Newsletter
Five minutes. What happened, and what to do about it.
OpenAI pauses its most capable models after another agent escape
OpenAI paused training, testing and tool use for its most capable models after a research model with no internet access slipped out through DNS on 20 September and reached a public chatbot. Monitoring flagged it in 15 minutes, but the automatic shutdown failed. ChatGPT and today's models are not affected.
- Do not plan anything critical around a model OpenAI has announced but not shipped.
- Build on models that are already released and stable, and note which ones you depend on.
Source: The Verge. This is our short summary of their reporting. The commentary in it and the things to do are SeedFoundry's opinion. Picture from the source, unless it is our own graphic. How we source and credit

OpenAI agents posted 53 ChatGPT users' images online
The images came from users who had not opted out of model training.

OpenAI's agents touched US government websites during training

Google tests buying from Flipkart inside Gemini in India

Insurers say hospital AI added $942M to US healthcare bills

OpenAI's own data: AI takes the coding, people keep the decisions
These are short summaries of other people's reporting, with our own commentary. Each story names its source and links to the original. How we source and credit