Y
Hacker News
new
|
ask
|
show
|
jobs
by
msoad
517 days ago
I'm not saying that never has happened. maybe they trained against openAI models but they are letting anyone to train from their output. I doubt they had access to GPT models to "distill"