My current team has had a great solution to this, which is to re-build in parallel. Build the new system alongside the old one, including the reporting and integrations. You’ll find the edge cases pretty quickly.

For a minute I thought QuickTime was back…

I actually use Vim to write all my code, but without IDE-ifying it, just syntax highlighting and some navigation tweaks (with Sublime3 for help with bulk edits). For most of my stuff an IDE is overkill.

I was just being facetious by suggesting another commandline editor. ;P

Yeah, like Vim. Who uses nano to code?

This is just an extension of the larger issue of people not understanding how AI works, and trusting it too much.

AI is and has always been about exchanging accuracy for speed. It excels in cases where slow, methodical work is not given sufficient time already, because the accuracy is already low(er) as a result (e.g. overworked doctors examining CT scans).

But it should never be treated as the final word on something; it’s the first ~70%.

My tip, “developer empowered” means that the buck stops with Engineering leadership when it comes to dev work. They have to be able to greenlight work, and also red-light it.

That gives engineers the flexibility to build what they see is missing, while still having executive leadership (e.g. “Director of Engineering” etc) in the loop who can rein in rabbithole-development projects.

But the second you allow “Product” teams to dictate dev work, you will lose that developer empowerment, and they will always just be “the ones who build what Product tells them to”.

Never used it for a website or web app, but I’ve used it for getting scripts for managing my servers started. It always requires a lot of refactoring to make it actually usable. Since I know when the code is wrong, I can coax it through fixing the code itself, but it’s faster to just change it myself. I don’t think a non-coder would get a whole lot of useful code out of it.