Imagine doing AI development in waterfall. You spend weeks writing your prompt, when you think you have it perfect, only then do you submit it to the AI. Then you wait a week or so, and see what it produced, expecting it to be exactly what you wrote.
Or, do you tell it the basic functionality you want, test it out, then add feature after feature that you want, sometimes dropping them and sometimes adding new ones that you thought of as your worked.
I don’t think that’s what it would look like at all. The first stages would be cheap - mostly requirements gathering and research, but a bit more focus on that. A bit more time would be spent up front, but then you’d see multiple proposals being built, from
that multiple plans being built, and finally multiple implementations to choose from. You might see A/B testing of multiple implementations or even products, and then a decision on which to pursue. You could move in multiple dimensions concurrently.
I’m not sure this is agile. I’m not sure it’s a waterfall.
We’ve got bounds on our infinite typing monkeys, but they increase every day
> Or, do you tell it the basic functionality you want, test it out, then add feature after feature that you want, sometimes dropping them and sometimes adding new ones that you thought of as your worked
the problem with this is long term maintainability. it works - and the engineer understands how it works - but a) the AI does not prioritize cleanup/organization/naming, and b) there's a blind spot/boiling frog type of phenomenon that can prevent the engineer from spotting the growing problem. the codebase becomes recognizable only to them. the engineer sees all features working, all bugs fixed, 90% test coverage, and submits it for a PR.
the engineer tasked with reviewing the PR will treat it as slop.
The vast majority of "AI is changing everything!" takes I read say more about people's fundamental misunderstandings of the software development lifecycle (the real one that companies actually do, not the one that people think they do or what companies say they do) than about anything AI is going to change about software eng.
If anything, their solving the complete wrong problems and being blind to the actual problems is probably a reason AI won't actually result in any real, top-level appreciable gains in shipping speed.
Waterfall came out when hardware and software had to be developed together, and appealed to traditional Engineering practitioners. You are right though, when the hardware constraints went away, software (more code) was cheaper and easier to ship in increments and iterate. But feature-rich products were still difficult to ship - and you had to pick and choose what things to spend your time on.
The SaaS-pocalypse is occurring versus investors don’t believe that to be true anymore.
I think they will still be wrong because ultimately people want people (particularly experts) to be held accountable for things - shipping high stakes software, running company ERPs/CRMs, and more.
Honestly posts like theirs are just indicative of someone who never understood their job/role.
People throw out terms like agile or waterfall, shit on agile etc. probably because they work at some worse than mediocre place let alone ever done their own thing.
The real problem is that company decisions will never be made as fast as software can be written (and rewritten) now.