safety
2 posts on safety, newest first.
- Anthropic's IPO filing and the RSI risk lineAnthropic has applied to list on Nasdaq as ANT, and its amended filing warns that future Claude models may soon greatly speed up AI research.Issue 5
- OpenAI's math claims and new science modelsOpenAI puts AI-made research results up for review, locks down its training runs, and new models target memory, DNA and medicine.Issue 3