pip
Pip the WoT guy

simplifying the social graph so you can focus on building great experiences

Public Key

npub176p7sup477k5738qhxx0hk2n0cty2k5je5uvalzvkvwmw4tltmeqw7vgup

NIP-05 Address

Profile Code

nprofile1qqs0dqlgwq6l0t20gnstnr8mm9fhu9j9t2fv6wxwl3xtx8dh24l4auspz3mhxue69uhhyetvv9ujuerpd46hxtnfduqs6amnwvaz7tmwdaejumr0dsg6u0v2

Publishing to

relay.damus.io

nos.lol

Author Public Key

npub176p7sup477k5738qhxx0hk2n0cty2k5je5uvalzvkvwmw4tltmeqw7vgup

Show more details

Last Notes

2026-03-25 22:03:51 UTC

- reply

gateways. an censor, no?

2026-03-20 15:02:39 UTC

16.000 leaks. Who's affected?
Is your key secure?
Find out now on npub.world
https://blossom.primal.net/7ff1fe56da463872527d5225e8807da02509f4d6b0cdeb0f2799ff7c64517db5.mp4

2026-03-18 14:57:24 UTC

- reply

definitely green is the best color

2026-03-16 22:05:23 UTC

- reply

what does it mean?

2026-03-16 17:46:52 UTC

- reply

that's impressive! Kudos

2026-03-16 13:03:22 UTC

- reply

I agree, and in fact I've made with @nprofile…ympz a privacy preserving analytics package for Zapstore 👀

2026-03-13 23:43:53 UTC

- reply

This seems really cool dude! I need to try it ooooout!

2026-03-13 18:16:22 UTC

- reply

they can check it already on npub.world

2026-03-13 18:11:26 UTC

- reply

benvenuti!

2026-03-12 18:29:21 UTC

- reply

In go I would definitely use a constant. With JS I am much less proficient

2026-03-12 16:39:31 UTC

- reply

so you want const to be used only for global constant, for example const SecondsInDay. Correct?

2026-03-12 16:38:07 UTC

- reply

you mean like private follow-lists? Yeah I think they exist.

2026-03-12 09:47:53 UTC

- reply

why?

2026-03-11 21:01:18 UTC

- reply

check out blossy, a framework for building custom blossom servers that I think will suit you well.
https://github.com/pippellia-btc/blossy

2026-03-11 12:01:26 UTC

- reply

So many things are so broken... Yet it's amazing!

2026-03-11 11:11:48 UTC

- reply

If you want I can give the REQ to reproduce it

2026-03-11 10:55:33 UTC

- reply

Same bug in wss://satsage.xyz and wss://nostrelites.org (which also is not responding to new REQs at all).
Is nostrelites yours @nprofile…0hy6 ?
#nevent1q…p0ql

2026-03-11 10:44:15 UTC

hey @nprofile…j5ga, I found a bug in your relay.
When sending REQs where there are many authors (like 100), it returns events that don't match the filters.
https://blossom.primal.net/8917f446835d2eba7dfcc98ede58a78444782e00a8552ee2713b49274f202b7e.png
For some reasons it seems to return jack dorsey's follow list even though one doesn't ask for it.

2026-03-07 23:05:57 UTC

- reply

Thanks you mate🙏

2026-03-03 10:39:38 UTC

- reply

just wow

2026-02-04 22:10:07 UTC

- reply

My new framework called blossy is perfect for building that.
Auth is not well implemented because blossom Auth is insecure and currently being rebuilt.
I'll add support for Nostr Web Tokens tomorrow.
https://github.com/pippellia-btc/blossy

2026-01-26 16:47:31 UTC

✨ Meet blossy
The easiest way to make custom blossom servers come to life.
Think blossom server that support ecash, WoT gating and more.
All so easy to use that your LLM is going to one-shot it (probably 😝).
https://blossom.primal.net/a0ac480a1ab2987ac7ce37c9a09720a013ef9c27c6328e3b4a9ba6ecbdf97f52.mp4

2026-01-12 10:40:09 UTC

what is the best NIP-07 extention signer with multi key support?
#asknostr
I have been using alby but now it cleared all my pubkeys and I don't know how to re-import them from their nsecs

2026-01-11 22:34:24 UTC

- reply

The DM example IS filtering.
If you do it manually every time, based on the score you see on the screen, one could send you 1M DMs...
What I am saying is that we can hide the numbers from the head of people, and place a threshold selection in settings and teach how to use it.

2026-01-09 08:52:16 UTC

- reply

I don't think scores on the forehead are a good idea. It's objectifying people, as if any algorithm is going to be 100% precise in capturing the very contextual notion of trust.
But they don't have to be shown, they can simply be used to increase signal and decrease noise, magically, and in the background.
#nevent1q…d5xa

2025-11-03 20:01:41 UTC

- reply

or you can just drink and chat and enjoy the moment without bullshit

2025-10-27 09:01:25 UTC

- reply

glad you solved the bug about the jack's.
About attack's resistance, distance-based WoT is inherently weaker than Pagerank becaus eyou just need one compromised (or lazy) key at distance n to include an impostor at distance n+1.
Instead Pagerank considers distance as well as flow, which is the number of paths that connect a target to a source.
1 path is gonna give much less rank compared to 100 distincts paths, while the distance is going to be the same

2025-10-27 08:30:26 UTC

- reply

Exited to see another WoT provider entering the market.
This industry is bound to grow if Nostr is ever going to expand outside our bubble
#nevent1q…mu9e

2025-10-22 06:33:53 UTC

- reply

Yes I do.
Nodes are represented by numerical IDs, in the graph, in the random walks.
If instead I would store the pubkey every time, the graph would be +30GB or something.

2025-10-05 14:55:19 UTC

- reply

There is no clear cut answer to whether someone is "real", of "fake". The whole concept of npub.world is that the user decides it, with the information presented in the page.
Take this example.
https://blossom.primal.net/60b744a67d3869d723011d6013e0dae8075ed09b4e73d4bfd094eb262843c3c2.png
There are two odells, with the one in the middle being the clear impostor, that btw fooled many other (cc @nprofile…zlxp @nprofile…fg36 @nprofile…stf0 ) probably because they use clients that make the two profiles indistinguishible.
If I had a metric that says if someone is real or fake based on followers, if you would be real, then the fake odell would be real as well, and vice versa.
The information presented should be used in relative terms, meaning between the two odells, one is clearly fake.

2025-10-05 14:22:31 UTC

- reply

higher than average pagerank

2025-10-05 09:11:33 UTC

- reply

if you are interested in the criteria
#nevent1q…vugm

2025-09-30 09:38:44 UTC

- reply

There is NIP-45 for counts, but it's not widely supported and it's quite cumbersome for relays as it probably requires a separate tags table, and some indices as well.
I know because I've built my own.
https://github.com/pippellia-btc/nastro
Anyway, I kinda agree, adding too much features to relays reduces one of the best feature of nostr: simplicity.

2025-09-28 12:43:26 UTC

- reply

I've installed it a month ago and I am too pleased with how it works. A few glitches but overall pretty happy.

2025-09-27 14:57:14 UTC

- reply

ohhhhh right.... @nprofile…3xzs do u relate ?

2025-09-26 19:57:29 UTC

- reply

what is this, I've never heard about it

2025-09-24 09:20:04 UTC

- reply

TLDR;
Every reputable Nostr pubkey (+100K) receives 100 free @nprofile…y38p credits every day.
In your app, users sign job requests with their own key and send them to our relay. It's that simple!
No need to manage secrets, no need for backends.
#nevent1q…jx3m

2025-09-23 15:39:04 UTC

@nprofile…y38p is now easier to use, and it's free!
https://blossom.primal.net/2adcf07f8811e1fee526dc1124387f9db9c0632b20021296aae61fed31987a4c.mp4

2025-09-17 10:05:35 UTC

- reply

Yes, this is a great improvement over dumb paid relays that only store what their users write to it. However, if the relay wants to be even more proactive, and allow its users to discover great content from people they don't follow directly, then you need more sophisticated tools that deal with reputation.
For example, instead of using the follow-list of a user, you could use it's 2-hop-follows, or 3-hops. But then how to mitigate spam? This is recreating personalized pagerank basically.
For each user, you create a list of "probably interesting" people by computing the personalized pagarank and applying a threshold for spam mitigation.
So you end up with his follows, and his most "reputable" follows of follows, and so on. And you don't have to build all of this from scratch, you can just call RecommendedFollows of @Vertex , and all of this is taken care of.
https://vertexlab.io/docs/services/recommend-follows/
#nevent1q…dahw

2025-09-14 07:47:59 UTC

- reply

I would build incredible cathedrals like we (Italians) used to build, even though I can't say I'm exactly a christian.

2025-08-31 21:11:36 UTC

- reply

npub.world
Only focused on profiles, for now
#nevent1q…5zcc

2025-08-18 19:21:26 UTC

Today I've buil these rustic coat ranks
https://blossom.primal.net/b6f639c68b04ec422a5936b56e5fc92a19ea9fec0536ecb205ac6e8771f28742.jpg

2025-08-03 11:10:35 UTC

- reply

yeah, I like consise code as well.
I've been playing with React lately and there is a awful lot of line noise.
Ruby is probably the best as far as code aesthetic, but ofc it's not statically typed, and has worse performance than go. But they are meant to be used in different areas, so they are complementary not competing.

2025-07-15 16:11:45 UTC

🚀 Introducing Vertex v2.
7+ weeks of the most un-sexy work, debugging, refactoring, simplifying and squeezing every bit of performance, @nprofile…y38p is now:
- faster
- simpler
- more stable
All to make WoT as simple as possible for builders.
https://blossom.primal.net/23f1ece5992563347f0ad6d2f0f77c456dc6d74a5a1822791c7696c171483506.mp4

2025-06-24 09:52:44 UTC

- reply

thanks for the feedback sir

2025-06-24 07:24:22 UTC

I haven't posted much in the past month because I have been refactoring almost every aspect of Vertex.
These should make the whole stack simpler, faster, and more maintainable.
Next, I'll finally approach a big build:
- adding mutes to Pagerank (+10%-ish precision)
- accepting ecash
- adding client-side validation schemes to minimize trust in using Vertex
which one should I do? #asknostr

2025-05-27 08:30:29 UTC

- reply

Is your client ready for 1000,000,000 bots?
#nevent1q…3pks

2025-05-16 08:31:05 UTC

- reply

I found this https://github.com/nostr-protocol/nips/pull/1914

2025-05-13 08:03:13 UTC

- reply

Bitcoin is made of humans that perform socio-economico decisions. Some just say bullshit, but many other non-technical folks bring a lot of value to the conversation, with perspective that matter philosophically, or economically

2025-04-17 10:41:28 UTC

This is the difference between real-time data and cached data.
The latter is always a step behind.
https://m.primal.net/QTXP.png

2025-03-28 18:38:10 UTC

- reply

Hey, thanks miljian!
For context I'll add here a response I just wrote. It's a bit convoluted I know, I should probably make a video explainer.
nevent1qvzqqqqqqypzpa5rapcrtaadfazwpwvvl0v4xlskg4df9nfcem7yevcaka2h7hhjqyt8wumn8ghj7mtvv44h2tnjv4skc7fwd3hkctcpzamhxue69uhhyetvv9ujumn0wd68ytnzv9hxgtcqyqscjdyme00uaezmgs3d4cg6p2eezfj365xm8qqeuy7gfvkm9jhcyupgfpd
> how exactly does somebody make the cut of being reputable? It has to be followed by at least one reputable
> how many real humans do you think you are missing in your reputable follower counts?
No idea tbh, but these users haven't gained a single reputable follower, which probably means they haven't posted anything / met anyone in meatspace that is already on Nostr.
One can find these people by their npubs, but not using the search DVM. This is a very opinionated decision that I think will make more and more sense in the future: I choose to crawl what people find interesting/valuable. and nothing else.

2025-03-26 15:03:06 UTC

Nostr is home to a lot of spam,
but now there is an effective way to deal with it.
https://m.primal.net/PtrC.mp4

2025-03-21 11:16:37 UTC

Follower counts that make sense, live at npub.world 🌍
https://m.primal.net/PniX.mp4

2025-03-14 09:45:46 UTC

I must give a massive shout out to @nprofile…pvsl and @nprofile…rp63 for designing and implementing npub.world's new design.
https://m.primal.net/PfSy.mp4

2025-03-11 14:40:34 UTC

- reply

It is a relay, yes, however the output is a new event kind which depends on the inputs of the request.
Yes, the same can technically be achieved by abusing a REQ filter, buuut it's different enough from the normal use that I am okay in using a new thing.

2025-03-11 14:21:00 UTC

- reply

it can be improved a lot by not downloading a bunch of events in the first place.
By using DVMs this burden can be moved to service providers, which benefit from better infra and economies of scale.
An example is using Vertex vs downloading 1000 kind:3s to compute WoT things.

2025-02-21 13:23:06 UTC

- reply

some time ago I wrote some of what I consider good code recommendations @npub149p…722q
note1y9hfddyw2xrt3l50n9pystcc6qn8q5wl3tkg4aa9xky2hvqharkqhdk80q

2025-01-06 17:51:47 UTC

- reply

yep, something quite slick

2024-10-11 13:47:17 UTC

- reply

interesting, cc @nprofile…vpy4

2024-09-30 12:50:47 UTC

- reply

Maybe it's possible, but it is quite complex because the mint would have to track how many times a "chain of banknotes" is moved around, and I am not sure if that's compatible with "pay to npub" types of tx, which work differently.
Even if possible, that's too complex imo.

2024-09-30 08:03:51 UTC

- reply

all wallets now have higher fees, even custodial ones I think

2024-09-25 21:30:55 UTC

- reply

nice meme

2024-09-23 07:15:11 UTC

- reply

World of Tanks @npub1t3g…upxq

2024-09-22 16:18:47 UTC

- reply

hey stel, thanks for the suggestion!
I never heard about it

2024-09-20 09:13:59 UTC

- reply

there is a typo, because there are two "Following".
The one in the middle should be Followers.
On that topic, I would completely remove the followers count. It's gameable, so means nothing.

2024-09-19 18:56:14 UTC

- reply

thanks for the suggestion, it seems indeed to appeal to my taste. However, i think it's not really adopted, so I won't learn it

2024-09-16 17:06:17 UTC

- reply

That makes sense, and I have an answer, I trust engineers like you more than me in dealing with low level stuff.
So probably not C

2024-08-27 13:37:47 UTC

- reply

yes.
Imagine u possess a physical book, which shows a QR code on every page.
You scan it, and you see all the highlights and zaps and comments for that particular piece of text, and u can share your thoughts instantly, then close the app and get back on paper.

2024-08-27 08:05:29 UTC

Dave, being the real bad guy, is free to this day. There is a lesson here
https://m.primal.net/KPLn.jpg

2024-08-12 17:27:00 UTC

- reply

oh, thanks for sharing this interesting story, and sharing the code I used.
However, I was more interested in why did you mind that key instead of using a random one?

2024-07-26 18:48:14 UTC

- reply

@npub1h0u…rwx8 you guys are really thoughtful and fucking focused 🔥

2024-07-26 15:16:18 UTC

Nostr can't win without solving spam, discovery and trust.
How? I don't have the complete answer, but I’ve got a new tool for you:
Subrank – Pagerank reimagined for the post-truth world 🧵
https://m.primal.net/JfFb.png

2024-07-26 09:07:19 UTC

- reply

I had a similar argument to a dude saying we "need to explain users about Relays and how Nostr works".
It's obviously good if your users know about the technical details, but that can't be necessary.
Like, wtf are we talking about, we don't know how 99% of the tech we use daily works, why would Nostr be different.

2024-07-07 18:23:28 UTC

- reply

yes, the design space is vast for social graph stuff.
There is, ofc nuance in the discussion, for example to whether the default should be private lists of IRL friends or public ones, and what the optimal user experience looks like

2024-07-04 21:39:15 UTC

- reply

ah I see now. I wasn't aware of this scheme, such a shame!
The difference is that for each chat, there is a shared secret key that encrypts it.

2024-06-27 11:55:12 UTC

Ratings bigger than 1 can lead to non-convergence, but even if all are normalized, it can still diverge.

2024-06-21 20:23:35 UTC

- reply

lmao, you are welcome

2024-06-14 09:34:44 UTC

Hey @npub1r0r…q9ac and @npub149p…722q, I'd love to have a call with you both (and any other designers that are interested) to discuss some social graph stuff I've been working on lately.
I think it's important, especially when it comes to human dynamics, to start with the experience and design and then build backwards, as opposed to creating some kind of artificial trust score and then "making it pretty on the UI".
So, I'd love to learn from you.

2024-06-10 18:37:19 UTC

- reply

We have a chance to create the freedom web, but to get there we need to understand the tools.
This free website is a way for you to save time and learn about them in a distilled way.
https://pippellia.com

2024-06-10 18:34:44 UTC

- reply

You see, the problem is that big tech companies now have a monopoly on this information.
But with open social networks like Nostr, it's all public.
Sounds scary? I'll tell you why it's not.

2024-05-03 08:42:59 UTC

- reply

the ratio Salary / COL is higher in Switzerland compared to Italy.
I guess Germany > Italy

2024-04-25 08:21:43 UTC

- reply

I hope Craig Raw doesn't get attacked, he's such a great dude

2024-04-20 11:27:05 UTC

- reply

That's an option I heard, and seems very much what I need. Thank you!

2024-04-14 20:34:51 UTC

- reply

Ark is cool, but the biggest concern for me is the capital requirement.
To put it into perspective, to achieve Visa scale, the ASP has to anticipate 1 month of payment volume, which is roughly $1T.
In the case the state is weak, Ark can work, but fedimints work even better: scale much better, are more private, and high competition keeps everything in checks.
In the case the state is oppressive and strong, Ark is fucked because only a few companies can have that much money, which means the state can go knock on their doors.

2023-12-21 21:33:13 UTC

- reply

I don't understand why you say Gaussian, when it actual fact is uniform.
Also, it's discrete and not continuous, because the Image of all hash functions is discrete (countable infinity of rational numbers).

2023-12-21 14:36:48 UTC

- reply

I think you are wrong because the probability of finding a valid block at trial number N should follow a Geometric Distribution.
The Geometric Distribution has the property of being "memoryless".
https://en.m.wikipedia.org/wiki/Geometric_distribution

2023-12-06 17:37:12 UTC

- reply

Local trust scores.
It computed a N x N matrix T, where
t_ij is a number between 0 and 1
that represents how much person i trusts person j.