fiatjaf

https://stuff.fiatjaf.com/

Public Key

npub180cvv07tjdrrgpa0j7j7tmnyl2yr6yr7l8j4s3evf6u64th6gkwsyjh6w6

NIP-05 Address

_@fiatjaf.com

NIP-57 Address

npub180cvv07tjdrrgpa0j7j7tmnyl2yr6yr7l8j4s3evf6u64th6gkwsyjh6w6@npub.cash

Profile Code

nprofile1qqsrhuxx8l9ex335q7he0f09aej04zpazpl0ne2cgukyawd24mayt8gprfmhxue69uhhyetvv9ujuam9wd6x2unwvf6xxtnrdakszxthwden5te0wpuhyctdd9jzuenfv96x5ctx9e3k7mgqz5704

Publishing to

relay.westernbtc.com

pyramid.fiatjaf.com

countries.fiatjaf.com

Author Public Key

npub180cvv07tjdrrgpa0j7j7tmnyl2yr6yr7l8j4s3evf6u64th6gkwsyjh6w6

Show more details

Last Notes

2025-11-13 14:30:52

- reply

Nothing is ever absolutely forgotten.
As I said, it's not like you keep remembering that specific thing, it gets mixed in with your other memories.

2025-11-11 17:48:16

- reply

Good. I'll do it.

2025-11-11 14:38:10

- reply

It was good while it lasted.

2025-11-11 12:28:05

- reply

@npub1jlr…ynqn I was going to invite @npub1syj…f6wl here but then I realized anyone else could snatch the invite code if I pasted it here.
Do we need a way to restrict invite codes to specific pubkeys?
Or maybe a process to just add someone without having to share a invite code?

2025-11-11 09:57:43

- reply

How can I generate an invite code from inside Jumble?

2025-11-11 09:56:16

- reply

How would that work? Do you have to provide a list of what you have somehow in order for it to work, right? How could that be done?
Currently there is `nak req --ids-only` which uses negentropy underneath but with an empty local set so you get all the ids from the server.

2025-11-07 01:38:17

- reply

Are you sure these users aren't all AIs?

2025-11-06 21:13:58

- reply

I imagine @npub1der…xzpc's army of robots can turn that JS server into a Rust server or whatever in a few seconds.

2025-11-06 21:13:01

- reply

Yes, I assumed they would work.
What I was saying is that they wouldn't work if redirection was done on the client like @npub1gcx…nj5z suggested.

2025-11-06 21:10:55

- reply

Would be nice if njump.to kept some short-term memory of the domains it last redirected each path to, because those would already be cached on the resulting server, so it would be better to send the same path always to the same server.

2025-11-06 21:09:21

- reply

I agree it was very broken for a long time, but it has improved a lot in the past weeks. At least I can reliably get URL previews on Telegram every time now.
But I insist that URLs/domains are a problem, if you redirect people to njump.me that requires that njump.me is up there to handle it. Even if it was much faster it could still become evil someday, or go down. Now if you send people to njump.to so it can split traffic you move the problem to Gigi.
In any case, a server that just splits the traffic and returns redirects can handle 1000x more traffic than a real njump server that has to do all the relay and local cache dance, so @npub1der…xzpc's JavaScript server is probably fine for now.

2025-11-06 20:59:37

- reply

There is no solution to the problem of the single point of failure when dealing with URLs. The web is rotten from the inside and that is why Nostr should be just native apps.

2025-11-06 20:48:55

- reply

The entire point of njump was always to render URL previews on other platforms (all the rest is fluff that maybe shouldn't have been added), and that breaks it.
If you're going to do anything on the client then I don't see why it isn't much better to just fetch the events on the client and just display them.

2025-11-06 20:46:27

- reply

I think it's some reverse proxy routing magic that detects if the path is a "/nevent", "/npub" etc and uses the njump server, otherwise uses the landing page.
It could be changed to detect email-like or domain-like URL paths too, and "/p/..." and "/e/..." paths that njump supports.
Or njump.to could do that work and not redirect to nostr.com when the path is a NIP-05.

2025-11-06 19:34:57

- reply

What post?

2025-11-06 17:24:25

- reply

Why is this a dig at Keynes? It's just truth that even Keynes's admirers accept.

2025-11-05 01:20:02

- reply

Bring it: #nevent1q…wrht

2025-11-05 01:18:20

- reply

I agree, Lightning sucks, activate Drivechain.

2025-11-05 01:17:29

The shortest complete NIP-34/GRASP pitch you'll ever get: #nevent1q…hqse
(Also GitHub is run by pedophiles.)

2025-11-05 01:13:57

- reply

Let's fire Core. Run the Drivechain activation client.

2025-11-05 01:13:13

- reply

Why do you love spam so much?

2025-11-04 12:23:17

- reply

Should be fixed on the latest master, or commit 210c0aa2820536f3bb947faed93a334c107961c9.
Please install it with `go install github.com/fiatjaf/nak@210c0aa2820536f3bb947faed93a334c107961c9` and let me know if it's still broken.

2025-11-04 12:12:18

If you love the terminal and you have a server and nothing else to do please try to run the totally new and much more fantastic version of the pyramid relay: https://github.com/fiatjaf/pyramid/issues/4
There will be a nicer installer later.

2025-11-04 11:32:05

- reply

So can we fix that if all apps start using these flags everywhere? @npub1gcx…nj5z @npub14ya…sn39 @npub1w4u…0jr5

2025-11-04 11:14:28

- reply

Weird, neither nak nor the client library impose any timeouts.
After writing the line above I figured out that there is indeed a timeout of 10s in a place I didn't remember because I'm using a weird HTTP library only for the blossom stuff.
I'll increase that and make it configurable.

2025-11-04 02:32:17

Why is Android so stupid that if you open an app B from a link inside app A then app B is somehow opened inside app A, not as an independent process?
Now you can't alternate between the two apps because there is only a single app B opened. In order to go back to app A you have to press "back" on app B many times and hope it closes and control is returned to app A.
Even if you try to reopen a new instance of app A it will just bring you back to app B that is running inside app A. It's impossible to do anything.

2025-11-03 16:23:31

- reply

(By which I mean they are capable of ingesting a lot of content then they can spit out a basically infinite number of sentences that are grammatically very correct but empty of content.)

2025-11-03 16:17:34

Are journalists the precursors of LLMs?

2025-11-03 16:13:02

- reply

I think of litter bins as free storage devices too, but I've come to the conclusion that litter bins are a horrible protocol because every time I put things there they either come back dirty or I can never get them back at all.
That's why I always advise people against using litter bins.

2025-11-03 09:39:33

- reply

What do you have against NIP-34 and GRASP?

2025-11-01 16:09:33

- reply

Yes, it should be a local thing, shouldn't be hard, store it temporarily or something. People have tried to do it with events in the past, but that is an overbloated complication.
Still, it would be nice if well-behaved clients stopped mentioning people after a while, it helps no one.

2025-11-01 16:07:20

- reply

Clients should nudge users towards using such relays for DMs, and that is happening right now in every client.
Relays were never meant to be dumb in the sense that they wouldn't have any internal logic, that was an unfortunate choice of words and a huge misunderstanding that damaged Nostr a lot, it's a surprise to me that people could have thought a relay would be really that dumb, do no filtering, no access control etc and Nostr would still work? Nostr would never work if it was like that.

2025-11-01 16:02:25

Would be nice if clients dropped "p" tags from replies if the referenced public key hadn't participated in the discussion for some number of notes.
Someone mentioned me in a thread once, then 20 notes later I'm still having to read @npub1gcx…nj5z and @npub1226…grkj insulting each other about encryption.
It makes me think encryption is bad by itself. Encryption and satellites.

2025-11-01 11:11:42

- reply

You do have a way of knowing: if your DMs are going to a semi-trusted relay and you have to AUTH to it in order to read them (as you should under NIP-17) that relay may inform you about who has read them, when and from what IP address.
If we implement https://github.com/nostr-protocol/nips/pull/1647 that will also kill the attacker ability's to read anything in the first place unless he wants to reveal himself by registering his bogus device or publishing new encryption keys.

2025-11-01 02:19:50

- reply

There are a dozen of issues with this scammy paper.
The most important is that it only worked on a couple of clients that didn't check signatures. These clients only connected to a static set of semi-trusted relays and changing the relays they connected to would require a manual typing operation from the user.
For the attack to work it required victims to manually type the URL of the attacker relay, which makes it completely absurd.
It's like telling someone to visit "verysecretnotscammywebsite.com" and type all their secrets there, then read their secrets because the website leaked them and write a paper claiming that the web is broken.

2025-10-30 19:26:47

- reply

Aren't you also against inscriptions? Shouldn't knots also filter inscriptions?
Maybe it does that already, I don't know.

2025-10-30 14:57:58

@npub1jjn…vknm when are you going to host a nice reliable durable cheap GRASP-powered git server for the 4 of us that use ngit?

2025-10-30 14:22:36

- reply

Now it's much harder to play, every move I try to make it complains about something bogus.

2025-10-29 18:42:39

- reply

I think it's a bug, it should have matched by id only. But your report will be ignored because I've changed these database implementations multiple times (uncovering many bugs that lead to duplicated copies being stored) and in the end I've decided that I don't want to support Badger anymore.
I'm surprised you even managed to export all the notes given that the version you're using would try to load everything into memory before printing and if you had too many that would cause a crash. On the other hand there was a hardcoded limit of a million notes per query when using the eventstore CLI that you couldn't bypass, I guess that prevents the crash.

2025-10-29 18:38:33

- reply

@npub1000…vwqk when this note is showed in the grouped view in https://fevela.me it doesn't include the URL, but also doesn't include the preview box, so it's weird.
I kinda wish the preview boxes could be disabled.

2025-10-29 16:47:56

- reply

This was a great solution, although I have no idea how it could be implemented.
Same for the code that decides who won. I was almost amazed at it when it committed a mistake:
https://image.nostr.build/d56d6da2af7dfae2e3de8dcc79cecdd24273d9fb9ea587f60b7202ef1d06f730.jpg
By the way, if you get back to it, I think we need more space around the squares still (and smaller squares).

2025-10-29 11:25:45

- reply

Classic, everybody always tries to do this.

2025-10-29 11:23:39

- reply

It's almost as if kind:2 is back.

2025-10-29 02:33:44

@npub1jlr…ynqn what is the difference between https://github.com/coracle-social/zooid and https://github.com/coracle-social/frith?

2025-10-29 01:46:00

- reply

Looking pretty good, now just some more space around and between the squares (maybe shrink the squares a little?), disallow going through the square and make the computer recognize a win condition and that should be it.

2025-10-29 01:43:14

- reply

I'm not sure it's worth playing seriously on the computer, but I think having something visual is good so people can learn and then play on paper when they're waiting for their food at a restaurant or something like that.

2025-10-29 01:41:41

- reply

I've attached a picture now.

2025-10-29 01:39:30

- reply

This is pretty good, but yes, without some pathfinding around the squares and avoiding existing lines it doesn't work: #nevent1q…h2ca

2025-10-29 01:36:31

- reply

I forgot to attach an example game to the explanation: https://image.nostr.build/4e564b953116ae68259614dd2b6ad870ffd1a0a4f6a2f2b0691c6316a13b8881.png

2025-10-28 22:21:56

If someone is a real vibecoder and wants to show it please make this game in a computerized version: #nevent1q…ha5l

2025-10-28 19:29:03

- reply

Yes, I would never merge huge PRs myself either. And more clients is always better, even (or specially) if they are very similar but with subtle differences.
The problem is that forking and keeping track of upstream is an impossible inglorious task.

2025-10-28 11:44:43

- reply

Time to join forces with @npub1m4n…c2jl and launch a dissident Jumble fork with a different, name, design and feature set.
And remove translations.

2025-10-27 19:10:34

- reply

Não vejo por que não. A física não permite nenhum argumento contrário a essa afirmação.

2025-10-27 10:33:21

- reply

After doing all of https://spotify-scrobbler.fiatjaf.com/ I learned that Spotify doesn't allow new integrations to connect to data of external users.
In order to enable that you must sign a contract with Spotify and they will only look at you if you have 250k monthly active users.

2025-10-27 09:47:11

- reply

ahahah

2025-10-27 09:46:39

- reply

Yes, I did. I once said this was like a 2D version of Nim, but no one understood because no one else had ever heard of Nim.
Wait, now that I'm thinking, when I was a kid my grandfather once played a game with matches with me, he put a bunch of matches on the table then we took turns removing either 1, 2 or 3 matches until one couldn't.

2025-10-27 03:24:25

- reply

Can you make a computer implementation of this game? #nevent1q…ha5l

2025-10-27 03:23:22

Joseph's Squares is a game to be played on pen and paper. It goes like this:
1. draw 2 squares;
2. player A draws a line between one side of a square to a side of the other square;
3. once a side of a square has been connected it can't be used again.
4. lines can't be crossed ever.
5. player B draws another line again connecting one unused side of a square to an unused side of the other square.
6. and so on until a player can't draw any lines, that player loses.
In a game with 2 squares the second player can always easily win, but once you get that you can start playing with 3 squares, 4 squares and so on.
It isn't meant to be the best game of all times, just better than tic-tac-toe.
I haven't tried to play with triangles or other shapes but could be interesting too.
You have to be a reasonable human and not try to draw confusing lines. These confusing lines cannot change the game in theory but they can confuse everything and ruin the game for both players (it's ridiculous that I have to say that but many people to whom I tried to teach the game had this stupid idea, so I'm saying here).

2025-10-27 02:56:51

https://www.youtube.com/watch?v=bpZvg_FjL3Q

2025-10-27 02:56:06

- reply

But aren't LLMs mediocrity machines by definition?

2025-10-27 02:29:29

- reply

Because everybody wants to be like everybody else, i.e. chase mediocrity.

2025-10-24 20:21:17

- reply

Exactly, we all assume too much based on our narrow views of reality.

2025-10-24 20:15:03

- reply

How can you even know what "to work" means if you don't know what the user wants and you don't give the user the option to express what he wants?

2025-10-24 20:13:54

- reply

The user must not know or see anything ever. That is good UX. How dare you.

2025-10-24 20:12:52

- reply

That's @npub1jlr…ynqn's vision.

2025-10-24 17:22:50

- reply

Thank you very much.

2025-10-24 16:16:46

- reply

Yes, if you can do that it will be great and I'll do the remaining ones. Most of these were stored on https://aegis.utxo.one/ blossom server, I didn't know that was down.
I've just removed Flycat and Lume. It's always a great feeling to remove broken stuff and add new exciting apps. Let me know.

2025-10-24 16:10:55

- reply

No, I didn't get anything in ngit.

2025-10-24 15:20:17

- reply

Seems to be working fine: https://image.nostr.build/ef87625d563b88fa9dc8540ce6553c4f0cd03ad087a9348f1ce34417aa31c6a6.jpg
In any case I wouldn't be able to accept such a giant vibecoded patch. What did you tell the LLM so it generated that?
The error page in your screenshot looks unrelated, you tried to open Lantern in the Chrome Web Store or some other chrome:// page and that isn't allowed by Chrome.

2025-10-24 14:17:50

- reply

Isn't it curious that not only Lantern is broken, but also the thing you used to submit the patch?
Can you tell me what is broken?

2025-10-23 22:26:36

- reply

This:
https://gitworkshop.dev/npub180cvv07tjdrrgpa0j7j7tmnyl2yr6yr7l8j4s3evf6u64th6gkwsyjh6w6/promenade
#nevent1q…4nfm
It's actually pretty streamlined already. Just run the server, then run the signers, use the accountcreator to distribute the shares.
Actually, no, in this case there is no distributed key generation, so for now someone will know the full key. But that part can be added.

2025-10-23 22:06:20

- reply

The relay still has to know who are the members, who can publish, who can read etc, otherwise we lose all the NIP-29 benefits.
I think all the metadata-related events can work the same. All the content events can be pretty much the same except it's all encrypted and wrapped in some other event kind that denotes that encryption status, then you add some other messages, sent to the group in the same way, to do the encryption key sharing magic.
This way the same group can have both encrypted-MLS-whatever stuff and non-encrypted stuff at the same time.

2025-10-22 19:22:50

- reply

Yes, exactly.
I don't think the fee market is super great (because it doesn't reward those that bear the biggest burden of carrying the data), but it's the only thing we have.

2025-10-22 19:20:12

- reply

https://lemire.me/blog/2019/12/19/xor-filters-faster-and-smaller-than-bloom-filters/ seems to be a good fit. I have the impression that there is a single design and everybody can implement it the same way. And it's made for static filters, which they will be if we share them in events.

2025-10-22 17:57:57

- reply

Please let me know if I'm wrong, but if both sides realized this objective truth we could finally take some real moves to reduce spam.
Ideally we would transition to some complex client-side-validated scheme like shielded CSV, Tachyon or Intmax2 or whatever (I stopped trying to understand these things deeply since it apparently makes no difference towards our chances of improving Bitcoin).
Since none of that is possible we can take the second-best choice: reduce the blocksize, activate BIP-300 and BIP-301, let people use sidechains with these magnificent antispam ultrascaling properties to make actual financial transactions without burdening anyone else with their spam.

2025-10-22 17:51:41

- reply

Maybe this is obvious but I don't see it mentioned anywhere: everybody cares very little about transactions from people they're not involved with paying other people they know nothing about, it's just spam from the perspective of people not involved.
Also some random message written in an OP_RETURN is still more valuable (for anyone) than, for example, a financial transaction between two objectively evil people, for example, or, say, from mortal enemies of whoever is judging. In this sense a normal transaction can be worse than mere spam.
All of that considered, even your own transactions are still spam after they have been acknowledged as received then spent. At that point they're not even valuable to you anymore.
The entire blockchain history is spam.

2025-10-22 17:46:47

The solution to the knots debate (or whatever it's called) is that the both sides must realize that every Bitcoin transaction is spam.

2025-10-22 13:02:15

- reply

Should be fixed in v0.3.3.

2025-10-22 11:51:30

- reply

What do you mean? It's a single badge for each profile. Or you mean it's too heavy to fetch a badge for every profile an app encounters?
Yeah, maybe.
In that case would be nice to have a standard for publishing lists of people as bloom filters, such that they are interoperable.

2025-10-22 11:36:15

I just realized that ATProto is not a social network protocol at all, it could never be. It is actually a collection of "tools" for building centralized platforms.
#nevent1q…p73y

2025-10-22 11:33:59

- reply

--Bob Dylan

2025-10-22 11:16:42

- reply

Also known as "Vite + React + TS" according to its <title>.

2025-10-22 11:15:46

- reply

By default you could trust badges from @npub1u5n…ldq3 or @npub176p…vgup or @npub1gcx…nj5z but a system like that would allow newcomers to offer different ways to rank that are better or not, or to just provide more up-to-date scores.

2025-10-22 11:07:59

- reply

Maybe issue badges for who qualifies as a human worthy of attention based on whatever algorithm, then client can just look for these badges from your pubkey, or some pubkey created just for this.

2025-10-22 10:59:04

- reply

I know this is probably an unsatisfying self-serving answer, but what is Reddit? Reddit is a list of communities, each of which provides its own NIP-50 search interface. As long as you have that list, search is, in fact, local.
But it is a good point. Maybe every time we think about these other platforms if we can reframe them in terms of relays and clients we can get new insight about how to do things in Nostr.
The current Reddit is in fact one big KYC relay with a list of NIP-29 groups in it, but on Nostr its functions can be abstracted onto multiple relays with different ownerships without much loss.

2025-10-21 23:21:48

- reply

"Global search" is really not that important in the context of social networks.
#nevent1q…qc3z

2025-10-21 23:21:37

I'm subscribed to over a hundred podcasts and I don't think I found any of those by searching randomly by keywords, it was always a comment I saw or a recommendation from someone else.

2025-10-19 19:56:36

- reply

Please add others.

2025-10-19 09:27:44

- reply

Done.

2025-10-19 02:16:58

- reply

https://image.nostr.build/be82a2233fdbb8038e57ccec73c0318ad7222d6043598b4e63b38f68e85dff48.gif

2025-10-18 22:42:31

- reply

I reread this and I mostly agree. I think I had agreed at the time too but couldn't answer because it was too much it was easier to not do anything.

2025-10-18 22:34:47

- reply

"Gravemente comprometido" de que maneira?

2025-10-18 22:31:22

- reply

It's every 6 hours, I haven't touched it in a long time, but it only publishes to wss://nostr.wine/, so it may be because your client is not fetching its notes properly (i.e. outbox), or maybe you only get them when someone republishes them to other relays which may not happen consistently.
Also I didn't get a notification for this reply (probably because your client didn't send to my "read" relays), I saw it by accident now.

2025-10-18 16:38:23

- reply

No, they wouldn't.
And I'm not sure we need a kind numbers repo, although that sounds like a good, harmless feature to me now.

2025-10-18 16:36:30

- reply

They should publish statistics about how many archived pages they have taken down due to legal requests and whatever reasons.
We're winning on that front with 0.

2025-10-18 16:20:19

- reply

What do you think of adding local storage of notes specifically for the following feed + building the feed directly from the local storage + background syncing of new notes from follows from relays? Would you accept a patch for that?

2025-10-18 16:16:01

- reply

What relays will you connect to in order to fetch "any notes with certain hashtags"?

2025-10-18 16:11:51

Lots of smart people in the world do not use ad blockers. How do you explain this?

2025-10-18 10:37:06

- reply

I don't think Cashu or Ark ever gave the impression that they would decentralized later, but for the others it's just willful ignorance or pure careless marketing.

2025-10-17 22:11:40

- reply

Just random people writing guides of "how to do whatever on Nostr" in any format they want.