4 ms·

I'd really like to see improvements like these: - Some technical proof that data is never read by open ai. - Proof that no logs of my data or derived data is sa

• anematode 6mo ago
  I don't think this is technically possible without something like homomorphic encryption, which poses too large of a runtime cost for usage in LLMs
  • I_am_tiberius 6mo ago
    They don't even try to proof it another way.