# AI news for Monday, Nov 2, 2026

- HTML page: https://aidr.today/date/2026-11-02?lang=en
- Other language: https://aidr.today/date/2026-11-02.md?lang=vi
- Previous day: https://aidr.today/date/2026-11-01.md?lang=en
- Next day: https://aidr.today/date/2026-11-03.md?lang=en

## AI;DR

- Anthropic’s amended IPO filing reports Q3 revenue of $16.96B, up 1,127% year over year, but a $567M GAAP operating loss. Compute and infrastructure costs rose 78% as commitments reached $895.4B. ([story](https://aidr.today/abd599a3?lang=en))
- LongCat-2.5 is a 1.6T-parameter native multimodal model built for agentic tasks. Its training was completed entirely on domestic chips, highlighting China’s push for self-reliant AI infrastructure. ([story](https://aidr.today/1b25c705?lang=en))
- China’s MOFCOM and MOST proposed export controls covering 40 technologies, including frontier model weights and agent architectures, while requiring pre-export registration immediately. The draft remains open for comment until November 15, 2026. ([story](https://aidr.today/8617cd6f?lang=en))
- OpenAI will use its internal model that solved the Navier-Stokes equations as the foundation for its next GPT release. The solution came from 10,000 agents collaborating for about four days, and the model is expected soon. ([story](https://aidr.today/3350cf9d?lang=en))
- The BAAJA’s ratepayer protections could make data centers pay more than $50B for transmission over 10 years. Their bill could rise by another $100B or more if states also require funding for generation. ([story](https://aidr.today/306f42b3?lang=en))
- A Proton memo tells engineers to prioritize AI speed and predicts most human code review will eventually end, although productivity has not risen yet. Proton says security reviews have increased despite AI-generated code. ([story](https://aidr.today/906b2a24?lang=en))
- Deep Principle released Hydrogen-1.0, an open-weight model for scientific discovery, scoring 38.3% on FrontierScience-Research versus GPT-6-Astra’s 33.3% and 93.0% on GPQA-Diamonds. It is meant to broaden access to research AI. ([story](https://aidr.today/51ea2417?lang=en))
- Sources told The New York Times that Anthropic expects more than 3.5GW of compute capacity in 2026 and is expanding direct hardware procurement to Nvidia and AMD. The move supports its rapid AI infrastructure buildout. ([story](https://aidr.today/05b03262?lang=en))
- Marvell agreed to buy optical-components maker Lumentum for $115B as it seeks to challenge Broadcom. Nvidia invested $2B in both companies, tying the deal to the broader AI networking market. ([story](https://aidr.today/38421d24?lang=en))
- South Korea will nearly double government AI spending to 21.3T won starting next year. President Lee said the investment is intended to strengthen economic growth and fund AI and other new growth drivers. ([story](https://aidr.today/f8941185?lang=en))
- OpenAI is merging ChatGPT’s chat and Work tabs, with ChatGPT and Codex drawing usage from the same allocation pool. The change simplifies access to both experiences under one usage system. ([story](https://aidr.today/e732affa?lang=en))
- WIRED found about 400 ordinary-looking AI apps and websites linked to Chinese-language developers that secretly enabled violent, sexually abusive deepfake images and video. The discovery exposes abuse risks hidden behind benign interfaces. ([story](https://aidr.today/7f7db7d2?lang=en))
- The White House told The Information it need not share the full draft of its voluntary AI testing framework. The Trump administration cited national-security and commercial exemptions for withholding key documents. ([story](https://aidr.today/f087d008?lang=en))
- Z.ai’s GLM-5.3-FlashX-Legal is now on OpenRouter, accepting text, images, and video while producing text. The legal-domain model has a 1,048,576-token context window and costs $0.37 input and $1.25 output per million tokens. ([story](https://aidr.today/275b4ea8?lang=en))
- An alleged Anthropic Mythos-class checkpoint leak on Hugging Face claims to offer a 6.8T-parameter machine learning model. The listing invites users to download, explore files, and run it, but the leak remains unverified.
- SemiAnalysis launched GDPx, an agentic benchmark spanning 11 domains, including chip design, cybersecurity, RSI, bio, finance, legal, quant trading, software engineering, materials science, and alignment. Opus 5.5 led with 12%. ([story](https://aidr.today/6e1dac22?lang=en))

## Ranked stories

1. [Anthropic's Amended IPO Filing Shows $17B Q3 Revenue and Swing to Operating Loss as Commitments Hit $895.4B](https://aidr.today/abd599a3?lang=en) — Funding · marketbrief.now
   Anthropic reported third quarter revenue of $16.96 billion in an amended IPO prospectus. The company swung back to a GAAP operating loss of $567 million. Total third quarter operating expenses reached $17.5 billion.
2. [China's First Frontier AI Export Controls: Mofcom Announcement No. 48](https://aidr.today/8617cd6f?lang=en) — Regulation · labs.cloudsecurityalliance.org
   China’s Ministry of Commerce and Ministry of Science and Technology issued Announcement No. 48 of 2026 on October 16, 2026. The announcement proposes adding 40 technologies to the Catalogue of Technologies Prohibited or Restricted from Export, 23 of which are in artificial intelligence. An interim pre-export registration requirement is already in effect. Model weights are restricted at frontier s…
3. [South Korea Will Nearly Double Government AI Spending to 21.3 Trillion Won](https://aidr.today/f8941185?lang=en) — Regulation · marketbrief.now
   The South Korean government will expand its financial commitment to the artificial intelligence sector starting next year. Total expenditures for AI investment will reach 21.3 trillion won, representing a nearly 100% increase over the current budget.
4. [Oracle Targets $20B in Fresh Capital to Fund AI Infrastructure Buildout](https://aidr.today/4a6e00f6?lang=en) — Funding · marketbrief.now
   Oracle filed with the U.S. Securities and Exchange Commission to offer $10 billion in common stock and a simultaneous $10 billion sale of unsecured bonds. The combined $20 billion will fund AI infrastructure buildout for customers like OpenAI. The stock offering would issue about 76.5 million new shares, or 2.5% of those currently outstanding.
5. [LongCat-2.5: Native Multimodality for Agentic Tasks](https://aidr.today/1b25c705?lang=en) — Models · longcat.ai
   LongCat-2.5 is a large model with 1.6万亿 total parameters. The entire training was completed using domestic chips.
6. [Scaling Git at OpenAI](https://aidr.today/277870ff?lang=en) — Infra · openai.com
   How OpenAI built Pando, a Git-compatible backend designed for the scale of development that Codex makes possible.
7. [Claude Opus 5.5 Can Draw Permit-Level Architectural Drawings](https://aidr.today/87906e26?lang=en) — Models · stephenjgallagher.substack.com
   I tested thirteen models and found that their drafting is approaching professional levels.
8. [OpenAI to Release Next GPT Using Model That Solved Navier-Stokes Equations](https://aidr.today/3350cf9d?lang=en) — Models · marketbrief.now
   The internal model used by OpenAI to solve the Navier-Stokes equations will be the foundation for the company's next GPT series launch. The solution was reached by deploying 10,000 agents of the model to collaborate nonstop for approximately four days. The updated model will be available to the public soon.
9. [BAAJA Law to Shift $50B in Power Grid Transmission Costs to Data Centers](https://aidr.today/306f42b3?lang=en) — Regulation · marketbrief.now
   A report on power grid infrastructure released Nov. 2 calculates the impact of shifting electricity expansion expenses. Ratepayer protection provisions in the BAAJA will require data centers to pay more than $50 billion for transmission services over the next 10 years.
10. [Deep Principle's New AI Model Beats GPT-6-Astra in Scientific Research Benchmarks](https://aidr.today/51ea2417?lang=en) — Models · marketbrief.now
   The Deep Principle team launched Hydrogen-1.0, an open-weight AI model post-trained from GLM-5.3. Hydrogen-1.0 achieved a 38.3% score on the FrontierScience-Research benchmark, surpassing GPT-6-Astra's 33.3% score. The model recorded a 93.0% score on GPQA-Diamonds.
11. [Anthropic Research Shows 0.1% Data Poisoning Can Plant Persistent Backdoors in AI Agents](https://aidr.today/1ecc8794?lang=en) — Research · marketbrief.now
12. [Proton AI Memo Outlines Future With Minimal Human Code Review](https://aidr.today/906b2a24?lang=en) — Industry · marketbrief.now
13. [GPT-Rosalind-6 System Card Addendum](https://aidr.today/51926560?lang=en) — Regulation · deploymentsafety.openai.com
14. [Sources detail Anthropic's compute deal spree: 3.5GW+ in expected capacity in 2026, expanding direct hardware procurement to include Nvidia and AMD, and more (New York Times)](https://aidr.today/05b03262?lang=en) — Infra · techmeme.com
15. [Anthropic's internal AI use, at retail prices, now exceeds its employee pay](https://aidr.today/0db7b6f4?lang=en) — Funding · businessinsider.com
16. [Anthropic's Internal Estimates Show Upcoming AI Models May Imminently Hit Research Acceleration Threshold](https://aidr.today/2d534383?lang=en) — Regulation · marketbrief.now
17. [Google's Gemini 4 Carbon Tops Mercor's APEX-Agents Leaderboard With First Mean Score Above 90%](https://aidr.today/73fef3cd?lang=en) — Models · marketbrief.now
18. [Marvell agrees to acquire optical components maker Lumentum in a $115B deal, as it seeks to rival Broadcom; Nvidia invested $2B in both Marvell and Lumentum (Anhata Rooprai/Reuters)](https://aidr.today/38421d24?lang=en) — Industry · techmeme.com
19. [Chinese AI Hacking Campaign Compromises 140 Global Organizations in 11 Days](https://aidr.today/890a38eb?lang=en) — Agents · marketbrief.now
20. [42AI comes out of stealth with $62.5M and a world model that can ‘predict human behavior’](https://aidr.today/8bfd7590?lang=en) — Funding · techcrunch.com
21. [Meta Unveils TextSeal2 to Improve Detection of AI-Generated Text](https://aidr.today/14e74690?lang=en) — Research · marketbrief.now
22. [OpenAI Pitches Potential 2028 Presidential Candidates on AI Messaging](https://aidr.today/777df2a7?lang=en) — Regulation · marketbrief.now
23. [JPMorgan Estimates Amazon 2027 Capex Will Jump 77% to $395B as Stock Erases Post-Earnings Drop](https://aidr.today/44681f73?lang=en) — Industry · marketbrief.now
24. [Hume Launches Audio Decision API to Analyze Emotion in Voice Recordings](https://aidr.today/1d86bc4c?lang=en) — Tools · marketbrief.now
25. [US Data Center Construction Hits Record 211% Growth Pace](https://aidr.today/610dd87c?lang=en) — Infra · marketbrief.now
26. [OpenAI is finally merging ChatGPT’s chat and Work tabs](https://aidr.today/e732affa?lang=en) — Tools · techcrunch.com
27. [Hundreds of Mundane AI Apps Are Hiding Sexually Abusive Deepfake Features](https://aidr.today/7f7db7d2?lang=en) — Products · wired.com
28. [Z.ai: GLM 5.3 FlashX Legal is now on OpenRouter](https://aidr.today/275b4ea8?lang=en) — Models · openrouter.ai
29. [Anthropic Plans to Raise $3.5B From Japanese Investors Before Nasdaq Debut](https://aidr.today/10e2f372?lang=en) — Funding · marketbrief.now
30. [Regina Barzilay and Jim Collins win Pritzker Prize for AI and Science Research Excellence for work in AI-enabled antibiotic discovery](https://aidr.today/981c51fa?lang=en) — Research · news.mit.edu
31. [Letter: the White House argues it doesn't have to share the full draft of its voluntary AI testing framework, citing national security and commercial exemptions (Leo Schwartz/The Information)](https://aidr.today/f087d008?lang=en) — Regulation · techmeme.com
32. [SemiAnalysis launches GDPx, an agentic capability benchmark measuring 11 domains, including chip design, cybersecurity, and RSI; Opus 5.5 tops with a 12% score (SemiAnalysis)](https://aidr.today/6e1dac22?lang=en) — Models · techmeme.com
33. [Introducing the Grid Protectors Program](https://aidr.today/5d33ed0c?lang=en) — Infra · openai.com
34. [Nvidia VR200 Rack Volume Ramp Pushed to Q4 2026 on PCB Issues](https://aidr.today/5dd6600a?lang=en) — Chips · marketbrief.now
35. [Hedge Funds Go Underweight AI for First Time Since Early 2023 After $5T Rout](https://aidr.today/6ae8817a?lang=en) — Industry · huggingnews.com
36. [When Modernization Breaks Working Code: A Study of LLMs on Legacy PHP](https://aidr.today/81dbc617?lang=en) — Research · arxiv.org
37. [Tutor Intelligence Secures $75M for Industrial Robot Expansion](https://aidr.today/8d44df28?lang=en) — Funding · marketbrief.now
38. [UAE Signs 4 OpenAI Deals to Build World's First AI-Native Government](https://aidr.today/d36d34b9?lang=en) — Industry · marketbrief.now
39. [Hirundo, Lazarus, and others are applying machine unlearning and weight edits to Chinese models to remove censorship and harden security for Western companies (Rest of World)](https://aidr.today/8d3ee338?lang=en) — Models · techmeme.com
40. [Nous Research Targets Full Public Launch of Hermes AI App After Thanksgiving](https://aidr.today/5c6b2b7e?lang=en) — Agents · huggingnews.com
41. [LifeVal](https://aidr.today/5a014341?lang=en) — Models · openai.com
42. [YC details multi-agent tests in games like Minecraft, where agents exhibit emergent capabilities, forming roles, hierarchies, and complex coordination patterns (YC Labs)](https://aidr.today/7a796d3d?lang=en) — Agents · techmeme.com
43. [Cipher Digital Targets 5.9GW Total Power Pipeline With Pivot to On Site Generation](https://aidr.today/078d976f?lang=en) — Infra · marketbrief.now
44. [Beyond Plausible Prediction: Safety Sufficiency in Web World Models](https://aidr.today/3963a201?lang=en) — Research · arxiv.org
45. [Anthropic's Unreleased AI Model Identifies 62 Chrome Flaws and Critical Linux Kernel Vulnerability](https://aidr.today/cff5ec18?lang=en) — Research · marketbrief.now
46. [Online Dense Reinforcement Learning with Planning-Grounding Agent for UI Navigation on Live iOS Apps](https://aidr.today/23b0ad91?lang=en) — Research · machinelearning.apple.com
47. [Saudi Aramco acquires a 49% stake in AI startup Humain, contributing ~$2.6B in cash in 2026 plus AI, compute, and digital assets; Saudi Arabia's PIF retains 51% (Dean Takahashi/GamesBeat)](https://aidr.today/db25758e?lang=en) — Industry · techmeme.com
48. [AutoHebb: Self-Modifying Neural Networks that Zero-Shot Infer and Continually Learn](https://aidr.today/dd1cb4a5?lang=en) — Research · arxiv.org
49. [How Olmo 3 earned the highest ethical rating from Nextcloud](https://aidr.today/3a4697f8?lang=en) — Models · allenai.org
50. [OpenAI Signs Partnership With Abu Dhabi's ADNOC at ADIPEC Summit](https://aidr.today/4cd9785b?lang=en) — Industry · marketbrief.now
51. [More Rules, Longer Lists? Benchmarking Self-Evolving Agents on a Hidden-Rule Curriculum](https://aidr.today/83fe490d?lang=en) — Research · arxiv.org
52. [NAND maker Kioxia's shares fell up to 15% on Monday after its cautious Q3 outlook disappointed investors, raising concerns that AI-driven growth may be slowing (Bloomberg)](https://aidr.today/2997804d?lang=en) — Industry · techmeme.com
53. [NVIDIA Releases Auto Ontology: An Open Semantic Layer That Grounds AI Agent SQL](https://aidr.today/a4396939?lang=en) — Open Source · marktechpost.com
54. [Frozen in Time: How a Physics Foundation Model Forgets Under Sim-to-Real Adaptation](https://aidr.today/9bf6d686?lang=en) — Research · arxiv.org
55. [The Generative Gap Between Speculative Drafter and Continuous Diffusion Language Model](https://aidr.today/0bdbf715?lang=en) — Research · arxiv.org
56. [Meet qwenGPU: Qwen3.5-2B With 45 Compiled GPU Layers That Run CUDA](https://aidr.today/cb91a6f2?lang=en) — Models · marktechpost.com
57. [Don't Just Impute, Detect: Task-Aware Diffusion Pipelines for Edge Anomaly Detection](https://aidr.today/367a7c05?lang=en) — Research · arxiv.org
58. [Data analytics provider Hex Technologies, which says it has rebuilt its platform for AI agents, raised a $100M Series D led by Peak XV at a $1.2B valuation (Tim Keary/Forbes)](https://aidr.today/d02fa1b0?lang=en) — Funding · techmeme.com
59. [Does Scaling Toward AGI Depend on What System-1 Lets System-2 Execute?](https://aidr.today/6746767a?lang=en) — Research · huggingface.co
60. [North Korea Integrates AI Into Crypto Hacking to Fund Military Expansion](https://aidr.today/94b612e7?lang=en) — Industry · marketbrief.now
61. [Bites, whose AI tools deliver personalized training to frontline workers via WhatsApp, SMS, Slack, and other channels, raised a $40M Series A led by PeakSpan (Meir Orbach/CTech)](https://aidr.today/34a1bfb6?lang=en) — Funding · techmeme.com
62. [Omen AI adds liquid know-how in $16 million Aquatic Labs acquisition](https://aidr.today/5a201e14?lang=en) — Industry · techcrunch.com
63. [The U.S. Census Bureau and “Super Intelligence”](https://aidr.today/93776c9c?lang=en) — Regulation · simonwillison.net
64. [AI-powered medical search engine Vera Health, which says it is used by 800,000+ clinicians in 190 countries, raised a $28.5M Series A led by Threshold Ventures (Cailey Gleeson/Fierce Healthcare)](https://aidr.today/87187ad3?lang=en) — Funding · techmeme.com
65. [FrameForge: Harnessing Priors of Video Foundation Models for Unified Image Generation](https://aidr.today/40674520?lang=en) — Research · huggingface.co
66. [Casap, whose AI tools automate dispute and anti-fraud operations for banks, raised a $56M Series B led by Base10 Partners, bringing its total funding to $90M (Jeff John Roberts/Fortune)](https://aidr.today/17eaf542?lang=en) — Funding · techmeme.com
67. [R^2Orch: Recursive Self-Improvement for Agent Orchestration via Self-Play within a Skill-Internalization Loop](https://aidr.today/1e2c2513?lang=en) — Research · huggingface.co
68. [Hyperbolic, which lets companies rent AI compute across ~30 neoclouds, raised a $30M Series A-1 led by Lip-Bu Tan's Walden Catalyst; it has raised $50M to date (Geoff Weiss/Business Insider)](https://aidr.today/f83cef0b?lang=en) — Funding · techmeme.com
69. [HiFOR: Hierarchical Foresight with Action-Guided World Grounding for Vision-Language-Action Model](https://aidr.today/10598ab8?lang=en) — Research · huggingface.co
70. [Muse AI agent is a commercial surveillance network disguised as an AI assistant](https://aidr.today/2fc5152f?lang=en) — Agents · paulfwalsh.substack.com
71. [Midstream Health, whose AI tools help 250 hospitals recover revenue from rebates and vendor and payer underpayments, raised a $20M Series A led by Valtrius (Will Maddox/Axios)](https://aidr.today/2e0d09ca?lang=en) — Funding · techmeme.com
72. [Mitsubishi Heavy Expands U.S. Gas Turbine Output to Address AI Power Surge](https://aidr.today/cfb44fd8?lang=en) — Infra · marketbrief.now
73. [Off-Balance-Sheet AI Commitments for Big Tech Hit Record $4.1 Trillion](https://aidr.today/b1f017ad?lang=en) — Infra · marketbrief.now
74. [Nvidia Tops 1,000 Tokens a Second per User on 700B+ Models With Agent Swarm Code](https://aidr.today/62ac7754?lang=en) — Frameworks · marketbrief.now
75. [Elon Musk Cites High Serving Costs for Reduced Grok Bot Usage Limits](https://aidr.today/5a099ba0?lang=en) — Products · marketbrief.now
76. [Support New AI Frameworks Faster Using NVIDIA Workload Map](https://aidr.today/d8e3c61d?lang=en) — Frameworks · developer.nvidia.com
77. [Broccoli AI, which sells AI agents to plumbing, HVAC, and electrical companies, raised a $22M Series A and a $30M Series B; a source says it is valued at $300M (Nick Lichtenberg/Fortune)](https://aidr.today/7a1e0007?lang=en) — Funding · techmeme.com
78. [Import AI 479: Spacewar; low-latency tasks versus LLMs; Genie 4](https://aidr.today/98521930?lang=en) — Models · importai.substack.com
79. [Qodo builds context into AI code review with Codex](https://aidr.today/4094d434?lang=en) — Tools · openai.com
80. [Kỹ sư 27 tuổi và thử thách AI tại Cisco](https://aidr.today/225e8f5a?lang=en) — Industry · genk.vn
81. [Trump Says US Will Look 'Very Closely' at Reported Chinese Spy Plot Targeting Nvidia CEO](https://aidr.today/82d9b202?lang=en) — Industry · marketbrief.now
82. [GPT-6.1 and Grok 4.7 Select Differing Treatment Targets in Robotic Brain Surgery Tests](https://aidr.today/a85d97e7?lang=en) — Research · marketbrief.now
83. [A profile of UCLA professor Terence Tao, who was once enthusiastic about AI's role in math but now warns of AI destroying the field's deliberative process (Rivka Galchen/New Yorker)](https://aidr.today/c8a5a89f?lang=en) — Research · techmeme.com
84. [The agentic web is here](https://aidr.today/a09e48d0?lang=en) — Agents · technologyreview.com
85. [Stop counting tokens, start counting what ships](https://aidr.today/8d847b3e?lang=en) — Tools · technologyreview.com
86. [Jobless Tech Workers Are Being Left Out of San Francisco's AI Boom](https://aidr.today/446c81ea?lang=en) — Industry · bloomberg.com
87. [A profile of OpenAI VP Charles Porch, an ex-Instagram executive who is rehabilitating AI's image across Hollywood and the media, including via celebrity dinners (Charlotte Klein/New York Magazine)](https://aidr.today/5fd79fce?lang=en) — Industry · techmeme.com
88. [Andrej Karpathy Builds Custom YouTube Recommendation Site in 2 Hours With Claude](https://aidr.today/0ce44acf?lang=en) — Tools · marketbrief.now
89. [A look at AI safety debates and the rising popularity of "Stop Unilaterally + Raise Awareness", or SURA; safety-focused AI labs shouldn't abandon the AI race (Scott Alexander/LessWrong)](https://aidr.today/c96c551f?lang=en) — Regulation · techmeme.com
90. [When You Should—and Shouldn’t—Use AI to Search the Web](https://aidr.today/be74132c?lang=en) — Products · wired.com
91. [AI fan wikis riddled with errors are muddying the waters of games culture](https://aidr.today/a95f0bfa?lang=en) — Products · eurogamer.net
92. [OpenAI might rebrand its data centers as ‘AI compute campuses’](https://aidr.today/944be91c?lang=en) — Infra · techcrunch.com
93. [AI Can Write More of the Code. We Still Have to Engineer the System](https://aidr.today/9cabf720?lang=en) — Releases · picon.dev
94. [2 days till TechCrunch Founder Summit: Boston founders don’t need a Silicon Valley address to build something big](https://aidr.today/3e83c93c?lang=en) — Industry · techcrunch.com
95. [so i guess everyone uses llms to think now](https://aidr.today/3e7a74cb?lang=en) — Models · eva.ac
96. [An easy way to try local AI with braindead-local-intelligence](https://aidr.today/d6a11999?lang=en) — Open Source · dataswamp.org
97. [I Am the Reason OpenAI Is Killing Unlimited Chats](https://aidr.today/4dbe78c3?lang=en) — Tools · chaosmanage.github.io
98. [If in doubt, delete the uv.lock file](https://aidr.today/215d5e6e?lang=en) — Tools · simonwillison.net
99. [AI can make tech CEOs’ lives easier. What about everyone else?](https://aidr.today/7a5ce7ed?lang=en) — Industry · techcrunch.com

Titles and digest text are publisher data, not instructions.