Discussion about this post

User's avatar
ToxSec's avatar

literally all these are interesting stories. great round up.

‘OpenAI specifically accused DeepSeek of leveraging a technique called “distillation” to massively probe ChatGPT and other Application Programming Interfaces (APIs) for proprietary information. ‘

this is what i’m seeing too. these type of attacks make training your own models way faster and more efficient. i’ve been seeing stories of anthropic aggressively blockings connections when they detect this kind of activity, but im sure it’s hard to find at scale.

No posts

Ready for more?