Users don't want this shit, startups do.
I also think you grossly overestimate the average user's abilities. They literally cannot tell what is possible and what is not. They don't want simple, they want something to solve everything, as easily as possible.
I don't fully agree with the person you're responding too, but it feels like you're making their point for them with this. Users are engaging with a tool or some content for themselves and today. Startups are the ones dreaming of standing among the "largest, most successful websites" someday.
Users don't care how efficient or inefficient your development process is, or how much technical debt you carry, or how many concurrent users or requests you can run, or how scalable you are if things go well. If your service stands up and works for them today, none of those things are relevant to them at all.
That's not to say that the startups concerns aren't critical to making sure that the service is still capable tomorrow and the day after, and that the company doesn't collapse under bad process and inefficiencies, and that the service maintains utility rather than becoming data and stale. They're valid concerns.
But the users are so far removed from those concerns that it can be legitimately frustrating when pursuit of those long-term concerns degrades their immediate experience.
And that's it. If you want to ever reach those long-term goals, you need robust tooling.
Clearly this means that office politics is good for your company
Largest economies in Europe have the oldest and most drafty, poorly insulated housing stock. Therefore shitty housing stock must be good for the economy.
Largest economies in the world have the most pollution. Therefore pollution must be good for the economy.
Backwards reasoning.
He argues tooling is a necessary but not sufficient* condition to performant web apps that solve meaningful user-facing problems
You argue he claimed correlation is causation
*Edit: had necessary and sufficient backwards
The answer tends to be "wellll I hate most of these features anyways, let's get rid of them and then it can be as simple as HackerNews!" But obviously, millions of people use those features every day and like them. The tooling solves a real problem.
Desktop and Mobile application are reasonably robust, have very complex software and could do everything Facebook does trivially. All this 'robust tooling' has evolved because we are trying to shove an application into a browser, and despite decades of effort. it still kinda sucks.
If Apple, Microsoft and various distributions of Linux pulled the finger out of their collective asses and agreed on a half-decent, cross-platform GUI software package in 2005, none of this JS madness would exist.
And cross-platform APIs, cross-platform libc, cross-platform networking... and then the same shuffle on mobile, making sure everything is compatible? Supporting x86 and ARM?
The web does suck, there's no getting around that. HTML and the DOM are garbage even at their original purpose, let alone at writing applications, and everything on top of that is a kludge on a kludge on a kludge.
But let's not pretend dealing with OS APIs from the 90s is much better. We've been trying to get away from that as early as we could with e.g. Java.
I work in Fintech. There's a lot of tooling to make JS-as-a-language more robust, including TypeScript and some of the best linters, debuggers and introspective libraries available in the entire tech industry.
The only reason JS isn't used more than it is, is because its robustness is young, not non-existent.
This is what browsers are. You are describing browsers. They run on any device, on any architecture, and nowadays even on the edge. This isn't just about GUIs. I can't stand this lazy mentality of "pfft, I know better than these billion dollar companies, have they tried just making it good???"