3 ms·
Relishing new Fickling features for securing ML systems
- nonrandomstring 3y agoOh this is going to get interesting, grab the popcorn. OpenAI: "Our stuff is safe. So long as normal, good people only use it for normal, good things." NYTimes: "But what about bad people using it for bad things?" OpenAI: "That could never happen with our 'safeguards' ". NYTimes: "Challenge accepted." hack, hack, hack.... OpenAI: "Oh noes! Our precious stuff!! You cheated, Filthy, evil hackers."