Not to be an LLM shill, but perhaps the reason we are not seeing a massive increase in apps being released and new domain names is due to economic factors stunting the potential AI. It's just a confounding factor I can think of.

If the author wants to steelman their argument, they could look into the total number of developers before AI and after AI. They could compare working hours (not officially recorded working hours, but actual "I am working on making code" hours).

There are other economic factors also. Large corporations will control the majority of developer working hours. And large corporations won't be making shovelware. Meanwhile people who are not so beholden to factors of economics aren't necessarily going to build public shovelware. They might just be building scripts for personal use.