Lenny Rachitsky: evaluation

Transcript context

…I think it will get better at taste and judgment and design. Yeah, I think we might be holding onto that a little bit too much and saying, "Oh yeah, a designer or somebody will always know the best thing to ship or the best version of this." But I do think AI's sense of taste will get better. At the end of the day, someone has to decide what is actually going to get built and what actually matters. And when I think about people saying, "Oh, AI is just going to build this software for us," a lot of the hard parts of building software are actually not building it. If you think about the hardest times that you've had at work, Lenny, it's probably things like, oh, you and some other person disagreeing about what should go into this feature or what shouldn't go into this feature. And those things still feel like, yes, AI can weigh in, but it can't necessarily solve this dispute between you and somebody else. And so there is something about deciding what actually goes into the things we build, which I guess is taste in some way, but maybe not taste in the way we think about aesthetic taste or whatnot. There's some sort of, it's judgment around what to do next. Just watching how quickly AI took over coding, which, I think a year ago, definitely two years ago, most people are like, "I don't think so. I don't think AI will get this good." And that the best engineers in the world trusted so much, they're not even looking at the code anymore. That's where we've gone. It just made me reevaluate all these assumptions I've had about, okay, AI will never be as good as really good PMs, designers, at judging what is great and deciding what to build. But I'm just starting to think, I think it will get there. Even an example you've shared, it could give these two people trying to make a decision, "Here's all of the data you need to make a decision and here's why this is the right answer and just press yes, press one, and I'll go ahead and build this. " So I think to your point, I think we undervalue just how good it'll get at this stuff. Okay. So your sense is it'll get better, but your sense is we'll still need awesome designers to be involved, us and PMs, to help make these decisions, engineers, of course? Yeah. Yeah. I think someone will still have to decide, oh, we want to build this kind of product. Or given what the AI is presenting us, someone still needs to be accountable for the decision. The same way that even though Claude can write all this code for you today, it is still an engineer who's accountable for, does that code actually work? Does this actually make sense in the product? So I think there's that decision making/judgment layer, which feels like maybe one day will come when we won't have to do that, but it still falls on us. Yeah.…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.