from January to August 2026