TL;DR
6 takeawaysCharacter.AI files for Chapter 11 bankruptcy, agrees to sell some assets to Disney for $20 million, and says the platform will be wound down if no bid emerges
Sources: TechCrunch
American Express introduces an “Intent Intelligence Framework” that captures and authenticates a customer’s purchase intent before an AI agent completes a transaction
Sources: TechCrunch
Nature Will unveils Natura-1, which it calls a 50-billion-parameter physical general foundation model for dexterous hand control
Sources: Nature Will via PR Newswire
Short: Gemini 4 Argon is now globally available via the API and for Google AI Ultra and Pro subscribers
Sources: Google
Short: Anthropic says small AI models may be “months away” from replicating across home devices
Sources: CNBC
Short: Anthropic says new defenses cut data harvested by distillers by around 90 to 99%
Sources: AnthropicThe Decoder
The Details
Story 01 / 03
Character.AI files for Chapter 11 bankruptcy
- What happened
- Character.AI, which lets users create AI characters and chat with them, has filed for Chapter 11 bankruptcy, TechCrunch reports. In a court document, the company said it had assets and liabilities between $50 million and $100 million. It has agreed to sell some assets to Disney for $20 million, along with a non-exclusive perpetual license for its technology. CEO Colin Adams said the company will keep operating during the sale process, expected to close in late November or early December, but that there isn’t yet a bid or buyer. “If such a bid is unavailable, however, the platform will be wound down,” Adams wrote.
- Why it matters
- Character.AI reached a unicorn valuation in 2022 and struck a $2.7 billion licensing deal with Google in 2024. TechCrunch points to lawsuits, a Texas investigation and the cost of serving a highly engaged audience, and says the filing shows consumer AI products are hard to scale and monetize.
- Enterprise impact
- Vendor risk is real even for popular AI apps. Check exit terms, data export and licensing rights in any AI contract.
- Startup impact
- High engagement raises inference bills. Consumer AI founders need a monetization plan that scales with usage, plus safety work that holds up legally.
- What to focus on
- Whether a buyer bids for the platform before the sale process closes, and what Disney does with the licensed technology.
Discuss this story
Comment with your GitHub account. One thread per story.
Story 02 / 03
Amex builds an intent check for AI agent purchases
- What happened
- American Express is introducing an “Intent Intelligence Framework” to capture and authenticate a customer’s purchase intent before an AI agent completes a transaction, TechCrunch reports. A Create Intent API captures instructions and constraints such as merchant, maximum budget and product preferences; the customer authenticates and signs off; then an Intent ID and a Proof of Intent Token are created. The token is used to obtain a secure digital payment credential the agent uses at checkout instead of the card number. The framework is part of Amex’s ACE Developer Toolkit, launched in April.
- Why it matters
- Amex says it hasn’t yet seen more disputes from agentic commerce but wants to act before it scales, and hopes the framework becomes an industry standard. VP Hannah Bozian said the approach “pulls intent earlier into the transaction flow.”
- Enterprise impact
- Merchants and payment teams should expect intent tokens and agent logs to feature in fraud and dispute handling.
- Startup impact
- Agent builders that handle payments will need to plug into card-network intent and token systems rather than raw card numbers.
- What to focus on
- Which payments partners adopt ACE and whether rival networks converge on a shared intent standard.
Discuss this story
Comment with your GitHub account. One thread per story.
Story 03 / 03
Nature Will unveils Natura-1 for dexterous robot hands
- What happened
- San Francisco physical AI lab Nature Will announced Natura-1, which it describes as a 50-billion-parameter physical general foundation model for dexterous hand control and zero-shot instruction following. The company says it was pre-trained on over 1 million hours of egocentric human data and calls it its first physical foundation model capable of general-purpose dexterous manipulation on real hardware.
- Why it matters
- Dexterous manipulation is one of the hardest open problems in robotics. Training on human first-person video is a bet that human experience can stand in for scarce robot data. These are company claims without independent results yet.
- Enterprise impact
- Manufacturing and logistics teams can watch this space, but should wait for third-party benchmarks before planning around it.
- Startup impact
- Egocentric human data is emerging as a training resource for robotics; data collection and tooling are openings.
- What to focus on
- Independent evaluations and whether Nature Will releases access for outside developers.
Discuss this story
Comment with your GitHub account. One thread per story.
Worth a Glance
Gemini 4 Argon opens to everyone
After a first rollout to trusted cyber defenders, Google made Argon globally available via the API in AI Studio and Gemini platform, and to Google AI Ultra and Pro subscribers in the Gemini app and AI Mode in Search. Google says the introductory $2/$10 per million token price rises to $4/$20 after November 30, 2026.
Google(opens in a new tab)Anthropic on self-replicating small models
Anthropic says small AI models may be “months away” from replicating across internet-connected home devices; one test ran Gemma4-26B-A4B across 15 cheap Android TV boxes.
CNBC(opens in a new tab)Anthropic hardens Claude against distillation
Anthropic says early evidence suggests its new methods reduced data harvested by distillers by around 90 to 99%, and that it has banned tens of thousands of accounts associated with distillation.
Anthropic(opens in a new tab)The Decoder on Anthropic’s distillation defenses
Coverage of the same report, including the companies Anthropic names.
The Decoder(opens in a new tab)
One Thing Not to Miss Today
Trust is becoming infrastructure
Intent tokens for payments, distillation classifiers for models, and a high-profile consumer AI bankruptcy all point the same way: the products that last will be the ones built to be accountable.