问HN：大型语言模型API的成本在一年内会变得微不足道吗？
嗨，HN。我们正在管理我创业公司的成本，而到目前为止，我们最大的开支是对Anthropic、OpenAI等公司的调用费用。我们考虑过建立自己的开源模型，但考虑到我们甚至还没有找到产品市场契合点（PMF），所以决定不值得去做。
不过，从乐观的角度来看，我注意到过去几年大型语言模型（LLM）的令牌价格大幅下降。你认为如果这种趋势持续下去，最终会变成微不足道的开支吗？还是说我们将永远被这些基础模型公司宰割？就像云计算（AWS、GCP等）的情况一样。
查看原文
Hi HN. We’re managing costs at my startup and by far our largest spend is on calls to Anthropic, OpenAI, etc. We’ve considered things like spinning up our own open source model but decided it’s not worth it considering we don’t even have PMF yet.<p>Optimistically though, I see that token prices to LLMs have been going down a lot in the past few years. Do you think if this continues that it’ll eventually become a negligible expense? Or do you think we will forever be gouged by these foundation model companies? (: Much like how cloud computing has went (AWS, GCP, etc.)