Interesting article, thanks for sharing. Reputation-based approaches have had a hard time so far in web3. CoLinks (https://colinks.coordinape.com/home) is an interesting recent attempt, a variation of the earlier FriendTech version, but is still figuring out what should be part of reputation and what should not. Govrn focused on itemizing contributions and building up reputation from there, but recording units of work is harder than we thought. Other attempts like praise have been around but haven't caught on for various reasons. I wonder if the idea of a "general" reputation is less productive than the idea of "specific" reputation: a specific "reputation for something". given reputation is inherently backward-looking (what did you do and not do), bounding the scope of what is considered may be the way to go.
I agree with @rolf.eth - narrowing the reputation scope might be helpful here. For example, Gitcoin Passport does a decent baseline reputation check. It isn’t designed to examine if you’re a good developer, a popular crypto influencer, or a scammer. It just uses data from Coinbase, Ethereum, Twitter, and so on to see whether you’re a bot. And it works pretty well. Also, re: article, I think the ability to “choose your own feed” is going to appear pretty soon. The main blockage point was the business model - when social apps make money on ads; they want you to use their engagement-farming feeds. But if you pay for social apps in a B2C SaaS manner, like on Farcaster or Kiwi, then the incentives are more aligned. You just want to make the user happy, and if it takes using some other feed, so be it. I think I even saw some Farcaster tool that lets you see the feed through some other user’s eyes. I really like the observation that “a feed in a social app is not only about content.” Since we can already do things in the feed (mint, buy, swap), I believe social apps might become more like super apps. That would unlock new monetization options for them (taking a % of each action) that might be more user-aligned. BTW, when I looked at visualizations of social graphs, it reminded me of how PayPal dealt with scammers over 20 years ago. They also visualized the graphs between the users and looked for malicious patterns.