To confidently offer what Primal offers every client would have to do what Primal is doing now, that is to say crawl and index everything. The same way Bing has to duplicate Google's (very expensive) work of crawling and indexing. There's just no getting around that at present, sans goodness-of-my-heart solutions. (Cooperative frameworks like this Espresso are well-envisioned but in early days, and address more the technical than the incentive-structure side of things.) This crawling and indexing also by nature applies to spam that has gotten past a given relay's own filter.

This isn't so much about performing a computational event each time a user makes a query, as the bulk of crawling and indexing is done in anticipation of a query; minutes before, days before, years before. Rather this is about maintaining a foundation upon which algorithms of a certain type can be run. So we're talking about a related but somewhat different set of tasks here.

Basically, in so far as this set goes, a 'Primal-like' user experience cannot scale on Nostr without either duplication of effort (every client sends out its own Googlebot or Bingbot) or a bandaid solution whereby the clients that are doing the costly work of crawling and indexing (the Primal or Primals) provide some kind of goodness-of-my-heart access to the fruits of their labour, and the clients that are making use of this access cross their fingers and hope things all carry on.