dex@dexhorthy我最喜欢🦄ai that works播客的一点，就是我和vaibhav意见不合的时候。 这次，我认为优秀的产品工程意味着：快速交付大量小东西、学习、然后迭代。 @vaibcode 的观点基本上是“如果你不是像@JeffDean那样提前规划整个系统，你就没戏了”。 你们怎么看——是优化学习速度，还是优化一次性搞定全局？
my fav thing about the 🦄ai that works podcast is when me and vaibhav disagree
in this case, I think good product eng means: shipping a lot of small stuff fast, learning, and iterating.
@vaibcode's take is basically "if you're not @JeffDean planning the whole system up front, you're ngmi".
WDYT - optimize for learning or optimize for one-shotting the whole thing??
来源：dex · x.com
我最喜欢🦄ai that works播客的一点，就是我和vaibhav意见不合的时候。 这次，我认为优秀的产品工程意味着：快速交付大量小东西、学习、然后迭代。 @vaibcode 的观点基本上是“如果你不是像@JeffDean那样提前规划整个系统，你就没戏了”。 你们怎么看——是优化学习速度，还是优化一次性搞定全局？
my fav thing about the 🦄ai that works podcast is when me and vaibhav disagree
in this case, I think good product eng means: shipping a lot of small stuff fast, learning, and iterating.
@vaibcode's take is basically "if you're not @JeffDean planning the whole system up front, you're ngmi".
WDYT - optimize for learning or optimize for one-shotting the whole thing??
来源：dex· x.com