That was a good skim. Already seeing wafer thin, and excellent, consumer apps built on stable diffusion.

I’ve been increasingly concerned that given now the wide spread skills to build consumer apps for example, the value will be eaten up too quickly.

Thai article makes the point that in fact this isn’t where (economic) value really sits. Building a moat requires more than just the ability to build.

Thus, it feels as though building simple consumer interfaces to the latest AI models is a short-term and largely thankless play.