Also: there are use cases that go beyond filtering and sorting in the background.

Suppose someone you don’t follow DMs you. How do you know whether this DM is worth your time and attention? One method is just to hide the DM if the WoT score is below some cutoff. But what should be the cutoff? There will never be one right answer. So a better method is to show the WoT score next to the author, obtained from Trusted Assertions as the 0-100 rank metric, like Amethyst is doing. The end user can decide whether a score of 10 or 50 or 95 is worth looking at. Maybe I have a lot of free time today, so I’ll read the message if rank is above 10. But tomorrow I’m busier so it’ll have to be 90 to catch my eye. And in the case of personalized GrapeRank, a 50 for me may not be a 50 for you, because you and I may have selected different GrapeRank parameters. My “rigor” parameter may be set to a low value (by my choice), so when I see a score of 50 on amethyst, it means “this user is followed/trusted by 5 moderately trusted users”, whereas your “rigor” may be set to a high value, so when you see the same score of 50 on amethyst it carries more weight, bc it means “this user is followed/trusted by 100 highly trusted users”.

We can’t give users control over their user experience if our philosophy is to hide the number crunching and make all the decisions for them.

At the same time, we ALSO need to make it super easy for the end user. If they don’t want to look under the hood, they don’t have to. Which is why GrapeRank comes with default preset parameters. But showing turn numbers, like I described here and like amethyst is doing, will oftentimes be better than simply filtering and sorting under the hood.

It’s easy for us to underestimate how much work it takes for devs to integrate these scores into their clients. Sure, a dev could roll out a new feature using the scores to sort content. But step 1 is simply to import the scores, make sure they’re being imported correctly, and educate the end user that this new tool is available, here is what it is, here is how it works.

The whole point of sovereignty is to put you in charge of your user experience. Which means putting you in charge of the algos that sort your content. Which means users need to be aware what algos are available and how they work. If we just sort content and say “trust me bro, we’re sorting it in the best way possible, please don’t examine the numbers or ask for details about how it works or look under the hood” — that’s the opposite of what we want to achieve.

Sign up at straycat.brainstorm.social and I’ll calculate your WoT so you can see scores on Amethyst.

@nevent1qqs...

The Brainstorm prototype at straycat.brainstorm.social was down but is back up now. If you sign up, I can calculate your Grapevine — your personalized and portable WoT scores — that can be viewed on Amethyst. CAVEAT: it’s super slow, which is one reason why we’re rebuilding it from scratch at @nprofile1q.... On the prototype I trigger the calculations manually, and the server can process a handful of customers in a day. The rebuilt Brainstorm will be better in a lot of ways, including much faster. For now, sign up on the prototype and I’ll DM you when your Grapevine is ready!

You sort and filter based on numbers which you’re using as a proxy for trust. The act of calculating and using personalized PageRank means you are ranking.

They can, but they shouldn’t, as I argue in my article. Why not? Main reason: portability.

@naddr1qq08...

For starters, the nostr ecosystem needs WoT Service Providers that calculate WoT scores that are *personalized* and *portable*. I’ve written about this in several of my long form posts over the past few months. Developing this ecosystem is the main thrust of the WoT hackathon at @npub1healt..., aka our #wotathon that goes through April. Importantly, these WoT Service Providers must be open source, which means you’ll have the option to do all the calculations yourself, as your own SP. Just like you can run your own BTC node and nostr relay if you so desire.

https://nosfabrica.com/wotathon

TrustRank evolved from PageRank. It was inevitable that it would become evil because the entire endeavor was a centralized, global scoring system from the very beginning. The solution is not to get rid of algorithms or scores, but to personalize them.

Do you consider mute lists to be “censorship”? I don’t, and I’m guessing you don’t either. In which case: why do you consider the WoT scores on Amethyst to fit into the censorship category?

I’d argue it’s censorship if it’s centralized, based on global trust scores. And yes, Bluesky is centralized. But if the system is personalized, like a mute list that you manage yourself or the metrics on Amethyst that are personalized to the end user, then no, it’s not censorship.

I should probably point out that the scores on amethyst are calculated using open source software that you can run yourself. You can delegate the calculations to someone else if you wish, but it’s not necessary.

You want to use these scores in the background but pretend like you’re not? So you can tell people you’re not objectifying them even when that’s exactly what you’re doing?

Google’s PageRank was magic. I remember when it first came out. Keyword search went from useless to amazing overnight.

But Google always calculated *global* PageRank. Meaning the PR scores were as seen by Google. The web: as viewed, as filtered, as scored, as judged —- by Google. The Freedom Tech way is for the algos and scores and point of view to be personalized to the end user. Be your own Google, so to speak.

The Primus relay (primus.nostr1.com) uses GrapeRank to power its WoT (aka WOA, Web of Access).

🍇 🧠 ⚡️

@npub10npj3...