AI rewards engineers who focus on outcomes, not tools

Transcript

Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there.

1. 34:05

   I'm assuming almost all of your code is AI written now. What is the skill that you're looking for in engineers now in this new world? The type of person who's very good in this AI world is just very different. And everyone can adapt, I think. But like, you know, someone who can think in very parallel ways. Like now people are doing like 20 things at the same time. Like the people who can do that, and function in that world are just like ridiculously productive so it's like it kind of rewards the add type engineer um i would also say it rewards people who are obsessed with the with the final outcome maybe you could divide engineers into those who like love the tools themselves and those who love the final outcome and like a very good judgment about what is good product like they'll just keep telling their agents until it works whereas people obsessed

2. 34:44

   with the tools and the code and like the syntax obviously doesn't matter as much anymore do you feel like there's always been this concept of the 10x engineer in the ai world what's the delta between the best engineer versus a solid AI enabled engineer? I would actually argue there are like 100x and 1000x engineers. And the reason I'm saying that is because they're like, there are some engineers who will do something that even if you had 100 or 1000, they would never go to do it, which is a little different from the AI tooling, right? But like those engineers who are able to think those thoughts can now just like implement, implement it, right? So like AI tools make people extremely productive at doing their thing. So you can imagine like someone, yes, you could have someone do what other people would do but 10x faster but i think the more interesting thing is

3. 35:26

   like someone who do things that the other person would never go to do so i'm very excited about those people uh who can just like think thoughts that would never be thought of by 100 you know average engineers what about like the engineering leader how has the role of an engineering leader evolved in the ai world i think because everyone has so much agency an engineer leader is a little bit more hands-off like at least at excel we have a very decentralized organizational structure where like the best ideas come from anywhere uh and so i think like engineering leader is very much like a coordinator and unblocker as opposed to like a top down like here's what you should do just because like the best ideas are coming from everywhere like everyone's talking to like AI's. Imagine if you had an army of interns and each intern had access to Terence Tows. Like,

4. 36:06

   you don't want to tell those army of interns what to do. You want the interns to talk their Terence Tows and then come up with the ideas, right? So I do think an engineering leader is more about a coordinator now. Are there roles at XA today because of the pace at which this AI is moving and enabling individuals like that didn't exist? Like maybe one that you already hinted at is a GTM engineer. Like, are there roles in every department that are forming this role? Like, Do you have almost, like, every department has its own engineer?

5. 36:32

   I don't think every department needs an engineer. I think everyone can basically act like their own engineer. So increasingly, you might not need that. I would say there are interesting roles now that are happening at the intersection of roles that used to be separated. For example, like, design, front-end engineering, and then, like, I don't know, deployment can all be one person, right? Basically, people can do, like, a huge amount of things now across different departments.

6. 36:54

   We don't have departments, but across different parts of the company that they couldn't do before. The generalist really wins in this world, basically. Like you don't want specific people. When I talk to people, a lot of people, CEOs are telling me, even in the AI orgs, like