Skip to content

Anthropic investigating unauthorised access of powerful Mythos AI model

ft.com
17 pointsfinghin8 comments
On HN

Comments

This is a puff piece designed to build excitement around Anthropic's products.

Perhaps, but it also introduced a lot of low hanging fruit about the security übermodel not being secured.

If this builds excitement for Mythos then we really are living in the weirdest of timelines.

This strategy of making unverifiable claims to possess a super-weapon is one of the most dystopian marketing initiatives ever. Makes Fritz Haber look like a deeply-committed humanitarian. Wild.

The claims are verifiable and the Firefox team published one verification just yesterday (https://blog.mozilla.org/en/firefox/ai-security-zero-day-vul...).

Open source project dependent on VC funding models verifies claims of VC funded corporation.

I tire of AI writing marketing puff peices for AI companies which are mostly read by AI. It's like an ouroboros of slop.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.