What about a browser extension doing exactly that: a distributed user populated database, karma based abuse prevention (because it will be abused) and a big indicator that turns on from yellow to red according to the level of garbage the site is throwing against its users.

What would stop them from abusing the karma itself and/or buying “influence” points from popular users?

Hopefully, the fact that the system would create its own loop, and would rely on the accuracy and usefulness of that loop. If it served inaccurate info to users, they might be tricked into opening something they wouldn't have, at which point they would lose trust in the tool upon seeing the thing it's meant to protect them from.

Not sure if we’re talking about the same. It’s clear that one could block reviews from a bad reviewer and e.g. subscribe to another one. But those who game the system will promote their own reviewer accounts and drown real useful ones in low karma via botnets, making them non-discoverable for new users. It means money to them, and it doesn’t cost much, so it’s basically free money in doing that for sites which would want to pay for a good position in such system. It’s inevitable and there are tens to hundreds of examples of third-party review sites out there. The loop is fragile by design.

Any reputation system is prone to manipulation, with its uselessness correlated to popularity.

I am optimistic that a social graph can help with this, using transitive trust and making it easy to transitively "cut off" bad parts of the graph with a blocklist. The problem is that commercial graphs do become useless, as you say.

But with the Fediverse we now have a graph that is resistant to this. I'm not aware of anyone building a reputation system on top of it, but this is something I expect will develop in the future. The system's opinion about a particular data point would then be personalized based on the people you say you trust to indicate that information, combined with the people they say they trust, and so forth.

I have an extension that lets me block websites from Google results.

I agree, we need a replacement. At the protocol level I would like life insurance policy returns and banking service as a place to store my money that allows me to pay any portion of currency I want at any time. So Facebook can be paid for the time I'm on the site, for instance. The returns from the insurance contract might even make it all free to me (it would, easily). Everyone is happy except those who rely on reaching you via forced advertisements. Oh, and the people who rely on enslaving others.

Any such protocol would be labeled as evil, filthy socialism.