> then that's a problem with the original question - not the solution itself
I think there's a good counterexample to this:
Atiyah/MacDonald proove the Nullstellensatz ultimately by using some trick involving determinants.
They give a very nice theoretical treatment of the content and context of the theorem. But the proof at one crucial point uses techniques that live conceptually outside of this context: While its possible to see that the argument is sound, it does not give a good explanation of _why_ it's true within the context of the theorem.
(You could of course argue that they did not give enough context
... but that's exactly my point: the trick makes the proof work but hides the explanation)
I think trying to iterate on a "spectral decomposition of your intent" - slowly working on increasinly refined breakdowns of what are the different aspects of your project are - both on the domain- and also the technical level; aka requirements and architecture. And then don't directly iterate on the code but rather regenerate/update the codebase based on the new intent and the old codebase... And a decomposition of the whole thing in terms of optics (open lenses, etc) where the decomposition respects the "spectral decomposition".
> A ProseMirror 2.0 with an incompatible interface would amount to the same but make it ambiguous what people mean when referring to ProseMirror. Trying to graft stuff on in a backwards-compatible way as an 1.x version would produce a compromised win32-style mess.
> If you look longer term, a society that does this will end up a lot poorer than one that doesn't.
Got data to back that up?
(On the serious level, I'm really curious; But on the polemic level I'll call BS - I highly doubt there ever was such a period in any capitalist country ever)
But I think there's a middle ground: You can definitely use GenAI to bring yourself to the page.
But that requires effort that goes beyond "draw me a pelican riding a bicycle".
I've used to create abstract art/(or "images that look like abstract art" if you prefer - let's not get distracted by this branch of the discourse) using midjourney: getting the AI to output something worthwhile would usually take me hours of iterating over a prompt - entering a feedback loop until two things happen: first, congruence between your intent and the output (both change during iteration! ); second, the output stabilizes with growing prompt length. and so generating the output turns from a slot-machine into something deliberate and personal
(lots of caveats of course but i think it's a worthwhile perspective)
PS -- What I forgot to mention: Its usually a hard fight to get out of the slop zone; The midjourney models have very boring default aesthetics and styles of composition (insultingly boring!)
> A vibe coder is someone who wants to test an idea by generating software as a prototype. A software engineer is someone who thinks about the entire software development lifecycle.
I don't think it's such a simple dichotomy; And dismissing the possibilites of agentic coding as inherently non-SWE is rather short-sighted: You CAN use agents as a software engineering tool.
It's just that it's often misaligned with the processes we're used to. But that does not mean that LLM-agents a bad tool.
I think there's a good counterexample to this:
Atiyah/MacDonald proove the Nullstellensatz ultimately by using some trick involving determinants.
They give a very nice theoretical treatment of the content and context of the theorem. But the proof at one crucial point uses techniques that live conceptually outside of this context: While its possible to see that the argument is sound, it does not give a good explanation of _why_ it's true within the context of the theorem.
(You could of course argue that they did not give enough context ... but that's exactly my point: the trick makes the proof work but hides the explanation)