Just like we didn't forget how to build lunar vehicles, we just decided that it wasn't worth it in its current state. It's still very possible to do another moon landing, it will just cost a ton of money that corporations won't risk and governments won't spend without a clear political goal like the old space race.
Space technology isn't degrading, it's becoming more cost-oriented. A SpaceX rocket only reached orbit but it was relatively very cheap. Cost efficiency helps space technology reach more people and create greater impact, ultimately making the technology more valuable to society.
It's the same with the web. You say that the web is bloated "instead of trying to improve people's lives" but the reason the web platform is so heavy is because the platform is rapidly adding features and APIs and tools that make it easier for developers of all skill levels to make complicated applications. Letting more developers make stuff in less time improves people's lives. Maintaining backward compatibility at the cost of bloat also improves people's lives.
Society has been constantly evaluating the web and it has consistently reached the same conclusions: RAM is cheaper than developers, a slower platform with a bigger audience is worth dealing with over multiple native platforms, and weird legacy stuff is fine if it means we can run 15 year old websites.
Infrastructure projects are important and what enables society to grow. Moon landings were plain awesome and had important effects on both the zeitgeist and growth of technology. That we're not incentivized to grow, or even maintain them, says a lot about us, and not much of it good.
And I do think GP is right when they say: "instead of trying to improve people's lives, we're just making things needlessly complex, buggy, and bloaty." Any actual improvement to people's lives that falls off this process on the web is incidental. The conclusions you say we've reached, like "RAM is cheaper than developers, a slower platform with a bigger audience is worth dealing with over multiple native platforms" are not conclusions. They're excuses. Excused for doing shitty engineering in order to get to the market first. And anyone trying to do the things the right way moves too slow to keep up with the hype train.
Ultimately, the complaints about the way technology develops today stem from this: some people see the suboptimal solutions created by our incentive structures, and also notice there are much better solutions available only a small step from the path we're on - solutions we could reach if only we could relax the market pressures a little bit. This is true as much for the web (e.g. for interoperability issues and bloat issues) as it is for transportation (where electric cars finally got viable after one player basically started to shove a product into peoples' throats until the market forces gave up and accepted it).
Today, for some reason, everyone thinks that writing software for a platform on a platform on a platform with no interconnecting standards or protocols is a great idea. Instead of trying to improve people's lives, we're just making things needlessly complex, buggy, and bloaty. You need 8 gigs of ram minimum just to browse the web, and god forbid you want to do something like back up your data. I was just in a meeting where nobody could get Google's video conferencing to work.
Instead of building internet technology, we build web technology. The web is harder to write software for. It's an ephemeral, inconsistent, difficult thing. But if we try really hard, we can turn the tide on the unnecessarily complex box we've forced ourselves into.
The proposed fixes are good ideas, but they're bandaids on axe wounds. There are much deeper problems going on that won't be fixed by a feature add or a pivot. We need a re-evaluation of all internet-based technology, how we develop it, how we incorporate it into our lives, and what we want as a society from it.