It seems like you have to ignore all value derived from networks in order to come to the conclusion that software is no more powerful than in the 90s.

In the 90s I couldn't have met with my team, with members in Moscow, California, Pennsylvania, and Texas, in any reasonable way...today I can chat, including video and sound, on a whim!

Managing source code today is massively more productive than in the 90s. CVS (or, heavens forbid, RCS) on a central server was how it was done back then, if you had revision control, at all. It's not merely a better revision control system (git), it's the web-based infrastructure around it (github/gitlab/web-based CI/whatever). That wouldn't be possible on any platform that's less connected and less widely available than the web.

The rise of package managers is another massive productivity booster that maybe goes unheeded (we all love them, but I think their productivity value is wildly underestimated...how else can you add 100,000 lines of code, that probably works, in a couple of minutes, and reliably allow every member of your team to do the same?). Web technologies have enabled that. There's a reason npm has the largest package selection the world has ever seen, and I think it's the massive interconnectivity of the web platform. (This feels sort of vaguely defined, I guess...but, there is a magic to the web platform.)

There's so many areas where we're more productive today because of the network effects of the web as a platform. Also, because the web is universal, I don't have to use Windows, ever. Everything I ever want to do has a Linux version. Anything that falls short of complete platform independence is probably a step backward, IMHO, even if it has other benefits like smaller/faster binary builds.

Also...WebAssembly is coming. We're going to see a fast/efficient web, long before a new platform could possibly be delivered.