At this point it’s clear that Google either isn’t trying or can’t produce a high quality LLM right now. A bunch of H1Bs at Xitter did it way faster starting at

At this point it’s clear that Google either isn’t trying or can’t produce a high quality LLM right now. A bunch of H1Bs at Xitter did it way faster starting at 0, though Jensen probably supplied all the GPUs Musk needed and there were likely no restrictions about safety or training on libgen.