> AI compresses implementation time for an individual engineer, but architecture decisions, design reviews, integration, testing, deployment, and production validation remain largely serial activities.

Integration, testing, deployment and production validation activities are also, depending on you setup, available for AI to work on. It depends on the type of stack that you have and how your hosting is set up but with an AI-friendly set up there's a lot of time to be saved here too.

▲

acdha 4 days ago | parent | next [-]

This is largely true to only the extent that you you have a machine-verifiable measure of quality. If not, you’re putting more pressure on the humans who can actually reason and understand business goals. This cascades: tests are what make LLM code generation work but if you have an LLM write the tests you are then pushing further down the road the point where you learn that the first LLM was going in the wrong direction.

▲

zombot 4 days ago | parent | prev [-]

GP: > With the current capabilities, models still need constant babysitting and course correction.

You'd be opening yourself up to even more babysitting. Would that really save time, or improve quality? I'm not convinced, to put it mildly.

	▲	andrekandre 4 days ago \| parent [-]
		well, that is a problem with automation in general i think; if you go too far instead of a simple process a complicated processes get encoded in the automation and if it hits an edge case or breaks for whatever reason the babysitting kills productivity (this is just my experience anyways) also, as far as quality goes, once something is automated that potential breakage puts a cap on how much improvement/quality you can achieve later... i'd guess llms are a bit more malleable there, but idk, i still see this issue even with skills and such...