Hacker News new | ask | show | jobs
by taeric 1755 days ago
It is almost certainly as bias trained. Such that if it is not stupid iterative with the witness, it is probably less likely to converge.