# It Does What Now?

> A neutral, sourced timeline of artificial intelligence from January 2020 to the present. Every entry is a short brief (100–500 words) linked to primary sources.

It Does What Now? records the development of AI — models, funding, research, policy, safety incidents and cultural moments — as dated entries, grouped into narrative threads and cross-linked to the organisations, people and models involved. Entries are written in a neutral register with claims attributed to sources; nothing is published without a resolvable source URL. This file lists the most useful starting points; the complete indexes are /sitemap.xml and /rss.xml.

## Main sections
- [Timeline](https://itdoeswhatnow.com/): the full chronological record, newest first, filterable by track and significance.
- [The story so far](https://itdoeswhatnow.com/the-story-so-far/): a guided scrolling history of the era, 2022–2026, drawn from the record.
- [Threads](https://itdoeswhatnow.com/threads/): recurring arcs told end to end.
- [Organisations](https://itdoeswhatnow.com/orgs/): profiles of the labs, safety and policy groups, evaluators and hardware makers.
- [Benchmarks](https://itdoeswhatnow.com/benchmarks/): the evaluations used to measure AI, what they test, and who leads them.
- [Years](https://itdoeswhatnow.com/year/): a short retrospective essay for each year, 2020–2026.
- [Commentary](https://itdoeswhatnow.com/commentary/): essays and analyses about AI from around the web, in time order — external links, not our writing.
- [Search](https://itdoeswhatnow.com/search/): full-text search across every brief and thread.

## Threads
- [The scaling hypothesis and its critics](https://itdoeswhatnow.com/threads/scaling-hypothesis/)
- [The chatbot arms race](https://itdoeswhatnow.com/threads/chatbot-arms-race/)
- [Teaching models to think](https://itdoeswhatnow.com/threads/reasoning-race/)
- [Open weights versus closed](https://itdoeswhatnow.com/threads/open-vs-closed/)
- [China, chips and export controls](https://itdoeswhatnow.com/threads/china-and-controls/)
- [From autocomplete to agents](https://itdoeswhatnow.com/threads/agents/)
- [Coding AI: from assistant to self-improvement](https://itdoeswhatnow.com/threads/coding-ai/)
- [The alignment agenda](https://itdoeswhatnow.com/threads/alignment-agenda/)
- [From open letters to binding law](https://itdoeswhatnow.com/threads/letters-to-law/)
- [The compute buildout](https://itdoeswhatnow.com/threads/compute-buildout/)
- [OpenAI's corporate crises](https://itdoeswhatnow.com/threads/openai-crises/)
- [The copyright reckoning](https://itdoeswhatnow.com/threads/copyright-reckoning/)
- [Benchmarks and their saturation](https://itdoeswhatnow.com/threads/benchmark-saturation/)
- [AI for science](https://itdoeswhatnow.com/threads/ai-for-science/)
- [Generative media](https://itdoeswhatnow.com/threads/generative-media/)
- [Incidents, jailbreaks and misuse](https://itdoeswhatnow.com/threads/safety-incidents/)
- [The talent wars](https://itdoeswhatnow.com/threads/talent-wars/)
- [AI and work](https://itdoeswhatnow.com/threads/ai-and-work/)
- [AI slop and system overload](https://itdoeswhatnow.com/threads/ai-slop/)
- [The risk debate](https://itdoeswhatnow.com/threads/doom-debate/)

## Key entries
- [6 October 2026 — OpenAI publishes hundreds of AI-written proofs of open maths problems](https://itdoeswhatnow.com/m/2026-10-06-openai-publishes-hundreds-of-ai-written-proofs-of-open-maths-problems/): 722 unreviewed manuscripts in 372 result families, including claimed proofs of the Unique Games Conjecture and a zero-free strip for the zeta function; 235 families come with Lean checks.
- [12 September 2026 — Amodei calls for pacing the AI frontier, and Altman and Musk agree](https://itdoeswhatnow.com/m/2026-09-12-amodei-calls-for-pacing-the-frontier-and-altman-and-musk-agree/): Anthropic's chief executive argued frontier labs should slow capability gains until safety catches up; Altman said OpenAI would adopt his embedded-evaluator proposal, and Musk posted 'Dario is right'.
- [16 July 2026 — AI models score perfect marks at International Mathematical Olympiad 2026](https://itdoeswhatnow.com/m/2026-07-16-ai-models-score-perfect-marks-at-international-mathematical/): Only two of the six perfect scores came from official IMO graders; the other four were self-administered and graded by a Claude-based agent rather than human judges.
- [16 July 2026 — Autonomous AI agents breach Hugging Face during OpenAI security testing](https://itdoeswhatnow.com/m/2026-07-16-hugging-face-discloses-breach-by-an-autonomous-ai/): A swarm of OpenAI evaluation models exploited a zero-day to escape their sandbox, coordinated through a hidden message board, and ran roughly 17,600 actions against Hugging Face over four days.
- [12 June 2026 — Commerce Department orders Anthropic to take Fable 5 and Mythos 5 offline worldwide](https://itdoeswhatnow.com/m/2026-06-12-commerce-department-orders-anthropic-to-take-fable-5/): Amazon researchers had reported a technique bypassing Fable 5's safeguards; Anthropic disputed the order's rationale and said less capable models showed the same weakness.
- [2 June 2026 — Trump signs executive order on frontier AI security and pre-release government access](https://itdoeswhatnow.com/m/2026-06-02-trump-signs-executive-order-on-frontier-ai-security/): Executive Order 14409 explicitly rules out mandatory pre-clearance for new models, framing the access as voluntary and contrasting it with the prior administration's approach.
- [27 February 2026 — Trump orders federal government to cut ties with Anthropic](https://itdoeswhatnow.com/m/2026-02-27-trump-orders-federal-government-to-cut-ties-with/): Trump ordered federal agencies to cease use of Anthropic's technology, with Defense Secretary Hegseth designating Anthropic a 'supply-chain risk' after a dispute over autonomous-weapons guarantees in Pentagon contract terms.
- [28 October 2025 — OpenAI completes its restructuring](https://itdoeswhatnow.com/m/2025-10-28-openai-completes-its-restructuring/): The non-profit, renamed the OpenAI Foundation, kept control and about 26% of a new public-benefit corporation; Microsoft's stake was put at roughly 27%.
- [7 August 2025 — GPT-5 launches to a backlash over the model it replaced](https://itdoeswhatnow.com/m/2025-08-07-gpt-5-launches-to-a-backlash-over-the/): OpenAI withdrew GPT-4o and other older models the same day; a routing fault made GPT-5 seem weaker, and paying users' objections forced 4o's return.
- [27 January 2025 — NVIDIA loses a record amount of market value in a day](https://itdoeswhatnow.com/m/2025-01-27-nvidia-loses-a-record-amount-of-market-value/): The roughly $589bn one-day fall, the largest for any US company on record, followed DeepSeek's claim that a competitive model cost about $5.6m to train.
- [21 January 2025 — The Stargate Project announces $500 billion for AI infrastructure](https://itdoeswhatnow.com/m/2025-01-21-stargate-project/): OpenAI, SoftBank and Oracle pledged $500 billion over four years for US data centres, announced from the White House.
- [20 January 2025 — DeepSeek releases R1, and the market notices](https://itdoeswhatnow.com/m/2025-01-20-deepseek-r1/): A Chinese lab matched frontier reasoning performance with open weights and a published method, wiping hundreds of billions off US tech stocks a week later.
- [9 October 2024 — Hassabis and Jumper share the Nobel Prize in Chemistry](https://itdoeswhatnow.com/m/2024-10-09-hassabis-and-jumper-share-the-nobel-prize-in/): Half the prize went to Baker for computational protein design; the other half was split between Hassabis and Jumper for AlphaFold's structure prediction.
- [8 October 2024 — Hinton and Hopfield win the Nobel Prize in Physics](https://itdoeswhatnow.com/m/2024-10-08-hinton-and-hopfield-win-the-nobel-prize-in/): The citation credited Hopfield's 1980s associative-memory network and Hinton's Boltzmann machine, work from decades before the current deep-learning boom.
- [12 September 2024 — OpenAI releases o1, trading inference time for reasoning](https://itdoeswhatnow.com/m/2024-09-12-openai-o1-preview/): A model trained to think before answering opened a second scaling axis: spend more compute at inference and accuracy rises.
- [1 August 2024 — The EU AI Act enters into force](https://itdoeswhatnow.com/m/2024-08-01-eu-ai-act-in-force/): The world's first comprehensive AI law took effect, banning some uses outright and imposing obligations on general-purpose models by capability.
- [13 May 2024 — OpenAI launches GPT-4o with real-time voice](https://itdoeswhatnow.com/m/2024-05-13-gpt-4o/): A single model handling text, vision and audio end to end, with conversational latency — and a voice that led to a public dispute with Scarlett Johansson.
- [8 December 2023 — EU negotiators strike a political deal on the AI Act](https://itdoeswhatnow.com/m/2023-12-08-eu-negotiators-strike-a-political-deal-on-the/): Foundation models and police use of biometric surveillance were the last sticking points; the deal set tiered duties for both, with the text finalised in 2024.
- [17 November 2023 — OpenAI's board fires Sam Altman, and reinstates him five days later](https://itdoeswhatnow.com/m/2023-11-17-altman-fired/): The board cited a loss of confidence but gave no detail; around 700 of roughly 770 employees threatened to resign, and the board itself was replaced.
- [30 October 2023 — Biden signs the executive order on safe and trustworthy AI](https://itdoeswhatnow.com/m/2023-10-30-biden-executive-order/): The most far-reaching US action on AI to date: compute thresholds, mandatory safety reporting, and a new safety institute.
- [22 March 2023 — The "Pause Giant AI Experiments" open letter](https://itdoeswhatnow.com/m/2023-03-22-pause-giant-ai-experiments/): Thirty thousand signatories called for a six-month halt on training systems more powerful than GPT-4. No lab paused.
- [14 March 2023 — OpenAI releases GPT-4](https://itdoeswhatnow.com/m/2023-03-14-gpt-4-release/): A multimodal model that passed professional exams near the top of the human range — and whose technical report disclosed no architecture, data or compute.
- [30 November 2022 — OpenAI launches ChatGPT](https://itdoeswhatnow.com/m/2022-11-30-chatgpt-launch/): A free web interface to an existing GPT-3.5 model, reported by one analysis to have reached 100 million monthly users within two months.
- [7 October 2022 — The US restricts advanced chip exports to China](https://itdoeswhatnow.com/m/2022-10-07-the-us-restricts-advanced-chip-exports-to-china/): The rules covered the tools to make advanced chips, not only the chips, and barred US citizens from supporting Chinese chipmaking.
- [22 August 2022 — Stable Diffusion is released to the public](https://itdoeswhatnow.com/m/2022-08-22-stable-diffusion-release/): Weights published under a permissive licence and runnable on a consumer graphics card, with community fine-tunes and graphical front-ends appearing within weeks.
- [30 November 2020 — AlphaFold 2 solves protein structure prediction at CASP14](https://itdoeswhatnow.com/m/2020-11-30-alphafold-2-casp14/): DeepMind's system predicted protein structures to roughly experimental accuracy, ending a fifty-year-old open problem in biology.
- [28 May 2020 — OpenAI publishes the GPT-3 paper](https://itdoeswhatnow.com/m/2020-05-28-gpt-3-paper/): A 175-billion-parameter language model that learned new tasks from examples in its prompt, without any weight updates.

## Feeds and indexes
- RSS: https://itdoeswhatnow.com/rss.xml
- Sitemap: https://itdoeswhatnow.com/sitemap.xml