https://x.com/kevinroose/status/2041577176915702169

NEWS: Anthropic’s new model, Claude Mythos, is so powerful that it is not releasing it to the public.

Instead, it is starting a 40-company coalition, Project Glasswing, to allow cybersecurity defenders a head start in locking down critical software.

Source: Thread by @kevinroose

It seems like the model has definitely caught some very interesting, old and potentially ruinable security holes. However, I am surprised that people are surprised by this. We need to only look to the past few weeks: when you have a tireless pattern matching bot looking for insecure patterns in code, you are bound to find the bugs that humanity left behind, especially in code that’s not been actively maintained in a really long time.

So, I am surprised that people are surprised. I am impressed at Anthropic parking some strategy credit here. They’ve leveraged the moment to gain some brownie points with the community and improve their brand.

Kudos for a win-win.