Perhaps because, amongst the other solid reasons already posted:

Demand is the real bottleneck: New tools expand who can ship, but they don’t expand how many problems are worth solving or audiences worth serving. Adoption tends to concentrate among "lead users" and not "everyone".

App store markets are power-law distributed (no citations sorry, it's just my belief): A tiny slice of publishers captures the most downloads/revenue. That’s discoverability, not "lack of builders".

Attention and distribution are winner-take-most: Even if creation is cheap, attention is scarce.

The hidden (recurring) cost as other commenters point out is maintenance: Tooling slashes first release cost but not lifecycle cost.

Problem-finding outweighs problem-solving: If the value of your app depends on users or data network effects, you still face the "cold start problem".

"Ease" can change the meaning of the signal: If anyone can press a button and "ship an app" the act of shipping stops signaling much. Paradoxically, effort can increase personal valuation (the IKEA effect), and a lower cost to the creator as seen from the outside kills the (Zahavi) signal.

And finally, maybe people just don't actually want to use and/or make apps that much? That's not to say that good apps aren't valuable, but the ubiquity of various platforms app stores implies that there's some huge demand, but if most app usage is concentrated amongst a small number of genuine day-to-day problem solving tools that already have heavy hitters that have been around for years, an influx of new things perhaps isn't that interesting.