I'm a CTO and I write code most days. Not production code. Not because I should be in the critical path. But because AI changed what "staying technical" actually means when you're running a team.

I run the technology function at a B2B SaaS company. Engineering, DevOps, InfoSec, sales engineering. I've got a Head of Engineering who runs delivery day to day, and a team of specialists who are better than me at every one of those things. My job isn't to ship features. It's to understand what we've got and keep us pointed at what's coming.

That used to mean reading architecture docs and sitting in design reviews. Now it means I can open Claude, point it at a repo I haven't touched in months, and have a real conversation about the tradeoffs in our implementation. I can evaluate an architecture decision before a planning session. Review a PR in a language I don't use daily. Write integration tests to understand how a service actually behaves, not how the docs say it behaves.

None of that pulls engineers off their work. None of it makes me the bottleneck. It just means the gap between "CTO who understands the tech" and "CTO who reads slides about the tech" got a lot smaller.

Everyone says CTOs should step away from the code. I think AI lets you step back in, if you're honest about why you're there.

I'm going to write about this. What actually works when you try to stay on the tools without getting in the way. What I've built, what broke, what the numbers look like.