Discussion about this post

User's avatar
Rene's avatar

Why is the subsidy calculation based on the price of the mini model, whereas the hardware described is that required to run the large model?

Expand full comment
John Allard's avatar

The most fun part about working at a place like OpenAI is seeing how wrong random substackers are when making claims about training and serving frontier models. The worst part about working at OpenAI is not being able to correct people without divulging inside knowledge.

I’ll say this: frontier LLM providers are running fantastic businesses. Your math is very wrong

Expand full comment
2 more comments...

No posts