ALLevie 写了一篇长文说 token 成本正在成为他接触的任何企业最热的话题之一。这对 AI 整体来说非常利好，因为意味着这些系统正以前所未有的规模被使用。
这催生了应用 AI 层的另一种差异化：model routing。当 token 占了任何一个工作流成本的很大一部分时，公司就会想确保钱花在最有效的 token 使用上。前沿智能在高端任务上（编码、法律、金融分析、医疗）始终重要，但你可以把个别任务剥离给更便宜的模型。
拥有最好的 evals、最强的路由能力、以及商业模式与客户财务目标一致的公司，会处于很好的位置。404 赞。
Levie：token 成本是企业最热话题。AI 被以前未想到的规模使用。Model routing 成为差异化。最好的 evals 和路由能力的公司赢。404 赞。
Levie writes that token costs are becoming one of the hottest topics for any enterprise. This is very bullish for AI in general because it means systems are being used at a scale that wasn't contemplated before.
This gives way to another form of differentiation for the applied AI layer: model routing. As tokens take on a significant amount of cost, companies will want to ensure dollars go to the most efficient use of tokens. Frontier intelligence will always be relevant for high-end tasks like coding, legal, financial analysis, healthcare. But you can peel off individual tasks to lower-cost models.
The companies with the best evals, the best ability to route workloads, and business models aligned to customers' financial goals will be in a great position. 404 likes.
Token costs are becoming one of the hottest topics for any enterprise I talk with right now. It's very bullish for AI in general because it means these systems are being used at a scale that wasn't contemplated before.
As tokens take on a significant amount of the cost of any given workflow, then companies will inevitably want to ensure that their dollars go into the most efficient use of tokens for the particular job at hand.
The companies with the best evals, the best ability to route the workloads, and those that have business models directly aligned to customers financial goals, will be in a great position.
