ai can make it quicker to get from an idea to a working prototype. that makes another part of the work more visible: deciding what is worth building.
the engineer i find most interesting in that environment understands the user problem, makes product tradeoffs, and stays involved after the code ships. that's how i think about a product engineer.
they own a loop: user problem → proposed solution → implementation → measurement. ai can help with the implementation. the judgment still has to come from someone who understands the problem and can assess the result.
what that looks like in practice
consider a hypothetical support dashboard. a request arrives to add an export button. it is easy to jump straight into building a csv download.
a product engineer first asks what the export is for. if the user needs to send the same weekly summary to a colleague, a saved report might be more useful. if they need to analyze data in a spreadsheet, the export may be exactly right.
the next step is a small, testable version. decide which fields belong in the report, check access permissions, and watch someone use it. after shipping, measure whether the task became easier: fewer manual steps, fewer errors, or less time spent preparing the report.
the code is part of the answer. understanding the workflow tells you which code to write.
technical depth still matters
i don't think product sense replaces technical expertise. reliable systems need people who understand architecture, security, performance, and the consequences of a shortcut. ai-generated changes need that scrutiny too.
what i want to emphasize is the connection between those skills and the outcome. an engineer who can explain a technical tradeoff in terms of the user's task helps a team make better decisions.
for small teams especially, there is value in people who can move between a user conversation, a prototype, and a production issue. that doesn't mean every team should shrink or every engineer should have the same role. it means ownership can extend beyond completing a specification.
build the thing, understand why it matters, and come back to see whether it helped.
inspired by the rise of one-pizza engineering teams (opens in a new tab).