undefined | Better HN

0 pointsexe341y ago0 comments

if you can do 100x more efficient training with open source, closeAI can simply take that and train a model that's 100x bigger/longer/more tokens.

0 comments

AKA why Unsloth is now YC backed for their even better (but closed source) fine-tuning.

j / k navigate · click thread line to collapse

AKA why Unsloth is now YC backed for their even better (but closed source) fine-tuning.

j / k navigate · click thread line to collapse