Moving up the stack

Yesterday I was working with a colleague on an idea. We needed to build out a new workflow for our users, and we'd already gotten a great start from prior work by another designer.

As we were wrestling through the interface, moving back and forth, changing things, and talking to Claude——I was struck by something.

The work we were doing was in direct purpose of serving our end users (who we're planning to talk to tomorrow by the way, actually getting software in front of those intended users is so critical).

At one point we had an idea to change a part of the design by switching out a preview state; making it dynamic instead of static. In our peer design session we described the need to Claude and waited for it to try out the new idea——and I had to laugh for a moment. That ask, that thing we requested, was something I've actually built out in the past with html/css, and at a good clip would have taken me half a dozen solid hours.

Now, we just needed to wait a moment and have Claude run the idea.

And that got me thinking.

Back when I got into product design, much of our time was spent creating to try and solve a problem. We'd iterate, design, code, and wrestle through hours on end; until we got a semblance of a prototype ready to go for review. Then we'd take feedback, run it through again——days, weeks, months of effort; and at the end of it we'd have to just wait and see if the numbers proved out our hypothesis. Yes, we could test with users ahead of time, make sure we were avoiding friction points, all the right things.

That's where the whole idea of design sprints was so crazy; taking an idea on Monday and showing to customers by Friday; it was revolutionary. Now that's doable any day of the week.

Product design is, at the end of the day, a guessing game mixed with intuition and science.

Now, in our peer work session, my colleague and I could wrestle through things in minutes that would have taken endless hours before.

I feel loss at that, at no longer needing all the manual time (and skills) that it would have taken me in the past. But I'm also excited about what that means——that I (and we) can spend time thinking about what the software we're building will do, what it means, who it's for, why it's working, how it's not working.

We can take those ideas, and in minutes show them to the end user, get feedback, change things live, and see what's working and not working.

That's exciting.

Yes, there's some things that no longer make sense to do with the advent of AI. But now my time can be spent thinking through the effects of our work, talking to users, analyzing data, playing with dozens of ideas, and then dogfooding to make sure it behaves the way we expected.

All that helps move up the stack toward ownership of the problem, toward helping to find a solution.

Earlier this week I ran into a small friction point in our software. I was reviewing a bunch of projects and wanted a way to see them all together with their statuses.

In the past that could have been a conversation with the team to talk through a design, plan out where development would fit into the schedule, and then move it through our process.

But now, with all the tooling we have setup, I could take the problem, spin up my localhost with Claude and try and fix it in code; validate that the new solution works, and then share it with the team for review and feedback.

A million paper cuts fixed by team members who are empowered to find the problems and solve them.

That's exciting, that's motivating, and it's meant the things I've given up doing have been replaced with time spent actually fixing and solving problems.

Now, one caveat on the small changes I made——I made sure that I understood the code that had been modified, then worked with a team member to review it. And the software wasn't user facing, it was admin software that directly benefitted the workflow of our team.

But that aside, it's pretty amazing what is possible now with the tools at our disposal——and how if used correctly, they can enable more thinking and more problem solving than ever before.