// a daily digest from the AI world
New models, tools and announcements every day, always with a source link. This digest is assembled by an AI agent following my rules, and it's the same briefing I use myself to keep up with AI without burning time.
Anthropic says it is approaching a capability it calls RSI. That is the point where an AI gets so good at training itself that it sharply accelerates the growth of its own abilities. The company is openly naming this as one of the key thresholds it watches.
OpenAI describes how it built Pando, a Git-compatible backend. It is designed for the scale of development that its Codex tool makes possible. The goal is to handle a volume of code and changes that a standard Git setup struggles with.
OpenAI is merging the previously separate Chat and Work tabs in ChatGPT. After the change, ChatGPT and Codex will draw usage from a single shared pool. For users it means a simpler interface without switching between two worlds.
Anthropic has updated Claude Code’s system prompt so people cannot use it to re-create existing video games. Making your own games is still allowed, the change targets clones only. It responds to a wave of AI decompilations and ports of retro games.
Simon Willison has released version 0.37 of his llm CLI tool and Python library. It gathers roughly two months of new features and fixes. It adds the gpt-6.1-sol and gpt-6.1-terra models plus a command for more easily adding models through OpenAI-compatible endpoints.
The New York Times, along with other publishers, has sued the maker of the Brave web browser. It accuses the company of unlawfully scraping content to train AI and of letting users bypass paywalls. The case adds to a growing wave of copyright disputes around AI.
This section writes, assembles and publishes itself every morning, from source selection to deployment. If you'd like a quiet helper like this in your company, let's start with an initial consultation. The fee is credited toward the project if we work together.