Lots of truth to this. Just start building. Build for yourself and real people, not imagined profiles of people. The more you build, the more you explore, the more you'll find low-hanging useful ideas. As you encounter them it’ll surprise you that the space is mostly empty.
Patio11 tweeted about using ChatGPT to immitate social signaling by rewriting content. I thought that was fun so I started doing it. That grew into Spencer Westford: https://vc.blankenship.io
That grew into a product: https://persona.ink
That keeps snowballing. Ever since building for myself my notebook of ideas is growing faster than I can keep up.
I've been doing this for years and all it has done is drill me further down into super-niche rabbit holes that the majority of people don't care about.
I mention this because I see this kind of advice all the time on here, but it so often comes bundled with the unwritten assumption that the kinds of things the reader will build won't stray too far from the mainstream.
After more than a decade of "building", I can state with confidence that profitable ideas don't spring up from countless hours invested in building e.g. a better comb filter for NTSC.
The reality of the situation is that if your interests are things like "doing cool stuff with AI" or "SaaSify yet another thing" or "shoehorn Merkle trees into yet another square hole", then this advice will likely work for you.
If you're the kind of dude (not me) to write ld-decode¹, then it will likely not.
I mean, shit, Donald Graft² has been "building" far longer than I have and yet none of his "building" has led to a business AFAICT.