Index of Print Codes
Every printed entry, with the master IDs in the historical edition that it draws on. 354 entries.
| Code | Date | Entry | Historical edition |
|---|---|---|---|
| January | |||
| P-2025-W01-01 | Reported January 1 | OpenAI misses its own date for a creator opt-out tool | 2025-W01-L1-4, 2025-W01-L2-4 |
| P-2025-W01-02 | January 1: instructions reported | Russia seeks closer AI cooperation with China | 2025-W01-L1-5, 2025-W01-L2-5 |
| P-2025-W01-03 | January 3 | Microsoft plans $80 billion for AI data centers | 2025-W01-L1-1, 2025-W01-L2-1, 2025-W01-L4-1 |
| P-2025-W01-04 | January 3 | Meta removes AI character profiles | 2025-W01-L1-3, 2025-W01-L2-3 |
| P-2025-W01-05 | January 5, Pacific time | Altman sets out expectations for AI agents | 2025-W01-L1-2, 2025-W01-L2-2, 2025-W01-L3-1 |
| P-2025-W02-01 | January 6 | Nvidia brings AI hardware and world models to CES | 2025-W02-L1-1, 2025-W02-L2-1, 2025-W02-L3-1, 2025-W02-L4-1 |
| P-2025-W02-02 | Reported January 7 | Anthropic is reported to seek $2 billion at a $60 billion valuation | 2025-W02-L1-2, 2025-W02-L2-3 |
| P-2025-W02-03 | January 8 | Microsoft releases Phi-4 weights | 2025-W02-L1-3, 2025-W02-L2-2 |
| P-2025-W02-04 | January 9 | Grok reaches US users through a standalone app | 2025-W02-L1-4, 2025-W02-L2-4 |
| P-2025-W02-05 | January 10 | Salesforce announces retail agents | 2025-W02-L2-5 |
| P-2025-W02-06 | January 10 | Zuckerberg predicts AI coding agents at the level of mid-level engineers | 2025-W02-L1-5 |
| P-2025-W03-01 | January 13; published January 15 | US rule sets new controls on AI diffusion | 2025-W03-L1-1, 2025-W03-L2-1, 2025-W03-L3-1, 2025-W03-L4-1 |
| P-2025-W03-02 | January 13 | UK adopts an AI action plan | 2025-W03-L1-3, 2025-W03-L2-2 |
| P-2025-W03-03 | January 14 | ChatGPT adds scheduled tasks | 2025-W03-L1-4, 2025-W03-L2-3 |
| P-2025-W03-04 | January 14 | Biden directs federal support for AI infrastructure | 2025-W03-L1-2, 2025-W03-L2-5 |
| P-2025-W03-05 | January 15 | MiniMax releases long-context models | 2025-W03-L1-5, 2025-W03-L2-4 |
| P-2025-W03-06 | January 16 | Apple pauses AI summaries of news alerts | 2025-W03-L3-2 |
| P-2025-W04-01 | January 20 | DeepSeek releases R1 reasoning models | 2025-W04-L1-1, 2025-W04-L2-1, 2025-W04-L3-1, 2025-W04-L4-1 |
| P-2025-W04-02 | January 21 | Stargate announces a large US infrastructure plan | 2025-W04-L1-2, 2025-W04-L2-2, 2025-W04-L3-2, 2025-W04-L4-2 |
| P-2025-W04-03 | January 21 | Google updates Gemini Flash Thinking | 2025-W04-L4-3 |
| P-2025-W04-04 | January 23 | Operator begins a browser-agent preview | 2025-W04-L1-4, 2025-W04-L2-3, 2025-W04-L3-4, 2025-W04-L4-4 |
| P-2025-W04-05 | January 23 | Trump orders a new AI action plan | 2025-W04-L1-3, 2025-W04-L2-4, 2025-W04-L3-3 |
| P-2025-W04-06 | January 24 | Meta raises its planned capital spending | 2025-W04-L1-5, 2025-W04-L2-5 |
| P-2025-W05-01 | January 27 | DeepSeek concerns trigger a sharp Nvidia selloff | 2025-W05-L1-1, 2025-W05-L2-1, 2025-W05-L3-1, 2025-W05-L4-1 |
| P-2025-W05-02 | January 27 | Janus-Pro combines image understanding and generation | 2025-W05-L4-2 |
| P-2025-W05-03 | January 28 | Qwen2.5-Max enters hosted competition | 2025-W05-L1-3, 2025-W05-L2-2 |
| P-2025-W05-04 | Reported January 30 | Microsoft makes DeepSeek R1 available on its cloud | 2025-W05-L3-2 |
| P-2025-W05-05 | January 30 | Ai2 scales its open post-training work | 2025-W05-L4-3 |
| P-2025-W05-06 | January 30 | Italy restricts DeepSeek’s processing of personal data | 2025-W05-L1-2 |
| P-2025-W05-07 | January 31 | NASA blocks DeepSeek on agency systems | 2025-W05-L3-3 |
| P-2025-W05-08 | January 31 | OpenAI releases o3-mini | 2025-W05-L1-4, 2025-W05-L2-3, 2025-W05-L3-4, 2025-W05-L4-4 |
| P-2025-W05-09 | February 2 | Deep research produces sourced reports | 2025-W05-L2-4 |
| P-2025-W05-10 | February 2 | The first AI Act obligations begin to apply | 2025-W05-L1-5, 2025-W05-L2-5 |
| February | |||
| P-2025-W06-01 | February 3 | Constitutional Classifiers face adversarial testing | 2025-W06-L1-4, 2025-W06-L2-3 |
| P-2025-W06-02 | February 3 revision; first posted January 31 | s1 explores reasoning with a small training set | 2025-W06-L1-5, 2025-W06-L2-4 |
| P-2025-W06-03 | February 4 | Google revises its AI principles | 2025-W06-L1-2, 2025-W06-L2-2, 2025-W06-L4-1 |
| P-2025-W06-04 | February 4 | Kakao and OpenAI announce a partnership | 2025-W06-L3-1 |
| P-2025-W06-05 | February 4 | Figure ends its collaboration with OpenAI | 2025-W06-L3-2 |
| P-2025-W06-06 | February 5 | Gemini 2.0 expands for developers | 2025-W06-L1-1, 2025-W06-L2-1, 2025-W06-L4-2 |
| P-2025-W06-07 | February 6 | Le Chat reaches mobile users | 2025-W06-L3-3 |
| P-2025-W06-08 | February 6–7 | Science opens the Paris summit program | 2025-W06-L3-4 |
| P-2025-W06-09 | February 9 | France announces €109 billion in prospective investment | 2025-W06-L1-3, 2025-W06-L2-5 |
| P-2025-W07-01 | February 10 | France links AI investment to electricity supply | 2025-W07-L1-2 |
| P-2025-W07-02 | February 10 | European companies launch AI Champions | 2025-W07-L2-2 |
| P-2025-W07-03 | Reported February 10 | OpenAI is reported to be finishing its first chip design | 2025-W07-L2-1 |
| P-2025-W07-04 | February 11 | InvestAI targets €200 billion in investment | 2025-W07-L1-3 |
| P-2025-W07-05 | February 11 | Vance argues for lighter AI regulation | 2025-W07-L1-4 |
| P-2025-W07-06 | February 12 | Firefly expands video generation | 2025-W07-L1-5 |
| P-2025-W07-07 | February 13 | Veo 2 adds generated clips to YouTube Shorts | 2025-W07-L1-5 |
| P-2025-W07-08 | February 13 | Publishers bring a case against Cohere | 2025-W07-L2-3 |
| P-2025-W07-09 | February 14 | OpenAI rejects the Musk-led offer | 2025-W07-L1-1 |
| P-2025-W08-01 | February 17, Pacific time | xAI introduces Grok 3 | 2025-W08-L4-1 |
| P-2025-W08-02 | Announced February 17; effective February 15 | DeepSeek pauses new downloads in South Korea | 2025-W08-L4-3 |
| P-2025-W08-03 | February 18 | Multi-Novelty tests more varied model outputs | 2025-W08-L4-4 |
| March | |||
| P-2025-W09-01 | February 24 | Claude 3.7 Sonnet offers two modes of response | 2025-W09-L2-2 |
| P-2025-W09-02 | February 24–28 | DeepSeek shares its infrastructure tools | 2025-W09-L1-1, 2025-W09-L1-2, 2025-W09-L1-3, 2025-W09-L1-4, 2025-W09-L1-5 |
| P-2025-W09-03 | February 25 | Wan2.1 opens video-generation code and weights | 2025-W09-L2-3 |
| P-2025-W09-04 | February 26 | Amazon announces Alexa+ | 2025-W09-L2-4 |
| P-2025-W09-05 | February 27 | GPT-4.5 enters research preview | 2025-W09-L2-1 |
| P-2025-W09-06 | Reported February 27 | Perplexity refreshes voice search on iPhone | 2025-W09-L2-5 |
| P-2025-W10-01 | March 4 | Moddy applies AI to changes across code repositories | 2025-W10-L2-2, 2025-W10-L3-4 |
| P-2025-W10-02 | March 5 | Agentforce 2dx extends agents into workflows | 2025-W10-L2-2, 2025-W10-L3-2 |
| P-2025-W10-03 | Reported March 6 | Manus previews an agent for multi-step tasks | 2025-W10-L1-1 |
| P-2025-W11-01 | March 12 | Research shows DiLoCo training scaling predictably | 2025-W11-L1-2 |
| P-2025-W11-02 | March 13 | OpenAI submits proposals for US AI policy | 2025-W11-L1-1 |
| P-2025-W11-03 | March 16 | Baidu releases ERNIE 4.5 and ERNIE X1 | 2025-W11-L1-3, 2025-W11-L2-2 |
| P-2025-W12-01 | March 17 | An experiment pits ChatGPT against federal judges | 2025-W12-L1-2 |
| P-2025-W12-02 | March 18 | Cosmos expands its physical-AI tools | 2025-W12-L1-HM2 |
| P-2025-W12-03 | March 18 | Skywork R1V extends reasoning to visual inputs | 2025-W12-L1-HM4 |
| P-2025-W12-04 | March 19 | METR measures the length of tasks agents can complete | 2025-W12-L1-HM3 |
| P-2025-W12-05 | March 19 | SoftBank agrees to acquire Ampere | 2025-W12-L1-1 |
| P-2025-W12-06 | March 20 | AlexNet’s original source code becomes available | 2025-W12-L1-HM1 |
| P-2025-W12-07 | March 20 | OpenAI adds transcription and speech-generation models | 2025-W12-L1-6 |
| P-2025-W13-01 | March 24–25 | DeepSeek updates V3 | 2025-W13-L1-5 |
| P-2025-W13-02 | March 24 | ARC-AGI-2 raises the reasoning challenge | 2025-W13-L1-4 |
| P-2025-W13-03 | March 25 | Google introduces Gemini 2.5 Pro Experimental | 2025-W13-L1-2 |
| P-2025-W13-04 | March 25 | GPT-4o adds native image generation | 2025-W13-L1-3 |
| P-2025-W13-05 | March 28 | Musk announces xAI’s acquisition of X | 2025-W13-L1-1 |
| April | |||
| P-2025-W14-01 | March 31 | Isomorphic Labs raises $600 million | 2025-W14-L1-2 |
| P-2025-W14-02 | March 31 | Amazon previews browser actions with Nova Act | 2025-W14-L1-3 |
| P-2025-W14-03 | March 31 | SoftBank agrees to lead new OpenAI funding | 2025-W14-L3-1, 2025-W15-L7-3 |
| P-2025-W14-04 | April 1 | Meta’s head of AI research announces her departure | 2025-W14-L1-1 |
| P-2025-W14-05 | April 2 | DeepMind sets out its approach to AGI safety | 2025-W14-L2-2, 2025-W15-L1-5 |
| P-2025-W14-06 | April 3 | AI 2027 publishes a detailed future scenario | 2025-W14-L3-2, 2025-W17-L2-5 |
| P-2025-W15-01 | Background: April 3; selected April 13 at the editor’s request | Research focus: MegaMath opens a mathematics corpus | 2025-W15-L5-1, 2025-W15-L5-2, 2025-W15-L5-3, 2025-W15-L5-4, 2025-W15-L5-5 |
| P-2025-W15-02 | April 7 | Stanford’s AI Index takes stock of the field | 2025-W15-L3-1, 2025-W15-L3-2, 2025-W15-L3-3, 2025-W15-L3-4, 2025-W15-L3-5, 2025-W15-L4-1, 2025-W15-L4-2, 2025-W15-L4-3, 2025-W15-L6-2 |
| P-2025-W15-03 | April 9 | Europe publishes its AI Continent Action Plan | 2025-W15-L6-1 |
| P-2025-W15-04 | April 9 | Cloud Next links computing capacity and agent tools | 2025-W15-L7-2 |
| P-2025-W15-05 | April 9 | OpenAI countersues Elon Musk | 2025-W15-L1-4 |
| P-2025-W15-06 | April 9 | xAI opens Grok 3 to developers | 2025-W15-L8-4 |
| P-2025-W15-07 | April 10 | OpenAI announces GPT-4’s retirement from ChatGPT | 2025-W15-L8-1 |
| P-2025-W15-08 | April 11 | Ireland’s privacy regulator opens an inquiry into Grok’s training data | 2025-W15-L1-3 |
| P-2025-W16-01 | April 14 | GPT-4.1 brings longer context to the API | 2025-W16-L1-1, 2025-W16-L4-1 |
| P-2025-W16-02 | April 15 | Veo 2 reaches Gemini Advanced and Whisk | 2025-W16-L1-2 |
| P-2025-W16-03 | April 15 | OpenAI revises its Preparedness Framework | 2025-W16-L4-3 |
| P-2025-W16-04 | April 16 | o3 and o4-mini combine reasoning with tools | 2025-W16-L6-1 |
| P-2025-W16-05 | April 16 | Wikimedia shares structured Wikipedia data on Kaggle | 2025-W16-L6-4 |
| P-2025-W16-06 | April 17 | Gemini 2.5 Flash makes thinking adjustable | 2025-W16-L6-2 |
| P-2025-W16-07 | April 20 | Hassabis sets out DeepMind’s ambitions on 60 Minutes | 2025-W16-L6-3, 2025-W16-L4-4 |
| P-2025-W17-01 | April 21 | Databricks expands conversational data analysis | 2025-W17-L5-2 |
| P-2025-W17-02 | April 23 | OpenAI opens its image model to developers | 2025-W17-L3-2, 2025-W17-L4-2, 2025-W17-L5-3 |
| P-2025-W17-03 | April 23 | US executive order establishes an AI education program | 2025-W17-L3-3, 2025-W17-L4-3 |
| May | |||
| P-2025-W18-01 | April 29 | A study questions Chatbot Arena’s ranking practices | 2025-W18-L2-6 |
| P-2025-W18-02 | April 29 | Meta launches a standalone AI app | 2025-W18-L3-1 |
| P-2025-W18-03 | April 29 | Meta says Llama has been downloaded 1.2 billion times | 2025-W18-L3-4 |
| P-2025-W18-04 | April 29; explanation May 2 | OpenAI rolls back an overly agreeable GPT-4o update | 2025-W18-L3-2 |
| P-2025-W18-05 | April 29 | Waymo and Toyota explore autonomous-driving collaboration | 2025-W18-L3-3 |
| P-2025-W18-06 | April 30 | Microsoft expands the Phi family with reasoning models | 2025-W18-L5-2 |
| P-2025-W18-07 | April 30 | Amazon releases Nova Premier | 2025-W18-L3-5 |
| P-2025-W18-08 | May 2 | An AI-generated image of Trump as pope draws criticism | 2025-W18-L4-1 |
| P-2025-W19-01 | May 5 | OpenAI keeps nonprofit control in its restructuring plan | 2025-W19-L3-5 |
| P-2025-W19-02 | Reported May 6 | OpenAI is reported to agree to buy Windsurf | 2025-W19-L2-3 |
| P-2025-W19-03 | May 6 | IBM announces tools for coordinating enterprise agents | 2025-W19-L2-5 |
| P-2025-W19-04 | May 7 | OpenAI proposes national AI partnerships | 2025-W19-L1-1 |
| P-2025-W19-05 | Plan reported May 7; rescission and non-enforcement announced May 13 | The administration moves to rescind the AI diffusion rule | 2025-W19-L2-2 |
| P-2025-W19-06 | May 7 | Amazon introduces touch-sensitive warehouse robotics | 2025-W19-L3-1 |
| P-2025-W19-07 | May 8 | FDA sets a timetable for internal AI deployment | 2025-W19-L1-2, 2025-W19-L2-1 |
| P-2025-W19-08 | May 10 | Pope Leo XIV links his name to the age of AI | 2025-W19-L1-3 |
| P-2025-W19-09 | Reported May 11 | OpenAI and Microsoft are reported to renegotiate their partnership | 2025-W19-L2-4 |
| P-2025-W20-01 | May 13–15 | US AI deals are announced across the Gulf | 2025-W20-L1-1, 2025-W21-L3-3, 2025-W21-L4-3 |
| P-2025-W20-02 | May 14 | AlphaEvolve searches for better algorithms | 2025-W20-L2-1, 2025-W21-L1-4 |
| P-2025-W20-03 | May 14–16 | xAI responds to off-topic political claims from Grok | 2025-W20-L1-2, 2025-W21-L3-2, 2025-W21-L4-2 |
| P-2025-W20-04 | May 16 | State attorneys general oppose a proposed AI-law moratorium | 2025-W20-L1-3 |
| P-2025-W21-01 | May 20 | Google expands AI Mode in Search | 2025-W21-L1-1 |
| P-2025-W21-02 | May 20 | Veo 3 generates video with sound | 2025-W21-L1-1, 2025-W21-L2-1 |
| P-2025-W21-03 | May 20 | Flow brings generative models into a filmmaking workspace | 2025-W21-L2-4 |
| P-2025-W21-04 | May 21 | Huang calls US chip export controls a failure | 2025-W21-L1-3 |
| P-2025-W21-05 | May 22 | Anthropic releases Claude Opus 4 and Sonnet 4 | 2025-W21-L2-3 |
| June | |||
| P-2025-W22-01 | Reported May 26 | Meta loses most of Llama’s original authors | 2025-W22-L3-5 |
| P-2025-W22-02 | May 27 | Salesforce agrees to acquire Informatica | 2025-W22-L2-1 |
| P-2025-W22-03 | May 28 | Amodei warns about entry-level employment | 2025-W22-L3-1, 2025-W23-L2-2 |
| P-2025-W22-04 | May 29 | The New York Times licenses content to Amazon | 2025-W22-L3-3, 2025-W23-L2-4 |
| P-2025-W22-05 | May 30 | AI dominates closing arguments in Google’s search case | 2025-W22-L1-3 |
| P-2025-W23-01 | Background: May 30; selected June 8 at the editor’s request | Research focus: BOND surveys the AI economy | 2025-W23-L3-1 |
| P-2025-W23-02 | June 2 | FDA launches Elsa for internal review work | 2025-W23-L1-3 |
| P-2025-W23-03 | June 3 | Bengio launches LawZero | 2025-W23-L4-1 |
| P-2025-W23-04 | June 4 | Mistral previews an enterprise coding assistant | 2025-W23-L1-2 |
| P-2025-W23-05 | June 4 | Markey challenges the proposed state AI-law moratorium | 2025-W23-L1-4 |
| P-2025-W24-01 | June 9 | Apple opens its on-device model to developers | 2025-W24-L1-3 |
| P-2025-W24-02 | Reported June 9 | OpenAI reports a $10 billion annualized revenue run rate | 2025-W24-L2-5 |
| P-2025-W24-03 | June 10 | Datadog introduces agents for software operations | 2025-W24-L2-3 |
| P-2025-W24-04 | June 11 | Google names a chief AI architect | 2025-W24-L2-2 |
| P-2025-W24-05 | June 11 | V-JEPA 2 learns predictive representations from video | 2025-W24-L2-HM1 |
| P-2025-W24-06 | June 12 | Meta invests in Scale AI as Wang joins its AI effort | 2025-W24-L2-1 |
| P-2025-W25-01 | June 16 | WhatsApp introduces ads in its Updates tab | 2025-W25-L2-4 |
| P-2025-W25-02 | June 17 | Jassy expects AI to reduce Amazon’s corporate workforce | 2025-W25-L2-3 |
| P-2025-W25-03 | June 18 | Midjourney releases its first video model | 2025-W25-L2-5 |
| P-2025-W26-01 | June 23 | Salesforce announces monitoring and integration tools for agents | 2025-W26-L4-5 |
| P-2025-W26-02 | June 23 and 25 | Two courts rule on fair use in generative-AI training | 2025-W26-L4-1 |
| P-2025-W26-03 | June 25 | US senators propose restrictions on foreign-adversary AI | 2025-W26-L1-2 |
| P-2025-W26-04 | June 25 | Nvidia regains the title of most valuable company | 2025-W26-L4-3 |
| P-2025-W26-05 | June 26 | Qwen previews image generation and editing in VLo | 2025-W26-L1-4 |
| P-2025-W26-06 | Reported June 27 | Microsoft’s next AI chip is reported delayed | 2025-W26-L1-1 |
| P-2025-W26-07 | June 27 | Project Vend tests Claude as a shopkeeper | 2025-W26-L1-5 |
| July | |||
| P-2025-W27-01 | June 30 | Meta brings its AI teams into Superintelligence Labs | 2025-W27-L1-2, 2025-W27-L4-5 |
| P-2025-W27-02 | June 30 | Microsoft tests an orchestrator for difficult diagnoses | 2025-W27-L4-2 |
| P-2025-W27-03 | July 1 | Cloudflare tests paid access for AI crawlers | 2025-W27-L4-3 |
| P-2025-W27-04 | July 1 | TreeQuest coordinates search across several models | 2025-W27-L4-4 |
| P-2025-W27-05 | July 3 | Ilya Sutskever becomes SSI’s chief executive | 2025-W27-L1-1 |
| P-2025-W27-06 | Reported July 6 | Isomorphic Labs prepares for human trials | 2025-W27-L1-3 |
| P-2025-W28-01 | Reported July 7 | Meta hires Apple’s head of foundation models | 2025-W28-L1-3 |
| P-2025-W28-02 | Reported July 8 | An AI-cloned voice impersonates the US secretary of state | 2025-W28-L3-4 |
| P-2025-W28-03 | July 9 | xAI releases Grok 4 with built-in tool use | 2025-W28-L1-2, 2025-W28-L2-3 |
| P-2025-W28-04 | July 10 | EU publishes the general-purpose AI Code of Practice | 2025-W28-L1-4 |
| P-2025-W28-05 | July 11 | Google hires Windsurf leaders in a licensing deal | 2025-W28-L1-1 |
| P-2025-W28-06 | July 11 | Moonshot releases Kimi K2’s model weights | 2025-W28-L2-2, 2025-W29-L3-3 |
| P-2025-W28-07 | July 11 | OpenAI postpones its open-weight model | 2025-W28-L2-1 |
| P-2025-W29-01 | July 14 | Meta outlines large computing-cluster plans | 2025-W29-L1-3, 2025-W29-L3-2 |
| P-2025-W29-02 | July 15 | Thinking Machines raises $2 billion | 2025-W29-L1-5 |
| P-2025-W29-03 | July 15 | Pennsylvania summit brings together AI and energy investment | 2025-W29-L4-3 |
| P-2025-W29-04 | July 16 | Google adds Gemini 2.5 Pro and Deep Search to AI Mode | 2025-W29-L4-4 |
| P-2025-W29-05 | July 17 | ChatGPT agent combines research with computer actions | 2025-W29-L1-1, 2025-W29-L3-1 |
| P-2025-W29-06 | Reported July 17 | Perplexity is valued at $18 billion | 2025-W29-L1-4 |
| P-2025-W29-07 | July 19 | OpenAI reports gold-level performance on IMO problems | 2025-W29-L4-1 |
| P-2025-W30-01 | July 21 | Gemini’s IMO solutions receive official assessment | 2025-W30-L2-1 |
| P-2025-W30-02 | July 23 | White House releases its AI Action Plan | 2025-W30-L1-1, 2025-W29-L1-2 |
| P-2025-W30-03 | July 23 | Google brings local AI processing and developer tools to India | 2025-W30-L1-3 |
| P-2025-W30-04 | July 24 | Nadella addresses layoffs amid record AI spending | 2025-W30-L2-4 |
| P-2025-W30-05 | July 24 | ASI-Arch explores automated neural-architecture research | 2025-W30-L2-HM1 |
| P-2025-W30-06 | July 26 | World AI Conference opens in Shanghai | 2025-W30-L1-2, 2025-W30-L2-2, 2025-W30-L2-3 |
| August | |||
| P-2025-W31-01 | July 28 | GLM-4.5 joins the open-weight model field | 2025-W31-L2-5, 2025-W32-L4-5 |
| P-2025-W31-02 | July 29 | ChatGPT adds a guided study mode | 2025-W31-L2-2 |
| P-2025-W31-03 | July 30 | Meta and Microsoft raise their AI spending | 2025-W31-L1-2, 2025-W31-L5-1, 2025-W31-L4-4 |
| P-2025-W31-04 | Reported July 30 | OpenAI’s annualized revenue reportedly reaches $12 billion | 2025-W31-L1-4, 2025-W31-L5-2 |
| P-2025-W31-05 | July 31; published August 26 | China approves its ‘AI Plus’ guidelines | 2025-W31-L1-3, 2025-W35-L1-5 |
| P-2025-W31-06 | July 31 | Stargate Norway sets out European computing plans | 2025-W31-L1-1, 2025-W31-L5-5 |
| P-2025-W31-07 | Reported August 1 | Anthropic cuts off OpenAI’s access to Claude | 2025-W31-L3-5 |
| P-2025-W31-08 | August 2 | EU general-purpose AI obligations begin to apply | 2025-W31-L2-1, 2025-W31-L3-3, 2025-W32-L1-2, 2025-W32-L2-2 |
| P-2025-W32-01 | August 5 | Grok Imagine renews the argument over guardrails | 2025-W32-L2-5 |
| P-2025-W32-02 | August 5 | OpenAI releases gpt-oss model weights | 2025-W32-L3-4 |
| P-2025-W32-03 | August 5 | Anthropic releases Claude Opus 4.1 | 2025-W32-L3-2 |
| P-2025-W32-04 | August 5 | Genie 3 turns prompts into interactive environments | 2025-W32-L3-3, 2025-W34-L3-2 |
| P-2025-W32-05 | August 6 | Jules leaves public beta | 2025-W32-L4-2 |
| P-2025-W32-06 | August 6 | Federal agencies gain $1 ChatGPT access | 2025-W32-L1-4 |
| P-2025-W32-07 | August 7 | GPT-5 combines fast responses with deeper reasoning | 2025-W32-L1-1, 2025-W32-L2-1, 2025-W32-L3-1, 2025-W32-L4-1 |
| P-2025-W32-08 | August 7 | Research focus: MiroMind opens a research-agent stack | 2025-W32-L5-1 |
| P-2025-W33-01 | August 11 | Nvidia and AMD agree to share China chip revenue with the US | 2025-W33-L1-2 |
| P-2025-W33-02 | August 11 | Cosmos Reason connects vision with physical reasoning | 2025-W33-L4-2 |
| P-2025-W33-03 | August 12 | Anthropic offers Claude to all three branches of government for $1 | 2025-W33-L2-5 |
| P-2025-W33-04 | August 12 | Claude Sonnet 4 offers a million-token context window | 2025-W33-L2-5, 2025-W34-L3-3 |
| P-2025-W33-05 | August 14 | Reporting exposes Meta’s chatbot rules for children | 2025-W33-L1-3 |
| P-2025-W33-06 | August 14 | Cohere raises $500 million for enterprise AI | 2025-W33-L2-1 |
| P-2025-W33-07 | August 14 | DINOv3 learns reusable features from images | 2025-W33-L2-2 |
| P-2025-W33-08 | August 14 | NSF and Nvidia back open models for science | 2025-W33-L1-4 |
| P-2025-W34-01 | August 18 | Texas investigates Meta and Character.AI over mental-health claims | 2025-W34-L4-5 |
| P-2025-W34-02 | Background: July 28; selected August 24 at the editor’s request | Research focus: what can an agent change? | 2025-W34-L1-HM1 |
| P-2025-W34-03 | Background: August 10; selected August 24 at the editor’s request | Research focus: learning from agent interactions | 2025-W34-L1-HM2 |
| P-2025-W34-04 | August 19 | Acrobat Studio combines documents, AI and design | 2025-W34-L4-2 |
| P-2025-W34-05 | Reported August 20 | Meta pauses hiring for its AI division | 2025-W34-L1-1 |
| P-2025-W34-06 | August 21 | DeepSeek-V3.1 combines thinking and non-thinking modes | 2025-W34-L3-1, 2025-W34-L5-5 |
| P-2025-W34-07 | August 21 | Google expands AI Mode to more countries | 2025-W34-L4-3 |
| P-2025-W34-08 | Reported August 22 | Meta signs a cloud deal with Google | 2025-W34-L2-2 |
| P-2025-W34-09 | August 22 | A specialized GPT model proposes reprogramming proteins | 2025-W34-L4-1 |
| P-2025-W34-10 | August 22 | Meta licenses Midjourney’s image technology | 2025-W34-L4-4 |
| P-2025-W35-01 | August 25 | Attorneys general press AI companies on child safety | 2025-W35-L1-4, 2025-W35-L5-1 |
| P-2025-W35-02 | August 25 | xAI sues Apple and OpenAI | 2025-W35-L7-2 |
| P-2025-W35-03 | August 26 | Nano Banana adds targeted image editing | 2025-W35-L9-3 |
| P-2025-W35-04 | August 27 | Anthropic forms a public-sector advisory council | 2025-W35-L1-2, 2025-W35-L3-1 |
| P-2025-W35-05 | August 27 | Nvidia’s results meet a market braced for more | 2025-W35-L7-1 |
| P-2025-W35-06 | August 28 | Microsoft previews two in-house models | 2025-W35-L1-1, 2025-W35-L2-1, 2025-W35-L9-2 |
| September | |||
| P-2025-W36-01 | From early September; reported September 8 | Hunger strikes target Anthropic and Google DeepMind | 2025-W36-L4-1 |
| P-2025-W36-02 | September 2 | Apertus opens a multilingual model for inspection | 2025-W36-L1-1 |
| P-2025-W36-03 | Updated September 3; first reported August 26 | PromptLock shows how language models could enter malware | 2025-W36-L1-5 |
| P-2025-W36-04 | September 4 | OpenAI plans a jobs platform and AI certifications | 2025-W36-L3-4 |
| P-2025-W36-05 | September 4 | A White House task force advances an AI education challenge | 2025-W36-L3-2 |
| P-2025-W36-06 | September 4 | Warner Bros. Discovery sues Midjourney | 2025-W36-L3-3 |
| P-2025-W36-07 | Reported September 5 | OpenAI is reported to be building a chip with Broadcom | 2025-W36-L4-5 |
| P-2025-W36-08 | September 5 | Research examines why models guess instead of abstaining | 2025-W36-L2-1 |
| P-2025-W36-09 | September 5; hearing September 8 | Anthropic agrees to settle authors’ piracy claims | 2025-W36-L3-1, 2025-W37-L2-2 |
| P-2025-W36-10 | Reported September 7 | OpenAI backs an animated feature made with its tools | 2025-W36-L4-4 |
| P-2025-W37-01 | September 8 | Google’s AI Mode adds five languages | 2025-W37-L3-5 |
| P-2025-W37-02 | September 8 amendments; passed September 13 | California passes SB 53, a frontier-AI disclosure law | 2025-W37-L2-1, 2025-W38-L3-2 |
| P-2025-W37-03 | September 11 | FTC opens an inquiry into companion chatbots | 2025-W37-L1-1 |
| P-2025-W37-04 | September 12 | Hassabis calls learning how to learn the key skill | 2025-W37-L2-3 |
| P-2025-W38-01 | September 16 | YouTube brings generated video and sound to Shorts | 2025-W38-L1-5 |
| P-2025-W38-02 | September 17 | DeepSeek-R1 research appears in Nature | 2025-W38-L1-1 |
| P-2025-W38-03 | Reported September 17; contest held September 4 | Gemini reaches gold-medal level in programming | 2025-W38-L1-4, 2025-W38-L3-1 |
| P-2025-W38-04 | September 17 | Meta announces display glasses with wristband control | 2025-W38-L1-3 |
| P-2025-W38-05 | September 17 | Delphi models long-term disease risk | 2025-W38-L3-5 |
| P-2025-W38-06 | Reported September 18; findings first published September 9 | Music publishers accuse AI companies of copyright theft | 2025-W38-L3-4 |
| P-2025-W38-07 | September 18 | Gemini expands its role inside Chrome | 2025-W38-L1-2 |
| P-2025-W38-08 | September 18 | Nvidia and Intel plan joint computing products | 2025-W38-L3-3 |
| P-2025-W39-01 | September 22 | Nvidia and OpenAI outline a large infrastructure partnership | 2025-W39-L5-1 |
| P-2025-W39-02 | September 22 | TII and Nvidia establish an AI robotics research lab | 2025-W39-L1-2 |
| P-2025-W39-03 | September 22 | GSA and Meta arrange federal access to Llama | 2025-W39-L1-3 |
| P-2025-W39-04 | September 22 | AppZen raises $180 million for finance automation | 2025-W39-L7-2 |
| P-2025-W39-05 | September 22 | DeepSeek updates V3.1 with Terminus | 2025-W39-L8-1 |
| P-2025-W39-06 | September 23 | Stargate announces five additional US sites | 2025-W39-L1-1 |
| P-2025-W39-07 | Reported September 26 | Altman puts a figure on AI’s share of economic tasks | 2025-W39-L7-1 |
| October | |||
| P-2025-W40-01 | September 29 | Claude Sonnet 4.5 arrives with new development tools | 2025-W40-L2-1, 2025-W40-L4-1, 2025-W40-L13-2 |
| P-2025-W40-02 | September 29 | California signs frontier AI transparency legislation | 2025-W40-L4-2 |
| P-2025-W40-03 | September 29 | ChatGPT adds checkout for selected purchases | 2025-W40-L11-2, 2025-W40-L13-3 |
| P-2025-W40-04 | September 29 | DeepSeek tests sparse attention for longer contexts | 2025-W40-L15-1 |
| P-2025-W40-05 | September 30 | Sora 2 combines generated video with sound | 2025-W40-L2-2, 2025-W40-L11-1, 2025-W40-L13-1 |
| P-2025-W40-06 | September 30 | Dragon Hatchling explores an alternative model architecture | 2025-W40-L11-4 |
| P-2025-W40-07 | October 1 | Samsung and SK join Stargate | 2025-W40-L11-3 |
| P-2025-W40-08 | October 1 | Stellantis and Mistral move from pilots to company-wide AI | 2025-W40-L2-4 |
| P-2025-W40-09 | October 2 | Perplexity makes its Comet browser free | 2025-W40-L13-5 |
| P-2025-W40-10 | October 3 | AI helps explain a new antibiotic for gut disease | 2025-W40-L11-5 |
| P-2025-W41-01 | Background: February 19; selected October 12 under the editor’s theme | Research focus: a co-scientist proposes testable hypotheses | 2025-W41-L3-1 |
| P-2025-W41-02 | Background: May 19; selected October 12 under the editor’s theme | Research focus: Robin links hypotheses with laboratory data | 2025-W41-L3-2 |
| P-2025-W41-03 | Background: May 24; selected October 12 under the editor’s theme | Research focus: AI-Researcher automates a computational workflow | 2025-W41-L3-3 |
| P-2025-W41-04 | October 6 | AMD and OpenAI agree on a six-gigawatt deployment program | 2025-W41-L6-3 |
| P-2025-W41-05 | October 6 | AgentKit brings workflow building and evaluation together | 2025-W41-L6-5 |
| P-2025-W41-06 | October 7 | IBM and Anthropic partner on enterprise development tools | 2025-W41-L6-2 |
| P-2025-W41-07 | October 8 | Europe sets strategies for AI adoption and scientific research | 2025-W41-L6-1 |
| P-2025-W41-08 | Scheduled for October 22; selected October 12 | Agents4Science creates a venue for AI-authored research | 2025-W41-L3-5 |
| P-2025-W42-01 | October 13 | OpenAI and Broadcom outline custom accelerator systems | 2025-W42-L1-1, 2025-W42-L3-2, 2025-W42-L6-5 |
| P-2025-W42-02 | October 15 | Claude Haiku 4.5 targets faster, lower-cost tasks | 2025-W42-L1-5, 2025-W42-L3-1 |
| P-2025-W42-03 | October 15 | Veo 3.1 adds control over generated video and audio | 2025-W42-L3-5 |
| P-2025-W42-04 | October 15 | A Gemma-based model proposes a route to make tumors more visible | 2025-W42-L6-1 |
| P-2025-W42-05 | October 16 | Agent Skills package reusable instructions and tools | 2025-W42-L1-5, 2025-W42-L3-1, 2025-W42-L6-2 |
| P-2025-W42-06 | October 16 | Windows expands Copilot voice and screen assistance | 2025-W42-L3-3 |
| P-2025-W43-01 | October 21 | ChatGPT Atlas puts an assistant inside the browser | 2025-W43-L3-1, 2025-W43-L6-1 |
| P-2025-W43-02 | Reported October 21 | Channel 4 reveals an AI-generated presenter | 2025-W43-L6-2 |
| P-2025-W43-03 | Reported October 22 | A public statement calls for conditions on superintelligence | 2025-W43-L3-2 |
| P-2025-W43-04 | October 22 | Applied Digital leases capacity at a campus under construction | 2025-W43-L1-4 |
| P-2025-W43-05 | October 23 | Copilot gains a face, group chats and memory | 2025-W43-L3-4 |
| P-2025-W43-06 | October 23 | Claude extends optional memory to Pro and Max plans | 2025-W43-L3-5, 2025-W43-L6-4 |
| P-2025-W43-07 | October 24 | SoftBank and Samsung plan AI research for mobile networks | 2025-W43-L1-3 |
| P-2025-W43-08 | Reported October 25–26 | Models resist a shutdown mechanism in new tests | 2025-W43-L1-1 |
| November | |||
| P-2025-W44-01 | October 27–28 | Qualcomm announces accelerators for data center inference | 2025-W44-L6-2 |
| P-2025-W44-02 | October 28 | Argonne plans two new AI supercomputers | 2025-W44-L4-1 |
| P-2025-W44-03 | October 28 | NVQLink connects quantum processors to GPU supercomputers | 2025-W44-L6-1 |
| P-2025-W44-04 | October 29 | Experiments test whether models can detect internal changes | 2025-W44-L6-3 |
| P-2025-W44-05 | October 29 | Nvidia becomes the first $5 trillion company | 2025-W44-L1-1 |
| P-2025-W44-06 | October 30 | Universal Music and Stability AI plan music-creation tools | 2025-W44-L1-3 |
| P-2025-W44-07 | October 30 | Aardvark enters private testing for software security | 2025-W44-L1-5 |
| P-2025-W45-01 | November 3 | OpenAI signs a $38 billion AWS computing agreement | 2025-W45-L1-1, 2025-W45-L2-4, 2025-W45-L4-3 |
| P-2025-W45-02 | November 3 | Microsoft gains licenses to ship advanced Nvidia chips to the UAE | 2025-W45-L1-5 |
| P-2025-W45-03 | Reported November 5 | Apple is reported to near a deal to use Gemini for Siri | 2025-W45-L1-2 |
| P-2025-W45-04 | Reported November 5 | China is reported to bar foreign AI chips from state-funded data centers | 2025-W45-L4-4 |
| P-2025-W45-05 | November 6 | Kimi K2 Thinking combines reasoning with tool use | 2025-W45-L2-1 |
| P-2025-W45-06 | November 6 | Google prepares Ironwood TPUs for general availability | 2025-W45-L1-3 |
| P-2025-W45-07 | Reported November 6 | Perplexity agrees to pay Snap $400 million for search in Snapchat | 2025-W45-L2-HM2 |
| P-2025-W45-08 | November 6 | OpenAI publishes a framework for teen safety | 2025-W45-L2-2 |
| P-2025-W45-09 | November 7 | Meta pledges $600 billion for US infrastructure | 2025-W45-L4-5 |
| P-2025-W46-01 | November 10 | Meta expands speech recognition to more than 1,600 languages | 2025-W46-L1-6, 2025-W46-L3-3, 2025-W46-L4-3, 2025-W46-L5-6 |
| P-2025-W46-02 | November 10 | Johns Hopkins announces two cancer AI research projects | 2025-W46-L1-2 |
| P-2025-W46-03 | Reported November 11 | Yann LeCun is reported to plan a departure from Meta | 2025-W46-L1-5 |
| P-2025-W46-04 | November 12 | GPT-5.1 updates reasoning and conversational controls | 2025-W46-L1-1, 2025-W46-L3-5, 2025-W46-L4-2 |
| P-2025-W46-05 | November 12 | Anthropic plans $50 billion in US computing infrastructure | 2025-W46-L4-5 |
| P-2025-W46-06 | November 12 | Nvidia sweeps MLPerf Training v5.1 | 2025-W46-L4-4 |
| P-2025-W46-07 | November 13 | Sparse networks make some model computations easier to inspect | 2025-W46-L1-7 |
| P-2025-W46-08 | November 13 | Anthropic reports misuse of Claude Code in an espionage campaign | 2025-W46-L5-1 |
| P-2025-W46-09 | November 16 | AI helps simulate the Milky Way star by star | 2025-W46-L3-1 |
| P-2025-W47-01 | November 17 | Grok 4.1 emphasizes conversational behavior | 2025-W47-L3-1 |
| P-2025-W47-02 | November 18 | Gemini 3 Pro enters preview across Google products | 2025-W47-L1-1 |
| P-2025-W47-03 | November 18 | Antigravity brings coding agents into a shared workspace | 2025-W47-L2-4, 2025-W47-L4-1 |
| P-2025-W47-04 | November 18 | Anthropic expands its Microsoft and Nvidia partnerships | 2025-W47-L1-2, 2025-W47-L2-2, 2025-W47-L6-1 |
| P-2025-W47-05 | November 19 | Codex-Max is designed for longer software tasks | 2025-W47-L1-5, 2025-W47-L2-1, 2025-W47-L5-1 |
| P-2025-W47-06 | November 19 | OpenAI offers a free ChatGPT for US teachers | 2025-W47-L1-3 |
| P-2025-W47-07 | November 19 | Nvidia reports $57 billion in quarterly revenue | 2025-W47-L7-1 |
| P-2025-W47-08 | November 20 | Olmo 3 opens more of the model-development process | 2025-W47-L1-4 |
| P-2025-W47-09 | November 20 | OpenAI and Foxconn plan US AI hardware manufacturing | 2025-W47-L1-3, 2025-W47-L2-3 |
| P-2025-W48-01 | November 24 | Claude Opus 4.5 targets coding and sustained agent work | 2025-W48-L1-1, 2025-W48-L3-3, 2025-W48-L4-2 |
| P-2025-W48-02 | November 24 | Genesis Mission sets a national program for AI-assisted science | 2025-W48-L1-2, 2025-W48-L3-1, 2025-W48-L4-3 |
| P-2025-W48-03 | November 24 | AWS plans up to $50 billion for government computing | 2025-W48-L1-4, 2025-W48-L3-4 |
| P-2025-W48-04 | November 25 | A bill would raise penalties for AI-assisted fraud | 2025-W48-L1-5 |
| P-2025-W48-05 | November 26 announcement; November 28 publication | USPTO revises guidance on AI-assisted inventions | 2025-W48-L1-3 |
| P-2025-W48-06 | Reported November 30 | Poetic prompts expose a weakness in safety testing | 2025-W48-L4-5 |
| December | |||
| P-2025-W49-01 | December 1 | DeepSeek releases V3.2 and a specialized reasoning variant | 2025-W49-L3-5 |
| P-2025-W49-02 | December 1 | Apple reorganizes its AI leadership | 2025-W49-L5-3 |
| P-2025-W49-03 | December 1 | Runway introduces Gen-4.5 video generation | 2025-W49-L5-2 |
| P-2025-W49-04 | Reported December 1–2 | OpenAI declares a “code red” to improve ChatGPT | 2025-W49-L3-2, 2025-W49-L5-1, 2025-W49-L6-2 |
| P-2025-W49-05 | December 4 | Gemini 3 Deep Think reaches Ultra subscribers | 2025-W49-L3-1, 2025-W50-L1-3 |
| P-2025-W49-06 | December 5 | Meta acquires AI-wearables company Limitless | 2025-W49-L3-4 |
| P-2025-W50-01 | December 9 | Accenture and Anthropic plan a larger enterprise practice | 2025-W50-L3-3 |
| P-2025-W50-02 | December 10 | Adobe brings editing tools into ChatGPT | 2025-W50-L5-5 |
| P-2025-W50-03 | December 10 | A 3D chip from a US foundry speeds AI workloads | 2025-W50-L1-5 |
| P-2025-W50-04 | December 11 | GPT-5.2 expands the model family for professional work | 2025-W50-L1-2, 2025-W50-L3-2, 2025-W50-L5-1 |
| P-2025-W50-05 | December 11 | A federal order targets conflicting state AI rules | 2025-W50-L1-1, 2025-W50-L3-1, 2025-W51-L3-4, 2025-W52-L3-1 |
| P-2025-W50-06 | December 11 | Google opens Deep Research to developers | 2025-W50-L5-3 |
| P-2025-W50-07 | December 11 | Runway announces interactive world models | 2025-W50-L5-4 |
| P-2025-W51-01 | Background: December 11; selected December 21; restored by the editor, October 2026 | Disney and OpenAI agree to bring Disney characters to Sora | 2025-W51-L3-2, 2025-W52-L3-4 |
| P-2025-W51-02 | December 15 | Nvidia acquires Slurm developer SchedMD | 2025-W51-L1-4 |
| P-2025-W51-03 | December 15 | Nvidia releases its Nemotron 3 open models | 2025-W51-L3-5 |
| P-2025-W51-04 | December 16 | ChatGPT Images adds a new editing model | 2025-W51-L1-1 |
| P-2025-W51-05 | December 17 | Gemini 3 Flash broadens access to the model family | 2025-W51-L3-1, 2025-W51-L5-2 |
| P-2025-W51-06 | December 18 | Agent Skills becomes an open format | 2025-W51-L1-2 |
| P-2025-W51-07 | December 18 | GPT-5.2-Codex targets extended software work | 2025-W51-L5-1, 2025-W51-L3-1 |
| P-2025-W51-08 | December 18 | Genesis Mission adds 24 research collaborators | 2025-W51-L1-3 |
| P-2025-W51-09 | December 18 | Britain publishes its first Frontier AI Trends Report | 2025-W51-L3-4 |
| P-2025-W52-01 | December 22 | GLM-4.7 adds continuity to coding workflows | 2025-W52-L1-2 |
| P-2025-W52-02 | December 24 | Groq licenses inference technology to Nvidia | 2025-W52-L6-4, 2025-W52-L11-1 |
| P-2025-W52-03 | December 27 | China consults on human-like AI interaction services | 2025-W52-L1-3, 2025-W52-L6-2 |
| P-2025-W52-04 | December 28 | NHS organizations use AI to forecast emergency demand | 2025-W52-L1-1, 2025-W52-L6-1 |
| P-2025-W52-05 | December 28 | Hinton says AI has progressed faster than he expected | 2025-W52-L11-4 |
| P-2025-W53-01 | December 29 | Manus announces that it is joining Meta | 2025-W53-L1-1, 2025-W53-L2-2 |
| P-2025-W53-02 | December 30 | xAI launches Grok Business and Enterprise | 2025-W53-L1-2 |
| P-2025-W53-03 | December 31 announcement; final funding December 26 | SoftBank confirms completion of its OpenAI investment | 2025-W53-L2-1 |
| P-2025-W53-04 | December 31 submission | OpenForecaster studies open-ended predictions | 2025-W53-L3-1 |
| P-2025-W53-05 | December 31 submission | AdaGReS selects context with less repetition | 2025-W53-L3-2 |
| P-2025-W53-06 | December 31 submission | RAIR tests relevance in product search | 2025-W53-L3-5 |