Yeah which is why this is a dumb statement from Amazon. But then again I don't expect C-suite managers to really understand the intricacies of their own companies.

Depends on the use case. Training local llms is a lot cheaper after Galore and there are ways to get useful local models with only a moderate amount of effort, see e.g. augmentoolkit.

This may or may not be practical in many use cases.

24 months is pretty generous but no doubt there will be significantly less demand for junior developers in the near future.

Of course not. It will be more expensive and they'll still have to pay developers to figure out what's wrong with their AI code.

Possibly. It’s hard to know without seeing the numbers and assessing output quality and volume.

Also it’s not unheard of that some bigwig wastes millions of company €€ for some project they fancy. (Billions if they happen to be Elon)