←back to thread

566 points PaulHoule | 2 comments | | HN request time: 0.494s | source
1. ceroxylon ◴[] No.44492535[source]
The output is very fast but many steps backwards in all of my personal benchmarks. Great tech but not usable in production when it is over 60% hallucinations.
replies(1): >>44493453 #
2. mike_hearn ◴[] No.44493453[source]
That might just depend on how big it is/how much money was spent on training. The neural architecture can clearly work. Beyond that catching up may be just a matter of effort.