ParentFull threadwmf·there’s no path to profitability for the buyers of these chips given the cost and capabilities of the current LLM architecturesDidn't Anthropic say inference is already profitable?View on HN