> There is no friction. It's trivial to write a program that can scrape a Gemini network.

It's not trivial at all. First, you have to want to do it, then you have to commit time and effort to doing it, then you have to maintain the solution you deploy specifically for Gemini in parallel to your web scraping architecture.

> If there is no one pulling data from Gemini servers yet, is not because it's difficult do it, but merely because it's still too small to be relevant.

Exactly. But it if was using web tech, all of the existing web scrapers could just be pointed at it with minimal effort. So using a separate, custom tech stack is what keeps the threshold of effort in front the threshold of desire.

And using a separate tech stack also creates intentional friction in terms of new user adoption, keeping it slow and maintaining the protocol's niche status. So this also helps keep that threshold of desire distant.