Comment by jackfranklyn
4 hours ago
Similar path here - studied physics, worked in accounting/finance for years, hadn't shipped code in forever. The thing that clicked for me wasn't the AI itself but realising my domain knowledge had actually been compounding the whole time I wasn't coding.
The years "away" gave me an unusually clear picture of what problems actually need solving vs what's technically interesting to build. Most devs early in their careers build solutions looking for problems. Coming back after working in a specific domain, I had the opposite - years of watching people struggle with the same friction points, knowing exactly what the output needed to look like.
What I'd add to the "two camps" discussion below: I think there's a third camp that's been locked out until now. People who understand problems deeply but couldn't justify the time investment to become fluent enough to ship. Domain experts who'd be great product people if they could prototype. AI tools lower the floor enough that this group can participate again.
The $100 spent on Opus to build 60 calculators is genuinely good ROI compared to what that would have cost in dev hours, even for someone proficient. That's not about AI replacing developers - it's about unlocking latent capability in people who already understand the problem space.
>That's not about AI replacing developers - it's about unlocking latent capability in people who already understand the problem space.
Feel like forums have turned into a grand Turing Test.
Turing Test is not really science (an infallible test, measurable outcome). An AI might never be able to pass TT for all humans. Just gets to be a high-def AI. Makes TT a technology.
True, as a threat to PM. Product management can't vibe their way out from a lack of domain expertise.