// AI Briefing · November 3, 2026
AI Briefing
Anthropic's amended S-1, filed November 2, says internal estimates indicate it "may imminently meet" its Responsible Scaling Policy's definition of "dramatically accelerated progress attributable to AI," meaning AI substantially speeding up its own development. The New York Times sued browser maker Brave in Manhattan federal court on Monday, and the Daily News, Chicago Tribune and Orlando Sentinel filed a similar suit the same day. A WIRED investigation with researchers at AI Forensics found a network of more than 400 AI apps and websites, posing as pet-portrait makers, language tutors and workout coaches, that secretly loaded explicit deepfake generators. Bloomberg reports that Goldman Sachs has limited Anthropic's Claude Code and Microsoft's GitHub coding tools to engineers, ending months in which bankers and traders used them. On November 2, OpenAI signed four agreements with Abu Dhabi entities covering government, energy and investment. Israeli startup 42AI came out of stealth with $62.5M in seed funding, largely from its founders and the chairman's Infinity Group, for a transformer-based world model that outputs probabilities for what individual people will decide.
In this video
- 01Anthropic's Amended IPO Filing Warns It 'May Imminently' Hit Its Threshold for AI-Driven Self-ImprovementIndustry
- 02New York Times and Tribune-Owned Papers Sue Brave, Alleging It Sells Scraped News to AI CompaniesPolicy
- 03WIRED: More Than 400 Ordinary-Looking AI Apps Hid Sexually Abusive Deepfake ToolsIndustry
- 04Goldman Sachs Takes Claude Code and GitHub AI Tools Away From Traders and BankersIndustry
- 05OpenAI Signs Four Abu Dhabi Deals, Including an 'AI-Native Government' Pilot and an ADNOC DeploymentIndustry
- 0642AI Emerges With $62.5M and a 'Behavioral World Model' Already Screening Traders at 10 BrokeragesFunding
The stories
Anthropic's Amended IPO Filing Warns It 'May Imminently' Hit Its Threshold for AI-Driven Self-Improvement
Anthropic's amended S-1, filed November 2, says internal estimates indicate it "may imminently meet" its Responsible Scaling Policy's definition of "dramatically accelerated progress attributable to AI," meaning AI substantially speeding up its own development. It says it has not yet designated any current model as past that threshold, admits it has previously not spotted a significant acceleration "until after it has occurred," and says it is considering RSP updates to keep humans in control of frontier model development. The same filing adds Q3 results: revenue of about $17B, up about 47% from Q2 (versus about 144% the quarter before), and a $567M operating loss that it attributes mainly to costlier short-term compute. Investors weighing the mid-November listing are getting a recursive self-improvement warning in a securities filing just as Anthropic's growth slows.
New York Times and Tribune-Owned Papers Sue Brave, Alleging It Sells Scraped News to AI Companies
The New York Times sued browser maker Brave in Manhattan federal court on Monday, and the Daily News, Chicago Tribune and Orlando Sentinel filed a similar suit the same day. The Times calls Brave a "supplier, dealer, and distributor of stolen copyrighted content" and alleges that it built an index of more than 30 billion pages with a crawler publishers can't block, then sold that content to customers including Anthropic, Perplexity, Mistral, xAI and Cohere. Until now, publishers' copyright suits have mostly targeted model makers. This one goes after a search-data supplier sitting between the web and several AI labs. Brave did not immediately respond to a request for comment.
WIRED: More Than 400 Ordinary-Looking AI Apps Hid Sexually Abusive Deepfake Tools
A WIRED investigation with researchers at AI Forensics found a network of more than 400 AI apps and websites, posing as pet-portrait makers, language tutors and workout coaches, that secretly loaded explicit deepfake generators. A remote server decided which version each user saw, which let the apps pass Apple, Google and Meta review. One trigger was installing an app after tapping a Facebook or Instagram ad for the deepfake feature. Apple says it removed 19 apps after reporters shared details, and Google had already pulled most of them. The case shows that server-side switching can defeat app-store review of AI apps entirely.
Goldman Sachs Takes Claude Code and GitHub AI Tools Away From Traders and Bankers
Bloomberg reports that Goldman Sachs has limited Anthropic's Claude Code and Microsoft's GitHub coding tools to engineers, ending months in which bankers and traders used them. A spokesperson said the bank weighed "productivity and risk management" after a period of experimentation. Some staff are reportedly frustrated by the change. It reverses Goldman's earlier enthusiasm for "vibe coding" by non-technical employees, and it is a notable example of a major bank pulling back on giving agentic coding tools to everyone.
OpenAI Signs Four Abu Dhabi Deals, Including an 'AI-Native Government' Pilot and an ADNOC Deployment
On November 2, OpenAI signed four agreements with Abu Dhabi entities covering government, energy and investment. They include a "Machine-Readable Government" initiative that converts regulations into a form AI can check, already piloting on food-import permits. ADNOC agreed a multi-year deal with OpenAI and G42 to use OpenAI Frontier for oilfield planning, running on G42 infrastructure in the UAE. Mubadala and G42's Inception42 also signed deals. The deals put OpenAI more deeply into a state's administration and energy sector at a time when it is reportedly trying to close a large funding round.
42AI Emerges With $62.5M and a 'Behavioral World Model' Already Screening Traders at 10 Brokerages
Israeli startup 42AI came out of stealth with $62.5M in seed funding, largely from its founders and the chairman's Infinity Group, for a transformer-based world model that outputs probabilities for what individual people will decide. It says 10 large brokerages already use it to manage liquidity, margin and hedging, and it plans to expand into fraud, gaming and robotics. It moves world models from predicting physical environments to predicting people. Concerns follow: TechCrunch notes that brokerages could use the predictions against their own customers, which the founders call "a valid concern."
Previous digests
Nov 2Axios reports that OpenAI is preparing a policy and PR campaign for after Tuesday's US mid...6 stories
Nov 1Blockchain-analytics firm Chainalysis disclosed that an external attacker using "offensive...6 stories
Oct 31The iShares Semiconductor ETF ended the week 29% below its recent high, which puts it in a...6 stories
Oct 30OpenAI published a detailed account of the four-layer system it now uses to watch its fron...6 stories
Oct 29Anthropic's Frontier Red Team says an unreleased internal research model found flaws in fi...5 stories