AI could eat itself: Competitors (..) steal their secrets and clone them

[pogue]

I'm curious how these 'distillation attacks' work and what the prompts look like to do them.

[MagicMoonlight]

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.

[sigwinch]

I’m not sure there’s even an adaptive pattern in prompts. The article sounds like a n>1E5 shotgun.

[jadelcastillo]

In my country we say: he who robs a thief has a hundred years of pardon.

[titaniumrain]

just blame china when things don't work out for them. china should have blamed openai for stealing too many yellow-skin talented chinese :))))