"not a single vulnerability was found by a US frontier model."

6 pointsposted 9 hours ago
by int32_64
(twitter.com)

3 Comments

thisisnotauser

6 hours ago

Those free models won't be able to keep up so easy now that the actual frontier models aren't giving away all the answers anymore.

Razele

3 hours ago

answers are what llms do

_aavaa_

5 hours ago

This is cope.

All of the Chinese labs are publishing papers describing their models and each one contains actually novel approaches and solutions. In a few months when they surpass OpenAI and Anthropic, what will be the excuse then?