I disagree with that. VisiCalc, MacPaint, and many modern “significant” apps started out as 1 person projects. I agree collaboration and communication are vital skills but you don’t need to make this grandiose statement to defend that.
I have done some very “significant” stuff, that was a big deal, when I released it, but would be considered “quaint,” nowadays.
That said, my significant work created a baseline that has been extended by a team, so it’s still “significant.”
The app I just released ain’t Facebook, but it’s pretty non-trivial; consisting of three backend servers (that I wrote), and two native frontends (that I also wrote).
Takes a while, though, and there’s a lot of moving parts. I’ve gotten used to working at this scale, but I’ve also been writing shipping software, for over 35 years.
Also consider games, back then one or two guys could do a game, now all but the outliers are done by large teams.
I would also add that with LLM assistance, one person being able to create a "significant" app is actually more likely today than five years ago.
If anything the opposite is true. Gamedev tooling is now so good and high level one or two people can make sophisticated games, in terms of audio, gameplay and visuals that an entire team could not make in the 80s or 90s.
Just one example, Signalis (https://rose-engine.org/signalis/). One developer and one artist if I'm not mistaken, stunning game that in say 2000 would have been a release from a major publisher.