That only works if you validate users, otherwise the trust database becomes 99% the result of the actions of SEO firms.

Doesn't cloudfare do this? (extremely extensively imo - just about every single click is a captcha of some type sometimes) As much as HN despises blockchains because of crypto, (no, that's not the only use of that tech) it does seem like verification of websites could be a decentralized consensus - along with DNS as well -and user verification could be a portion of the consensus of website quality. This could even help with making self hosted search engines more accessable

Hmm, or you outsource the trust problem to your users. Let them select which domains to trust or who to trust to weight trustworthiness of domains for you.

Imagine that user A can create a list of domains with trustworthiness score. User B can then use that list by going to user-a-awesome-curation.koogle.com?q=nice+shoes

It might create filter bubbles but it would be transparent filter bubbles. You could even wikify/open source the curation.

A lot of this could be solved if we could signal intent or context before searching. But that would require that you know how to use the tool which is a gargantuan user barrier from Google's point of view. Meanwhile, we have to hack about trying to signal context.

pretty much what we're doing at https://breezethat.com, a topic search engine

- we currently curate domains / pages internally based on trust - we'll be adding an open low-code way for others to vote / moderate