Nolan Lawson offers an approach to using AI in code that cuts against the usual trends, use it to go slow. Have AI take a look at your code, analalyze it for bugs from several different angles and then address the most critical bugs right away, under advisement.
When I use this technique, I haven’t necessarily seen my velocity go up. If anything, the review process often finds pre-existing bugs, so I end up on a tangential side-quest where I’m writing unit tests and fixing subtle flaws that pre-date the PR. This is the opposite of the “10x productivity” slop-cannon style of development that most people imagine when they think of vibe coding, but I find it very satisfying.
This doesn't necessarily mean being less efficient. In fact, this is a way of working most coders dream of. Instead, it's a way of moving slower, in order to go fast. It's something I've been thinking about in the work of my team as well.