AI News for November 3, 2026
Browse any day on a calendar, install it to your home screen, read offline.Read in the app →-
1TechCrunch06:14
Anthropic warns that it's very close to RSI capability, aka self-building AI
Anthropic says its systems are approaching recursive self-improvement, the threshold at which a model becomes good enough at training AI to sharply accelerate its own capability gains. The company presents this as a safety boundary that requires new controls rather than a product milestone, which puts pressure on how frontier labs define and enforce capability thresholds.
Read original ↗ -
2TechCrunch00:30
42AI comes out of stealth with $62.5M and a world model that can 'predict human behavior'
42AI exited stealth with $62.5M in funding and a world model the startup says can anticipate how people will act, including in situations the model has not seen before. The claim puts it among a small group of companies building predictive world models rather than text-first systems, though the approach has yet to be validated publicly.
Read original ↗ -
3TechCrunch03:19
New York Times and other outlets sue web browser Brave over alleged content scraping
The New York Times, joined by the New York Daily News and the Chicago Tribune, sued Brave over alleged scraping of copyrighted articles that the browser company then sold on to AI customers including Anthropic and Perplexity. The suits also accuse Brave of letting users bypass paywalls, extending publisher litigation from model developers to the data supply chain. The Verge also reported the filings.
Read original ↗ -
4TechCrunch07:13
Trajectory raises $45M to make AI model training accessible to nontechnical experts
Trajectory, founded a year ago, raised $45M at a $300 million valuation for tooling that lets domain experts without engineering backgrounds train and adapt models. The round reflects continued investor appetite for the layer between foundation models and vertical deployment.
Read original ↗ -
5The Verge02:01
OpenAI pitches data center power plants as a way to protect nearby areas
OpenAI will pilot a "Grid Protectors Program" at a planned data center in Effingham County, Georgia, promising to release up to 1,000 megawatts to the local grid during extreme weather or emergencies. The pitch comes as local opposition to data center buildouts grows over power and water demand.
Read original ↗ -
6TechCrunch00:23
Elon Musk throttles Grok Bot limits due to 'unsustainably high' costs
Musk said xAI is tightening usage limits on Grok Bot because serving costs have become unsustainably high, days after calling it the fastest-growing product he has seen. The move is a concrete data point on how far inference economics still lag consumer demand for agent-style products.
Read original ↗ -
7The Verge02:56
Anthropic clamps down on Claude’s AI video game clones
Anthropic updated Claude Code's system prompt to block the tool from re-creating existing video games, while still allowing original game development. The change responds to a wave of AI-made decompilations and ports circulating in the retro gaming scene. Ars Technica argues the restriction is unlikely to slow the broader cloning trend, since no single model provider controls enough of the market.
Read original ↗ -
8TechCrunch23:00
Omen AI adds liquid know-how in $16 million Aquatic Labs acquisition
Omen AI is acquiring Aquatic Labs for $16 million, combining two startups working on water consumption in data centers. Cooling and water use have become a recurring obstacle to permitting new AI capacity, making the segment an increasingly active area for consolidation.
Read original ↗ -
9TechCrunch09:49
Apple may try to increase revenue by doing...something with third-party AI apps
On its Monday earnings call, Apple hinted at unspecified "opportunities" to generate revenue from third-party AI apps without detailing what form they would take. Any such arrangement would matter to AI developers distributing through the App Store, given Apple's existing control over payments and placement.
Read original ↗ -
10TechCrunch21:57
OpenAI is finally merging ChatGPT's Chat and Work tabs
OpenAI is consolidating ChatGPT's separate Chat and Work tabs into a single interface. After the change, ChatGPT and Codex will draw from the same usage allocation, which affects how teams budget their quota across conversational and coding workloads.
Read original ↗ -
11TechCrunch11:30
Google takes Dreambeans to India in bid to bring AI-generated personal stories to the masses
Google expanded Dreambeans, its AI storytelling experiment, to India with support for 13 Indian languages. The rollout is a test of whether consumer generative products can find scale in a market where multilingual coverage, not model quality alone, determines reach.
Read original ↗ -
12TechCrunch08:35
Y Combinator's Garry Tan uses AI agents so much, he spends up to $200K a year on tokens
Speaking on Walter Isaacson's podcast The Builders, Y Combinator's Garry Tan said his personal AI agent usage costs as much as $200K a year in tokens. The figure is an outlier, but it illustrates the spending ceiling heavy agent workflows can reach for a single user.
Read original ↗
