GeistHaus
log in · sign up

Mythos finds a curl vulnerability

daniel.haxx.se

yes, as in singular one. Back in April 2026 Anthropic caused a lot of media noise when they concluded that their new AI model Mythos is dangerously good at finding security flaws in source code. Apparently Mythos was so good at this that Anthropic would not release this model to the public yet but instead … Continue reading Mythos finds a curl vulnerability →

1 page links to this URL
The harness is the product, not the model

Cloudflare’s Project Glasswing write-up landed today and the Hacker News thread is mostly arguing about whether the prose was written by Mythos or by Opus. It is a fair complaint and an irrelevant one. The diagram halfway down the page is the actual deliverable, and almost nobody is talking about it.

0 inbound links article en ai cloudflaresecurityagentsclaude