Oddbean new post about | logout
 Toxic model. It should be about differences in value of content, moreso than differences in value of people  
 You assign trust/influence/value-to-you to accounts along contexts - not just "about the person" (which I agree is a problem in the bad version of WoT). 

So you are ultimately scoring their content. If I assign low trust to Laeserin on the "economics" context but high on "religion", then when she writes content on those two topics, it will be filtered (for ME, not anyone else) accordingly.
Now, if you attest about me on these two topics, my view of Laeserin will slightly influence your view of Laeserin (only on these topics), along with your own view of her and others you've attested about. 

The process is entirely voluntary and reputation flows from you. Nobody else can influence your graph if you don't trust them to. There is no global state of influence.  
 hello from the nostr army knife 
 What is this  
 We've discussed this before, what you're describing is useless because it's relying on blanket judgements of entire people which most human brains aren't good at during a population crash. You need to rely mainly on judgments of individual pieces of content. Likes and reports exist. Don't know why this is so hard for you people to wrap your heads around 
 Friend... likes and reports on a given piece of content could be exactly the things that feed your web of trust about the author of the content on the contexts they wrote about!

In fact, an early prototype here does just that. Upvote/downvote on wikifreedia articles are used (crudely for now) to build your personal attestations: https://brainstorm.ninja/#/dashboard 
 You're not making sense. Upvotes and downvotes on wikifreedia are attestations, if you're using them to "build attestations" it sounds like you're doing something wrong. Like using houses to build a house. 
 Using them as data points in a personal algorithm to build up a subjective worldview about many accounts and many contexts, taking in other data points as well. 

Not "using a house to build a house", using numbers to build other numbers. AKA "math". 
(Or if you prefer your strange analogy context "using rooms to build a house") 
 Web of trust wouldn't rely on an algorithm to develop a worldview. That sounds more like "web of algorithmic worldview." A web of trust would rely on an algorithm to sort content. The words you're using aren't making sense because you're trying to mask bullshit. 
 Listen man, I don't know what's up with you today but you're being extremely hostile for no reason. I'm not masking any bullshit and the code I'm referring to that I linked to is all open source. You can see for yourself that there's no big bad secret. I didn't even write the thing, I'm just telling you about it. 

"Algorithm" here just means "processes", not "evil social media voodoo". I'm just saying that a bunch of data is combined to generate content rankings from your point of view. That's "an algorithm"

Please relax  
 A subjective web of trust where attestations flow from you outward and you only get information from those you trust cannot be gamed. 

nostr:nevent1qqs29rzrzghl4veh43mvzttx4wlqgwwrrggs7c8gctvdngx7526jm3spr3mhxue69uhkxunewp6x7am0d3nzumn0wd68yvfwvdhk6tczyqh04fc4hw6xm4d7dd7634msqfndz9n5hyfms9u2mk6u9e3anpenzqcyqqqqqqg98hk6n