Hacker News new | ask | show | jobs
AI could eat itself: Competitors (..) steal their secrets and clone them (theregister.com)
12 points by myk-e 123 days ago
3 comments
pogue 122 days ago
I'm curious how these 'distillation attacks' work and what the prompts look like to do them.
link
MagicMoonlight 122 days ago
I’m pretty sure they just mean DeepSeek was paying to generate text, and then using this text as training data.

Which of course is completely wrong and totally hacking and illegal. Very different from using book torrents or people’s google docs to train your AI.

link
sigwinch 122 days ago
I’m not sure there’s even an adaptive pattern in prompts. The article sounds like a n>1E5 shotgun.
link
jadelcastillo 122 days ago
In my country we say: he who robs a thief has a hundred years of pardon.
link
titaniumrain 122 days ago
just blame china when things don't work out for them. china should have blamed openai for stealing too many yellow-skin talented chinese :))))
link