This summer I did a 3-month residency in Berlin. I attended a program called The Sandbox.
One of the most memorable days was a trip, organized by Janne, to the warehouse of a prominent logistics company. It handled tons of products from various vendors, and the size of the place, the number of packages, and the number of people working there were all really impressive.
But one thing impressed me the most. Much of the software inside this warehouse was built with either Lovable or Claude Code by non-technical people. They had made over 15 Lovable apps for themselves, without any engineers or IT. The apps ranged from the check-in system we used to register on arrival to very sophisticated dashboards that the people running daily operations built to see the status of the warehouse and come up with new ideas on how to improve processes.
I've spent the last 15 years mostly in San Francisco and New York. Most of my circle are engineers or otherwise highly technical people.
Engineers usually know how to build their own software from lower-level primitives, and they often don't like higher-level, easier tools. The canonical example here is Drew Houston's Dropbox announcement post, where the top comment infamously dismissed the product as unnecessary (because a user could build a system like this themselves) and as having no product or financial potential.
As a software engineer, I was skeptical about Lovable too. The warehouse was the moment I understood the hype and validation behind it (the value of Claude Code was clear to me from day one, because it's a tool for software engineers).
The first major insight was about who builds software. I do believe that people who have domain expertise and access to actual knowledge should be able to build their own tools. Before 2025 and the rise of Claude Code and Lovable, it was really hard, because domain experts had to find engineers and convince them to help.
The other major insight was quality. The apps they were using were better than 90% of the internal apps I've seen in my career. Famously, most companies neglect internal software. Until recently, you would have some brilliant backend engineer who would produce absolutely unusable internal tools, because they had no feel for UI and UX and didn't know or care how to make usable software.
All of that has changed. I believe we are entering an era of better internal software and more personal, custom software. It is no longer gatekept, and no longer built by people who have neither the domain expertise nor the incentive to make it better.