Started with kolektiva.social/@dan_ballard/111111540352943785 a few weeks ago:

So Google and the like search results seem to have been degrading for a while, and now with a deluge of AI generated sludge hitting them they are really sinking below the surface of usefulness.

Seems like it’s time for Yahoo’s curated directories to return to victory eh?

More seriously the direction my mind always goes is something P2P, everyone collecting links, tagging them, sharing it. users subscribing to friends and trusted curators and using that new overlay to navigate and explore the web

Does authentication on the web work? It relies on central chains of authority which suck, and in the end, no one checks TLS certs for details, and “bad guys” can slap a cert on their fake site about as easily, so from a user perspective how does it help with anti spoofing?

The topic of trust, identity, verification, and authenticity also came up in the last few weeks discussing anti spoofing of onion sites and what that could even look like. It’s another area where folks extremely don’t want to rely on any central authority, but simply allowing self identification adds little to noting. Especially since v3 onions are already self authenticating. It’s why you don’t need to use TLS for hidden services even tho folks seem to feel the itch to try. And it’s why Tor Browser retooled its icons a bit in the URL bar to convey that a .onion hidden service is as secure and authoritative for its name as any site over SSL. Yet name camping and spoofing of similar names, worse with onion names, persists as a problem.

What’s been skulking around in the back of my head for a while is the idea that this is all the wrong approach. Central authorities cannot scale, and anywhere you can apply AI, AI can be weaponized as well to game it, as we’re seeing with Google results and answer degrading more rapidly for instance.

All my brain wants to do is try p2p distributed bookmarks and related webs of trust. Take us a back to a time when folks were book marking more, starting with directories like Yahoo, and up to del.icio.us. Have folks keep their own bookmarks, and optionally ratings, and let you publish if you want, and let folks subscribe to whomever they trust: friends or trusted community pillars for example, and let that info be integrated to website ratings.

If I got to my bank.com weekly and it’s shows that many of my friends, and thousands of strangers all have it marked, that seems good. If I get phishing email sending me to bonk.com even if it’s got an SSL cert for bonk.com, if no one I know has listed it, my tools could be like “hey, this seems like a new and possibly sketchy site, you sure you want to log in?”.

Sadly, hard to monetize p2p tools so there aren’t a bunch of VC funded startups making these tools, same old story.

In the end, if I had time for even more side projects, I’d love to dive into p2p link shares, decentralized and localized trust systems and book marking in general. I’ve long been interested in book marking but always itchy and creeped out by central systems wanting to host that info, apply AI to it, and sell it to advertisers and aggregators so never engaged much. I’d even like it just so I had a way to sync bookmarks between my browsers without relying on a centralized service!