I've become insanely happy with how I write end-to-end tests now and run them with Aspire.
This is remarkable because I used to hate e2e testing. In traditional development, it's the vegetables of software engineering. Everyone knows they're important. Nobody enjoys writing them. They're brittle, slow, and a pain to maintain.
With AI, that entire dynamic flipped.
I describe the user flow in natural language. The AI generates the test code. I run it, see what failed, iterate. The cycle that used to take hours takes minutes. And because the AI can generate variations easily, I can cover more scenarios than I'd ever bother to write manually.
But the real magic is Aspire. It turns out Aspire is perfectly designed for this workflow. You can spin up your entire distributed system, run tests against it, tear it down. The framework tells you what failed, why it failed, and what might fix it. The agent can listen to that feedback and adjust.
I've gone from viewing e2e tests as a necessary evil to actually enjoying the process. That's not a small shift. That's the sign that a fundamental constraint has been removed.
When testing becomes fun instead of tedious, it means something structural has changed about how we build software.
Part of the #100DaysToOffload documenting agentic development in 2026