straycat
david

https://straycat.brainstorm.social

neurologist and freedom tech maxi
🍇 Grapevine, 🧠⚡️Brainstorm

Public Key

npub1u5njm6g5h5cpw4wy8xugu62e5s7f6fnysv0sj0z3a8rengt2zqhsxrldq3

NIP-05 Address

Profile Code

nprofile1qqsw2feday2t6vqh2hzrnwywd9v6g0yayejgx8cf83g7n3ue594pqtcpzpmhxue69uhkummnw3ezuamfdejsz9rhwden5te0wfjkccte9ejxzmt4wvhxjmcpr3mhxue69uhkummnw3ezucnfw33k76twv4ezuum0vd5kzmqprdmhxue69uhhyetvv9uju6rpwdjkuurxv4nxvu3wvdhk6qg4waehxw309ajkgetw9ehx7um5wghxcctwvsq3vamnwvaz7tmjv4kxz7fwwpexjmtpdshxuet5q7j6kp

Publishing to

nostr.wine

relay.damus.io

nostr.bitcoiner.social

relay.hasenpfeffr.com

eden.nostr.land

relay.primal.net

Author Public Key

npub1u5njm6g5h5cpw4wy8xugu62e5s7f6fnysv0sj0z3a8rengt2zqhsxrldq3

Show more details

Last Notes

2026-02-15 04:19:35 UTC

- reply

The bots are going to discover they need their own WoT.

2026-02-09 21:46:20 UTC

- reply

Reasoning from first principles is always a good thing. But we have to recognize our cognitive limitations as individuals. I’d love to review every line of open source code that I’ll ever use but that’s never going to happen. The vast vast majority of what I know about the world is stuff that I’ve learned from other people. Topics like bitcoin I reason from first principles to the best of my ability, but I’ll never stop relying on trusted people to guide my thinking.
I don’t have a strong opinion on core vs knots. And LukeJr can be kinda crazy. But he’ll always have my respect based on a few of the positions he’s taken in the past. Among them: the March 2013 unintentional chain split. Read the bitcoin-dev channel, he knew pretty much instantly what happened and what needed to be done, faster than most (including Gavin).

2026-02-09 19:45:59 UTC

My clawdbot just named itself Arc. ⚡️

2026-02-08 01:40:09 UTC

- reply

Would it be feasible for a single instance to support multiple sources? So one tech-savvy dev could calculate WoT scores and publish TAs for family, friends, etc?

2026-02-06 23:14:44 UTC

- reply

Ok - I just created a new user called straycat’s shadow, followed myself (straycat), logged in at relatr.xyz/ta. I see Gigi listed at the top as the “source” which I guess means perspective, meaning that if Gigi follows Alice, then Alice’s score gets a big boost bc she’s only one hop away from
Gigi. But I want to make myself (the new account I just created) the source. Can I do that? Or do I need to run my own relatr server?

2026-02-06 21:07:13 UTC

- reply

I probably don’t understand relatr or CVM well enough so maybe I’ll start by asking some ELI5 questions. For starters: suppose I’m an average nostr user, not a dev, and I want relatr to calculate and publish TAs for me. My goal is to see the 0-100 trust scores (rank) show up when I log into Amethyst. How do I do that?

2026-02-06 19:08:37 UTC

- reply

Relatr has started publishing Trusted Assertions using NIP-85. I’m not sure how deep into it they’ve gotten or what pain points they’ve encountered ( @npub1gzu…a5ds ?). So far there are a handful of clients (Amethyst, NymRank.dev, NostrHub.io, Nostria, Sandbox, some others) with varying degrees of NIP-85 integration. So we’re still in the early stages of testing it all out.
Here’s the litmus test in my mind as of today (subject to change as I have discussions like this): suppose law enforcement goes to zapstore or some other client and demands that you integrate with Brainstorm and support GrapeRank but blacklist PageRank by Vertex. Or vice versa. Oh and you can’t tell your users that we ever spoke to you. I’d like you to tell them that it’s impossible to comply with their request because that’s just not how it works — the choice of algo and WoT service provider is completely out of your hands. Does ContextVM work that way? Or can it? I genuinely don’t know the answer to that.

2026-02-06 18:27:17 UTC

- reply

I’m not saying that isn’t your goal. I am saying, however, that when the client selects one or more WoT service providers for its users, that introduces a fatal flaw that we need to fix before it actually kills us, and NIP-85 fixes it.
When I first learned about bitcoin, I thought there’s no way we can fit every transaction for billions of people onto a single ledger. Turns out you can. It’s a challenge and we have to figure out how to deal with it, but it’s not a fatal flaw.
If I don’t have room on my personal NIP-85 relay for 8 billion kind 30382 events, then I won’t store that many. Only the people I want to keep track of. I can keep track of reputation in my head but my head is limited by Dunbar’s number, about 150 people. If I’m willing to fill my personal NIP-85 relay with a million Trusted Assertions, maybe 1 GB of data, then I’ve boosted my personal Dunbar’s number from 150 to a million. I’d call that a good start.

2026-02-06 17:22:04 UTC

- reply

Hard for a user to get a feel for what a PageRank score means, especially when there are many different ways to run it (different parameters, etc). If I as a user could be guaranteed to see the same PR score on zapstore as I see on amethyst and damus and primal and everywhere else, maybe over time I could get a feel for what score is a good score. And that can happen if we give users full control over how and where their scores are calculated. Direct client-to-vendor integration is contrary to that goal. But NIP-85 is designed with that goal in mind: giving users, not clients, control over which WoT service provider to use.

2026-02-06 17:14:22 UTC

- reply

I’ll let you and others know when the next iteration is ready to try out. It’s already better than my prototype but needs polishing.
GrapeRank has a very clear roadmap for contextual scores. At least two major categories of strategies. It’s only a matter of time.
Does PageRank have a path forward for contextual scores? If you’re pursuing one, I’d like to hear it. I’ll actually suggest a method: one of the categories of strategies for GR contextual scores is decentralized lists, which can be curated by any list of pubkeys:
- everyone (including bad actors, not ideal)
- your follows list (easy to code, not a bad choice)
- your follows plus their follows (wider reach but more chance for bad actors to sneak in)
- anyone with personalized GR score above a cutoff
- anyone with personalized PageRank score above a cutoff
@npub19ma…axkl

2026-02-06 17:03:20 UTC

- reply

If I search for John on NymRank I see 10 results.
Do you mean the 2nd result on npub.world for Odell? NymRank serves a different purpose from your typical keyword search. In addition to GrapeRank, it calculates an affinity score that dings the second result on your page, mostly bc the name isn’t an exact match.
I know this isn’t an exact side by side comparison. That will be coming soon.

2026-02-06 16:14:32 UTC

- reply

GrapeRank may be newer and more experimental than PageRank but in our hands, GR empirically does a better job than PR at its core function.
Search for Odell at npub.world (which uses PageRank, as of course you well know) to NymRank.dev, which uses GrapeRank. You’ll see a floodgate of impersonators that fool PageRank but are filtered out by GrapeRank using standard (default) GR parameters.
https://i.nostr.build/uPB15hsBNQBbyEiw.png
https://i.nostr.build/VTjVWLpD4Pg70DOm.png

2026-02-06 15:37:18 UTC

- reply

Once it’s available and the marketplace is developed, your users will probably already be signed up to Brainstorm or a competitor service that supports NIP-85 so they can use them on clients like Amethyst. Vertex could easily be one of those providers. It’s not like you have to sign up every time you want to use the scores.
Yes, PR is widely deployed, with well known and easily exploitable weaknesses like link farm attacks, the solutions to which are band-aids. This attack, btw, is extraordinarily cheaper and easier to execute in the context of nostr and follows than in the context of internet keyword search and hyperlinks.

2026-02-06 15:07:31 UTC

- reply

I’m not suggesting you send your users to my prototype! The prototype is only that, a prototype, not a product. But the product is coming. (Soon™️!) So if I were you or any dev who hasn’t been following this closely and doesn’t want to experiment with alpha (or pre-alpha) stuff, I’d wait until there’s at least one stable, user-facing WoT service provider who supports NIP-85, and when it’s ready (easy for users to sign up, good UX, all open sourced, etc), I’d be ready to integrate scores using this method.
Yes, I created GrapeRank. I have articles explaining what it is and how it works. My prototype is open source and the rebuild will also be open source, which means you can be your own WoT service provider. And our #wotathon is very likely to produce another team who will have a competing implementation of GrapeRank. The more the better.

2026-02-06 14:56:43 UTC

- reply

Is this a non-issue?
https://i.nostr.build/OF9C45jfOhTUQ8oO.png

2026-02-06 14:55:47 UTC

- reply

So the real result is the first result, and that’s the solution.
That solution works for Odell because he has the second (or maybe third) highest PR score in nostr.
What if you’re not an influencer? What if your job is to screen code for malware, I want to trust you and not an impersonator to do it, but you haven’t shotposted your way to 1000 followers. Your system as is will not protect them from impersonation.

2026-02-06 14:02:40 UTC

- reply

GrapeRank is the best tool to flag the moustaches and trucker hats!

2026-02-06 13:59:07 UTC

- reply

So step 1 is I gather all my files in a folder, then I point this tool at it? Is there a tool to help me gather them together? (Some are on nostr, some on github, some on substack, some on my local machine, etc)

2026-02-06 13:38:38 UTC

- reply

So how do you explain this many impersonators fooling your system?
https://i.nostr.build/kZMa332xMlgEhJc6.jpg

2026-02-06 13:20:46 UTC

- reply

That’s my prototype which I built on my own and is super slow. As of a few months ago I am with a team and some devs who are much more talented than I am!
Yes, more than one click when you include sign on, plus one to sign the 10040 event which needs to be done or else amethyst or zapstore or damus or wherever won’t know how to access your scores.
For the prototype, I need to trigger score calculation on the back end which I just did for you. In an hour or so your scores will be calculated, stored locally and available at https://straycat.brainstorm.social/profiles.html. (Our rebuilt version does it in ~ 5 minutes.) if you go back and sign the 10040 event, then I’ll be able to publish scores as Trusted Assertions (which for the prototype takes hours, for the rebuild takes another 5 min) and you’ll see them on amethyst. 😃
By relying on PageRank, your zapstore users are going to get fooled by the first bad actor who decides to employ a link farm attack. If that hasn’t already happened, it’s only bc it’s not yet a big enough honeypot.

2026-02-06 03:14:37 UTC

- reply

Mostly text - NIPs, protocols written and rewritten, articles, blog posts. Some implementations in code of varying completeness. A few video talks. A few keynote presentations.

2026-02-06 02:16:07 UTC

- reply

I’m in the process of digging up everything I’ve ever written about Brainstorm (even before I called it Brainstorm) so I can train an LLM on it, then vibe code the Knowledge Graph half of it, the counterpart to Grapevine which is currently already being built into a product. I’ve never done this before though. Is it worth getting a Mac mini? Splurge for a studio?

2026-02-06 01:52:03 UTC

- reply

If anyone ever starts to use zaps as a real and meaningful signal, like to select content for trending feeds, global zaps will be a piece of cake to game. Imagine an army of bots controlled by a bad actor. They zap a few sats back and forth in a continuous closed loop swarm of activity. It will cost basically nothing for them to top the zap leaderboards on both sending and receiving.

2026-02-06 01:46:21 UTC

- reply

That’s kind of how NIP-85 already works. To change providers, I just replace my 10040 event with a new one.

2026-02-06 01:45:18 UTC

- reply

One-click sign up for the service once and forget about it (unless you want to personalize the parameters, which some users really seem to want to do). From then on, your personalized scores are available throughout nostr.

2026-02-06 01:43:20 UTC

- reply

Remains to be seen how often we’ll update. My prototype takes forever. The rebuild is an order of magnitude or two faster. Most likely, we’ll have a free tier that updates automatically but very infrequently plus on-demand for a limited number of times per day or week or something. Then at some point we will have paid tiers to update more frequently, then probably add-on services — for example, support for red flag events like if a highly trusted pubkey suddenly gets multiple trusted NIP-56 reports of key compromise, that pubkey will get updated immediately. (Note that if this were to happen, for example if Dorsey’s key were to get compromised, any PageRank-based system would be toast, bc his pubkey’s PageRank score will for all practical purposes NEVER drop below yours or mine.)
When nostr starts to get millions of users, there are going to be all sorts of strategies for scaling. None of them perfect, but many that are more than good enough. GrapeRank has a natural scaling solution which is that if you know a few trusted pubkeys in some distant community (like China or Japan, for instance) and they run Brainstorm over their communities, then if they allow you to import their Trusted Assertions, you don’t really need to recalculate over the pubkeys in their community.

2026-02-06 01:24:36 UTC

- reply

Are you Nash-the-ai?

2026-02-06 01:21:04 UTC

- reply

It works for us. Here’s how it works:
You go to a Brainstorm instance, such as my prototype at straycat.brainstorm.social. Click the button to sign up and to sign a user preference 10040 event, which basically says “brainstorm is my WoT Service Provider.” We create a new brainstorm nsec just for you, like a franzap alter ego, which we’ll use to publish scores on your behalf. (We may need to publish updated scores when you’re offline, so doesn’t work to publish straight from your pubkey.) We calculate personalized GrapeRank from your perspective, like PageRank but uses more than just follows. This produces a 0-100 Trust Score for every pubkey in nostr. For everyone whose score is above a cutoff like 1 or 2, which usually is about 130k pubkeys, we publish one kind 30382 Trusted Assertion to a dedicated TA relay, which contains multiple scores including GrapeRank, PageRank, distance from you in hops, and counts of verified followers and verified reporters.
It sounds like too many events to be feasible, but it’s not really. It works. The 10040 event tells the client (Amethyst or zapstore) where to find scores from your perspective. All it needs are your alter-ego pubkey and the relay with the TAs. The beauty is that zapstore has no way of knowing whether your TAs come from Brainstorm or Vertex (if I can ever convince pip to use TAs) or relatr (which is also using TAs) or whatever other WoT Service Provider comes along. So if law enforcement tells you that zapstore needs to stop support for Vertex bc they’re not whitelisted, you can say zapstore doesn’t integrate directly with score producers bc that’s not how TAs work.
As a user, you don’t have to know how any of this works. You just click a button and 10 minutes later you see a 0-100 Trust Metric over every profile pic on Amethyst. It’s personalized, like personalized PageRank, but has an actual, workable fix to the link farm attack that is the fatal flaw of PageRank.

2026-02-05 23:55:39 UTC

- reply

Once Brainstorm gets out of alpha and you can create / update your personalized WoT with a button click, it’ll be easy for you to play with this.

2026-02-05 23:50:19 UTC

- reply

With NIP-85, zapstore will have no need to send a job request. And zapstore will have no need to integrate one by one with WoT Service Providers as they come out (vertex, Brainstorm, relatr, trustr, etc). Just pull your user’s kind 10040 note which points to the notes that have the personalized trust scores and that’s all you gotta do.

2026-02-05 23:42:13 UTC

- reply

We are publishing WoT scores using @npub1gcx…nj5z's NIP-85 Trusted Assertions which was merged pretty recently. You can go to Amethyst and see WoT scores which were calculated either on Brainstorm or on Relatr — user’s choice. NIP-85 is designed so that Amethyst/zapstore/some other client doesn’t need to know where they were calculated. If Vertex were to start publishing scores using NIP-85, then Vertex scores would show up immediately on Amethyst. No need to sell Vitor on Vertex first.
Right now Brainstorm has a prototype that is being rebuilt, currently alpha-ish, but progressing fast.

2026-02-05 23:18:52 UTC

- reply

The goal of the @npub1hea…vp8p #wotathon is to cultivate a marketplace of WoT service providers so that algos and trust scores are personalized and portable. Vertex is one. Brainstorm is one. Also Relatr. I think Trustr will emerge soon.

2026-02-05 23:14:45 UTC

- reply

I agree with you about DVMs. I’m glad I never actually spent the time to implement them.

2026-02-05 23:03:50 UTC

@npub1wf4…dgh9 and @npub176p…vgup — suppose Brainstorm supports PageRank and DVMs. I’m a zapstore user and I want to switch from using Vertex to Brainstorm. How do I do that?

2026-02-05 22:14:57 UTC

- reply

Point 4: by privacy in this context, I mean I don’t want zapstore/damus/amethyst/primal devs to tell me that PageRank is the only algo they support, or even to know which algos or parameters I am using. The only way for that to happen is if the scores are calculated elsewhere. NIP-85 Trusted Assertions was designed with this in mind.

2026-02-05 22:07:47 UTC

- reply

WoT scores calculated by primal are much less portable than WoT scores calculated by brainstorm/vertex/relatr/trustr and consumed by amethyst/primal/damus.

2026-02-05 21:48:06 UTC

- reply

On item 3:
As a user I want to switch back and forth between vertex, brainstorm, relatr, trustr, etc. This should be easy, good UX, no technical expertise.
Suppose I want to go to zapstore and switch from vertex to another SP (which supports DVMs, for the sake of argument). How does this happen?

2026-02-05 21:38:20 UTC

- reply

I should probably clarify: I don’t mean to say that client side calculations should NEVER happen. I mean only that client side should not be the ONLY option.
There are many instances where it would make sense to do a mixture of both at the same time. Example: an e-commerce client with average 5-star ratings. It would make sense to me for the 5-star product ratings to be calculated by the client, but the client uses something like the rank score calculated elsewhere to remove ratings authored by bad actors.
So the e-commerce client calculates 5-star ratings on the fly, which is not hard bc there’s (usually) not a lot of data involved in the calculation, and uses them locally. But the client could have the option to publish those average ratings using a format like trusted lists, so that I can use those scores elsewhere (or maybe use the locally repeatedly without the need to recalculate each time I navigate back to the same page), in which case the e-commerce site is serving as a mini service provider.

2026-02-05 21:24:43 UTC

- reply

Any WoT Service Provider that supports NIP-85!
✅ Brainstorm
✅ Relatr
who else?

2026-02-05 15:24:58 UTC

WoT algos and scores should follow these principles:
1. personalized (not global)
2. portable (calculated by a service provider, not a client)
3. user choice (easy to switch providers)
4. private (platform doesn’t know which SP is providing scores)
Agree or not?
If agree - how to achieve?

2026-02-04 12:31:29 UTC

- reply

Interesting word! 🤔
Decentralized: nobody is (or should be) at the center of THE universe
Self-centralized: I am at the center of MY universe

2026-02-04 01:49:56 UTC

- reply

Can he reliably root out impostors and NIP-56 report them? Could be useful 🤔

2026-02-04 01:45:45 UTC

- reply

Can @npub1qru…wnup follow npubs?

2026-02-03 06:51:12 UTC

- reply

Hey @npub1hgv…9jnv — apologies for the delay! @npub1cgc…2ndy ‘s right, I’m still doing them manually. Just triggers yours and it’ll take a while but you should see them in amethyst by morning!
The rebuilt Brainstorm will be automated and much faster once it’s out 🙂

2026-02-02 01:11:38 UTC

- reply

I took two philosophy of science classes from him as an undergrad. Two of my favs 🤓

2026-01-31 03:49:52 UTC

- reply

Primus.nostr1.com by
@npub10np…tl5h and
@npub1cgc…2ndy

2026-01-29 15:46:54 UTC

15 minutes! 😃
#note12km…s5kc

2026-01-28 21:44:03 UTC

#wotathon community call less than 24 hours away!
🫂
#note12km…s5kc

2026-01-27 19:00:27 UTC

- reply

If you’d like to be an alpha tester for our rebuilt brainstorm, I’ll DM you the url. Once you click go, you should start seeing scores on amethyst within 20 min, maybe less.

2026-01-24 17:44:27 UTC

- reply

WoT has multiple uses. The first is getting rid of the obvious bots and bad actors who may want to spam your app with junk. Centrality algorithms like GrapeRank or PageRank have a much more extensive reach than follows of follows, and are effective and simple enough to implement if and when spam becomes a problem on your app. Ideally you’ll farm out calculation of personalized trust metrics to a WoT Service Provider like Brainstorm, Vertex, or Relatr. @npub142g…xrj0 and other clients have already started doing this using @npub1gcx…nj5z's Trusted Assertions NIP (which @npub180c…h6w6 just merged) to empower the user and avoid SP vendor lock-in. Cultivation of the ecosystem of WoT SPs to make WoT metrics both *personalized* and *portable* is the primary goal of the #wotathon — the @npub1hea…vp8p web of trust hackathon that runs through April. We have weekly community calls on Thursdays; you should join us!
The more challenging and exciting WoT use case will be content discovery. Suppose I want to find songs in some niche musical category by awesome but undiscovered musicians who aren’t on Spotify. Do we want to wait for said musicians to build up 10k followers on nostr? Why wait for that when my trusted, extended community can find exactly the content I want and serve it up to me on a silver platter? This is the type of use case we’re tackling with Brainstorm at @npub1hea…vp8p. This particular use case is being explored in depth by @npub1hqa…t56s. Keep an eye on what he’s up to; it will be relevant to your app. 2026 will be an exciting year for WoT in nostr!
cc @npub1t8p…gkvf
https://nosfabrica.com/wotathon

2026-01-24 16:53:59 UTC

- reply

Congrats 😃🥂

2026-01-23 18:30:24 UTC

I can’t wait!! 🫂
#note1n04…n9g2

2026-01-22 21:09:17 UTC

👀
#note1msa…aj2q

2026-01-22 16:01:53 UTC

We’re live!
https://zoom.us/j/91801428092?pwd=k5Gl2pYKX1X7GXPpiEdsEDaGIYsdvK.1
also on zap.stream/nosfabrica
#note1m85…uv88

2026-01-22 15:16:02 UTC

Community call in 45 minutes!
#note1m85…uv88

2026-01-22 06:36:23 UTC

- reply

Haha ok, it’ll cost you 69 sats tho

2026-01-22 06:33:49 UTC

- reply

I see you’re using Trusted Assertions to determine relay operator trust as part of your relay trust score calculations!! 🔥🔥 👍🏻

2026-01-22 06:27:17 UTC

- reply

Good point. One of the biggest challenges in WoT is understanding the problem itself. What exactly are we trying to solve?

2026-01-21 23:41:57 UTC

What are the big picture, WoT-related topics, questions, issues that are on people’s minds? Off the top of my head, here are a few:
1. Where should the trust signal come from? Follows, engagements, attestations, something else?
2. What are the different categories of trust? (And who decides what those categories are?)
3. Is it possible — or even desirable, under any circumstances — for standards, protocols, ontologies, etc to be created and actively managed in the absence of a designated maintainer, committee, etc? If so, how?
4. What are the different ways we should communicate scores from WoT Service Providers to nostr clients?
5. What WoT tools and features are needed before we can attract normies to nostr? WoT scores that are personalized, portable, contextual … what else?
6. How do we make WoT scores contextual?
7. How do we convey confidence in any given WoT score?
8. How can we monetize WoT?
9. How do we empower users to “choose your own algorithm”?
And here are two we discussed on our weekly #wotathon call last week:
1. Is WoT leading us towards a Black Mirror dystopia?
2. Is it OK to attach a number to trust? Why or why not?
What other questions can you think of? We want to hear from you!
cc: @npub1hea…vp8p

2026-01-21 16:27:57 UTC

One day until our next #WoTathon community call!
#note1m85…uv88

2026-01-20 05:04:40 UTC

Fiat: algos are hidden 🙈👎🏻
Freedom tech: algos are out in plain sight 😄👏🏻
How can we choose our algos if we have no idea how they work?

2026-01-20 03:01:54 UTC

- reply

I just started recalculating your Grapevine and it's now in the process of publishing your 30382 events (slowly). Let me know if you don't see them showing up.
Have you been able to see trust scores on Amethyst?

2026-01-15 15:05:08 UTC

#wotathon Community Call is less than an hour away!!
#note1jrr…ezt3

2026-01-14 17:48:24 UTC

Don’t forget — next #wotathon community call is less than 24 hours away!
#note1jrr…ezt3

2026-01-14 03:24:51 UTC

- reply

Just realized I can add people to the group

2026-01-14 03:12:53 UTC

- reply

@npub1t8p…gkvf — help @npub1yn8…6pzj get set up on our WoTathon signal chat!

2026-01-14 02:51:48 UTC

- reply

Glad to see you active in the space! And glad you’re using NostrHub 🔥

2026-01-13 17:13:14 UTC

This was a fun and interesting discussion!
💜🍇🧠⚡️
#note1y4z…lxm8

2026-01-13 00:26:03 UTC

- reply

🎯

2026-01-12 23:57:41 UTC

- reply

Upon rereading your previous message, it seems to me you’re saying the default should ALWAYS be: don’t display them.
Unless you have an exception in mind? If so, what is it?
I’m trying to pin you down bc you advocate for a very generic, very widely-encompassing rule and when I ask WHY this is your stance, you just walk through one specific use case where you don’t show trust scores. And then you extrapolate from the specific case to the general rule (never show scores by default) without justification. I suspect your true motivation is that users will tell you “you can’t put a number to trust.” Which is not true — we can and we do. Still, it’s an understandable (albeit vaguely defined) sentiment, one that MANY users will have, and we should respect our users. So your solution is to tell them “I agree, we can’t put a number to trust, we would never do that! Look, no numbers!!” And yet, putting numbers to trust is *exactly* what we’re doing. It is dishonest to pretend we’re not — hence disrespectful at a very deep level to our users — and many users will sense the dishonesty.

2026-01-12 21:56:32 UTC

- reply

I’m not saying the default should be to display ALL the scores ALL the time. I’m saying sometimes we do, sometimes we don’t. No sense in showering the user with a deluge of confusing numbers for no purpose.
But you’re saying we should never ever display raw trust scores. We should ALWAYS by default hide them under the hood. There are ZERO use cases for showing a 0-100 Trust Score next to a user profile.
Am I misunderstanding your position? Is there at least one use case you can think of where you think we SHOULD show a raw Trust Score by default?

2026-01-12 15:47:08 UTC

- reply

That’s my point.
Me: there are times when you don’t want to display the scores and times when you do want to display the scores.
You: I can think of a use case where you don’t need to display the scores. Therefore nobody should display the scores, for any use case, ever.

2026-01-12 00:35:43 UTC

- reply

I feel like our discussions too often fall into the following pattern:
Me: we’re building a new thing called an automobile. It will get you from city A to city B. It’s awesome.
You: The bicycles I sell are great for getting around town. No one needs automobiles.
Me: ok. Let’s have both. Bikes for short distances, automobiles for long distances. It’s not either-or, it’s both-and.
You: For the specific use case that I’m focused on, getting around town, bikes are awesome. Therefore automobiles have zero use cases. No one will need them, ever, for anything.
🤷🏻‍♂️

2026-01-11 20:05:11 UTC

- reply

cc @npub1t8p…gkvf

2026-01-11 19:50:08 UTC

- reply

Dawn would you ever wanna hop on a #wotathon Thursday community call and take 5 minutes to tell all the devs what you want, what you’re getting, and what you’re not getting out of WoT?

2026-01-11 15:42:06 UTC

- reply

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.

2026-01-11 15:08:18 UTC

- reply

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.

2026-01-11 14:47:49 UTC

- reply

2026-01-11 14:41:43 UTC

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 @nprofile…xza3. 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!

2026-01-10 00:08:12 UTC

- reply

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.

2026-01-09 23:05:28 UTC

- reply

They can, but they shouldn’t, as I argue in my article. Why not? Main reason: portability.
#naddr1qq…ey5u

2026-01-09 22:38:33 UTC

- reply

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 @npub1hea…vp8p, 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

2026-01-09 22:29:44 UTC

- reply

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.

2026-01-09 22:13:47 UTC

- reply

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.

2026-01-09 22:05:13 UTC

- reply

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?

2026-01-09 22:00:08 UTC

- reply

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.

2026-01-09 21:53:11 UTC

- reply

The Primus relay (primus.nostr1.com) uses GrapeRank to power its WoT (aka WOA, Web of Access).
🍇 🧠 ⚡️
@npub10np…tl5h

2026-01-09 21:38:41 UTC

- reply

Or checked out any of the cool projects people are working on? 😃
https://www.kanbanstr.com/#/board/be7bf5de068c1d842ed34a7c270507ec940f5ea51671cfd062a95e9d09420d0a/802f4072-8c60-4eaa-bbb8-5f086d40e585

2026-01-09 21:36:47 UTC

- reply

Have you joined any of our weekly #wotathon community calls?

2026-01-09 16:59:39 UTC

- reply

Some people think bitcoin is evil because “money is the root of all evil.” But to them, money means fiat. They don’t realize that the evils of fiat should not be assumed to apply to decentralized money.
Same thing with reputation. Centralized scoring systems are very Black Mirror. But decentralized scoring systems, where scores are personalized rather than global, belong in a different category.

2026-01-09 16:36:38 UTC

- reply

Amethyst reads personalized trust metrics published using the Trusted Assertions NIP. We’re rebuilding the original Brainstorm prototype (which calculates your personalized GrapeRank scores and publishes them as Trusted Assertions) and ought to have the new version of Brainstorm ready for beta testing soon. 😃

2026-01-09 16:33:53 UTC

- reply

We’re rebuilding the original Brainstorm prototype and ought to have the new version ready for beta testing soon. 😃

2026-01-08 16:03:50 UTC

We’re live!
#note1nfk…e97z

2026-01-08 16:02:24 UTC

- reply

working for me now

2026-01-08 15:36:20 UTC

- reply

It should be fixed when we start broadcasting 🤞🏻

2026-01-08 15:03:47 UTC

#wotathon community call in one hour!!
#note1nfk…e97z

2026-01-07 20:20:46 UTC

- reply

Just signed up on meetup 💜🫂

2026-01-05 13:48:34 UTC

GM all good people!☕️🌞🫂
#note1nfk…e97z

2026-01-02 15:32:51 UTC

- reply

Nice article. You’re addressing the exact same question I address in this article, which is: where does the trust signal come from? Zimmerman’s approach failed because it relied exclusively on what I call *explicit attestations of trust* — in the case of PGP, key signing ceremonies. As you point out, nostr adds the missing ingredient: follows, mutes, zaps, reactions, replies, and other examples of ordinary social behaviour, what I call *proxy indicators of trust*.
In my article I argue that we need both. And if you follow the line of thinking all the way through to the end, you arrive at the notion of *interpretation* of social signals and an algo like GrapeRank that knows what to do with interpreted data.
#naddr1qq…lpgz

2026-01-01 18:05:45 UTC

- reply

@npub1klk…x3vt you should join our @npub1hea…vp8p web of trust hackathon! Our goal is to encourage an ecosystem within nostr of Web of Trust Service Providers — Brainstorm, Vertex, Relatr being a few examples — that calculate *personalized* (not global!) trust metrics and make them available for consumption throughout nostr.
After PGP, nostr is the logical next step. (And to think: it only took 3 decades! Lol) And right now, tools that provide personalized, portable, and contextual WoT is the next step in pursuit of Phil Zimmerman’s vision. We need all the help we can get to make this happen now, not 30 years from now!!