> If they're not losing money on inference, then why do they need to keep raising absurd amounts of money? I recently heard someone say that ~"state of the art

> If they're not losing money on inference, then why do they need to keep raising absurd amounts of money? I recently heard someone say that ~"state of the art LLMs are the most rapidly depreciating asset in history." This seemed accurate sounding to me. Anyone else have thoughts on this?