Survey of developers experiences and opinions of AI tools

The gap between "96% don't fully trust AI output" and "only 48% always verify" is concerning. We're essentially gambling with code quality because verification feels like friction.

I've started treating AI suggestions like code from a very productive but junior developer - it needs review every time, no exceptions. The productivity gains disappear quickly if you're debugging AI-generated bugs weeks later.

Not sure what the selection bias for this report is, perhaps that we care about code and believe in the value of static code analysis. Some interesting results in there either way.