An alleged Anthropic Mythos-class model checkpoint appeared on Hugging Face under username 'burner22' in a repository named 'numbat', according to a Hacker News thread. The leak suggests either a data breach or unauthorized model release.
Topics #
Sources #
- Social[HN thread](https://news.ycombinator.com/item?id=50326727)
- Official[Read article](https://huggingface.co/burner22/numbat)
Go deeper #
This intelligence is sourced automatically from public sources across the web and synthesised by the Prefactor AI pipeline. Stories are reviewed before publication.
source & further reading
getreadyforagents.com — original article
Google's spam filters miss AI-generated outreach email at scale, founder reports
AI agents deployed in production require structured process engineering rather than flexible ad-hoc workflows
OpenAI trains models on Daybreak user sessions despite explicit opt-out selection