Ethan Mollick：关于 AI 行业的一件需要了解的事情是，拥有前沿模型的实验室可以发布半成品产品，而这些产品出奇地好用，因为 AI 能够自…
关于 AI 行业的一件需要了解的事情是，拥有前沿模型的实验室可以发布半成品产品，而这些产品出奇地好用，因为 AI 能够自行解决问题并即兴发挥。这就像在产品中内置了一位前线部署的工程师和客户服务代表。
他们发货的任何产品的一个破损的早期版本仍然比许多时候精心打造的传统产品更好，因为大型语言模型（LLMs）在处理各种各样的事情上只是异常有效的工具，而人们和公司正在学会只是顺其自然。
AI Labs 的每个应用都有奇怪的毛边，文档不足，有时会崩溃，而且毫无通知或更新日志就进行更新，界面也随时大改……
……但它们背后的模型真的很棒，所以每个人都只是耸耸肩，继续使用它们。
对照原文
A thing to know about the AI business is labs that have frontier models can release half-built products and they work surprisingly well because the AI can just figure stuff out and improvise. Its like including a forward deployed engineer & customer service agent in the product.
A broken early version of whatever product they are shipping is still better then what a well-built traditional product a lot of the time because LLMs are just unreasonably effective tools at doing a vast array of stuff and people & companies are learning to just roll with it.
Every one of the apps from the AI Labs have weird edges and are underdocumented and break sometimes and update without any sort of notice or changelog and change UIs on a dime... ... but the models behind them are really good and so everyone just shrugs and keeps on using them.