AI: misinformation about Anthropic Claude Mythos, Nicholas Carlini, and FreeBSD
Recently posted to Artificial Intelligence Made Simple by Devansh:
He misinforms his readers:
… Nicholas Carlini published a 15-round reverse shell exploit here, …
– false. Devansh:
- links to the post-CVE work by Calif
- misattributes the work to Carlini.
Carlini used Claude Mythos before publication of CVE-2026-4747. The work by Anthropic is properly described here:
- Claude Mythos Preview \ red.anthropic.com (2026-04-07)

