Paying for LLM inference by the kilowatt-hour instead of per token

Y	Hacker News new \| ask \| show \| jobs

	Paying for LLM inference by the kilowatt-hour instead of per token (coinerella.com)
	4 points by willy__ 36 days ago

1 comments

AI is an energy problem.

Marketplace success will likely be determined by energy costs as much if not more so than the technical aspects of model development.