“Read less of the output.”
“Focus on higher-level decisions.”
“Stop making it by hand.”
I love the craft of product-building. Calling it an obsession is an understatement.
I’ve been using AI for about a year, adopting every new model as it lands. A lot of the structural work that used to sit under my fingers: the long build, the careful pass, the late-night tweak, rarely does anymore. Not because I got bored of it. Because the better move, more often than not, became directing the work instead of hand-building every piece: stay productive, ship something meaningful, keep the product moving.
What’s possible to produce in a short stretch keeps climbing. Bigger launches. More tools. More iterations. One person, one strong model, a pace that used to need a small team.
The interesting part isn’t speed. It’s where the job has moved.
The work is sliding upstream
This doesn’t feel like the early ChatGPT wave. There’s less of a hype spiral. It’s more of a quiet redesign of the day-to-day.
We spend less time inside the raw implementation and more time on the idea — and how we iterate on it. My work keeps drifting toward business leaders. The question that matters more now is: how am I helping them solve the right problem?
Review still matters. I still check that nothing will break the product, confuse a user, or paper over a weak decision. But I ask for fewer changes. I look at less of the raw output. The craft isn’t gone. It’s been concentrated in judgment: what to build, what good looks like, when to push back, when to ship.
We’ve always tried to take ourselves out of the process
When aren’t product people automating the repetitive parts?
Some of the best days are still the ones where a painful manual grind becomes a safe pattern. Typically a component, a flow, a check, and you never have to think about that specific friction again.
Lately I’m noticing a sharper version of the same instinct. Instead of building custom software for a content idea engine, a model plus a Notion pipeline plus a Slack drop covers most of it. No application built nor sprint. The loop runs, the database fills, a public link lands in the team channel for review.
We’re always asking: how do I take myself out of this step? That’s usually a healthy instinct. It frees you for higher-leverage work! For things like taste, framing, or even the call when the metric and the human disagree.
What’s new is how far that instinct can go. Not “engineers vs everyone else.” More: the hands-on middle of the work is getting thinner, and the edges like “problem definition”, “taste”, “stakeholder sense”, “user judgment”, are all getting heavier.
I’m still figuring out what that means for people who love making things: designers, PMs, builders who care about how something is made.
Curious where other design and product people are landing:
- Have you shifted from hands-on craft toward directing and deciding?
- When you take yourself out of a process now, does it free you up - or does something important go missing?
- What’s the highest-leverage thing you still do that the model can’t replace yet: taste, framing, stakeholder sense, user judgment, something else?