Coding agents and the growing 1% problem
2026-04-19
The mainstreaming of AI workflows will cause a massive hiring boom because the more AI-assisted work you do, the more work you create for yourself.
The “1% problem” with AI is that you can spend a couple of minutes, 30 minutes, or an hour and get something interesting and passable. But in order for you to really share it with other people in your organization and for you to be able to guarantee the quality of the experience, the accuracy of the data and so on, you have to put in orders of magnitude more work, 10x or 100x, than it took to get the first draft out.
To make something out of this that you could publish in your application and share with real end-users you would have to spend 100x more time and effort on it.
When you can do more with less effort you easily end up doing more and getting exhausted.
The other aspect of the 1% problem is that every feature you launch increases the complexity of your codebase, your product, and even your organization as someone will at some point be able to answer customer questions about it, market it and sell it and so on. You’ll also have to maintain it and further develop it as your codebase evolves. It’s a factor of at least 10x the effort it took to launch the first version.
If everyone builds their own agentic workspaces and workflows, adds a chatbot or AG-UI to their applications, then someone will need to maintain this. And this means new hires and why AI will result in a massive hiring boom, at least across software and likely across almost every industry though not at exactly the same time.
This also means that we need to go back to thinking before we build something because otherwise we will either end up with massive debt and increasing hiring needs, or churned users.