The problem is many founders build things they /think/ people would want. Like "nice sounding problems" that actually no one cares about.
The great thing about building something for yourself is at least you know it's a real problem. It may potentially be too small of a market, but at least it's a real problem.
Also, it's much easier to build something that you deeply understand. Sure, you want a lot of users feedback and talk with them all the time, but deeply understanding the problem really help you focus on what's important rather than building features that users ask but that aren't that important.
The great thing about building something for yourself is at least you know it's a real problem. It may potentially be too small of a market, but at least it's a real problem.
Also, it's much easier to build something that you deeply understand. Sure, you want a lot of users feedback and talk with them all the time, but deeply understanding the problem really help you focus on what's important rather than building features that users ask but that aren't that important.