It's hard to compete with Google. Back when Google started, you could fit the world wide web on a single computer. Today you can only really get started the same way Google did if you focus on an unfancy subset of the web that hasn't been react'd and wasm'd into exabytes of opacity. https://wiby.org/ does it I think by refusing to index any HTML page that has <script> tags. Websites like https://millionshort.com/ provide pretty nice results sometimes, but I don't think they actually index the web on their own; they probably use the Bing or Google API on the backend. With million short if I want bike reviews for example, I get the authentic blog posts from 2010 with amazon affiliate links to bike listings that are no longer being sold lol.
The problem is I don't think the Wiby model (rejecting JavaScript) or the MillionShort model (rejecting popularity) have done a good job capturing what we loved most about the old web and systematizing its curation. There's definitely an opportunity for someone to come along and create a focused niche alternative that's better.
One way I suppose it could be done, is if you convinced a bunch of trustworthy high status trendsetter type people to subscribe to a paid service with a browser toolbar that lets them click a button for each website they visit to say "I like this" or "I dislike this" and then use that information to train a neural network that divides seo spam from content. Mix that with classic page rank and you might have something good. I'm not sure if it'd ever appeal to a more general public audience though.
Anyone who does it is also going to want to make a deal with the archive.org to somehow get a snapshot of the old web, to recreate those original experiments. Or possibly even resurrect an old build of it. Plus Gutenberg. There's enough content from 2005 web alone and all the books published before to last anyone several lifetimes. That's actually one of Google's blind spots. They're so good at up-to-the-minute indexing of current events that sometimes if you just want to get the text to something like Seneca it's like pulling teeth.
The problem is I don't think the Wiby model (rejecting JavaScript) or the MillionShort model (rejecting popularity) have done a good job capturing what we loved most about the old web and systematizing its curation. There's definitely an opportunity for someone to come along and create a focused niche alternative that's better.
One way I suppose it could be done, is if you convinced a bunch of trustworthy high status trendsetter type people to subscribe to a paid service with a browser toolbar that lets them click a button for each website they visit to say "I like this" or "I dislike this" and then use that information to train a neural network that divides seo spam from content. Mix that with classic page rank and you might have something good. I'm not sure if it'd ever appeal to a more general public audience though.
Anyone who does it is also going to want to make a deal with the archive.org to somehow get a snapshot of the old web, to recreate those original experiments. Or possibly even resurrect an old build of it. Plus Gutenberg. There's enough content from 2005 web alone and all the books published before to last anyone several lifetimes. That's actually one of Google's blind spots. They're so good at up-to-the-minute indexing of current events that sometimes if you just want to get the text to something like Seneca it's like pulling teeth.