Post #4149595
2026-07-27 23:12 UTC
@glyph@mastodon.social Oh I think it remains dangerous. We know open weight models can find vulns just fine.
https://aisle.com/blog/aisle-matches-anthropic-mythos-on-freebsd-zero-days
And let's assume that capacity maps roughly onto offensive capability.
So between rapid discovery of new vulnerabilities, exploitation of existing ones, and orchestration of attacks, the danger doesn't go away.
Also when the subsidies end, a collapse will mean a collapse in the price of hardware for local compute.
I really think we'll have to deal with the consequences of unleashing this poison for some time.
Replies (2)
-
@glyph@mastodon.social 2026-07-27 23:14
@mttaggart@infosec.exchange using a chatbot for attack orchestration still feels like strapping a shotgun to a tilt-a-whirl and hoping it'll mostly go off while pointed at the enemy
-
@glyph@mastodon.social 2026-07-27 23:54
@mttaggart@infosec.exchange I am inclined to trust your opinion, so this isn't me saying you're necessarily wrong, but just for comparison, today I'm looking at this: https://support.apple.com/en-us/128067 and that is an enormous, record-setting number of vulnerabilities; it would clearly be foolish to say that nothing is happening, something is definitely happening. but there's also a huge, record-setting amount of resources being poured into this right now. like an apocalyptic amount of resources