I recently had a conversation where I discussed my recent experiences with AI Coding. As we talked about my seemingly contrarian experiences with agentic coding — it’s still not a great experience, and still doesn’t feel productive (to me). I explicitly clarified:

It’s not the quality of the code, it’s achieving the outcome at all.

This steered us onto the topic of “quality” of output, and they said something closely approximating:

If it works, and it did it faster than a human would have, does it really matter? Ok, it’s not how you might have done it — or how sustainable you feel it may be. But unless you have concrete reasons why it’s not ‘good enough’, does it really matter if it works?

Fam, this got me. My whole career flashed before my eyes. And those of every single developer who has ever been told ‘write better code’, ‘write cleaner code’, or been put through the ringer of an interview where their code was ‘messy’ or ‘inefficient’ and they didn’t get the job.

Now we’re saying it’s OK? Now?

This is with the ‘background radiation’ that ‘architects should be writing more code’. Bro, what. I have to write more code so I stay sharp? But the thinking sand is YOLO? Why do we even bother with in-depth PRs? Why do we follow style guides? Why do we slow it all down, to make sure we get it right when it’s a human, but maybe it doesn’t really matter when it’s a robot?

This person wasn’t being douchey, or aggressive. It was a totally reasonable conversation. But in that moment — that singular moment — I saw the matrix.