Laan Tungir

If this is the case [and my guess is it will become increasingly so], then I'm thinking more about how to protect our connections to our agents.
#nevent1q…3m8j

2026-07-18 14:40:53 CEST

https://blossom.laantungir.net/a55c8991973a0f16bf0675e6ef3ddea8fc64981df692843f85ca2a1b9e837990.jpg
#naddr1qv…x44z

2026-07-18 13:09:29 CEST

I need some [black coffee] to wake up.
#naddr1qv…x44z

2026-07-18 12:37:05 CEST

GM
https://blossom.laantungir.net/bdc95fe9bd6fec19914894ec8923b8c8613a6e2ac900247bc1538f6984176ea9.jpg

2026-07-17 20:55:55 CEST

https://blossom.laantungir.net/7c2df85979ef3ba47ba86d8034b319eabc378d71a0ecb0f101262fc57cd4474b.jpg

2026-07-17 17:03:39 CEST

- reply

Agreed. Much improved.

2026-07-17 16:21:30 CEST

https://blossom.laantungir.net/55e4380cefc4fab23776c55008ae153df862798ef29217ff356fd05094ebf1ea.jpg

2026-07-17 16:11:19 CEST

@npub1get…0nfm Hey guys, I think getalby.com is down.

2026-07-17 16:08:06 CEST

- reply

Nice!

2026-07-17 13:32:03 CEST

GM
https://blossom.laantungir.net/669f821f26e56e3f296121848b9bc134c58f6efa7105a00a80d6f847d7bbe391.webp

2026-07-17 02:28:23 CEST

- reply

AU9913 looks like he knows what he is talking about.

2026-07-17 02:25:12 CEST

- reply

KIMI

2026-07-17 02:24:48 CEST

Anybody use KIKI K3 yet?

2026-07-16 21:04:45 CEST

https://blossom.laantungir.net/26002b42cc589cef7a739cf4fe73a245ccac6f692dd85e513b91506920c89b72.jpg

2026-07-16 17:47:44 CEST

- reply

That's awesome.

2026-07-16 16:17:04 CEST

https://blossom.laantungir.net/ba9282826b8dab85271b631a43b14ec99e2ff06d3504f68a3d8c55711ea88fc3.webp

2026-07-16 15:18:55 CEST

- reply

Yes. I think that you have to accept that Bitcoin is fixed supply, and go down with the ship if it is going to sink.
If you think a coin needs tail emission to be viable long term, then switch coins or make a new one.

2026-07-16 13:41:44 CEST

- reply

To say that this is hard to think about is an understatement. I had that belief as well, where now I honestly don't know.
Let me give you somewhat of a counter argument.
There are "emergent" phenomena that don't appear dependent on lower levels of explanation. They appear at higher levels of explanation.
For example, you take two atoms and bump them into each other. They follow deterministic dynamical laws. Fancy way of saying they are like billiard balls on a table. You send them in a particular path, they bump, and will follow a particular path out. The path out depends on the path in.
If you took a film of the two atoms bouncing off each other, and reversed it, you could not tell which film is the forward film and which is the reverse. The laws of motion are symmetric. There is no way to distinguish between the two. You can not see the direction of time in the film.
Great.
But now take a whole bunch of these atoms and put them together, and somehow the second law of thermodynamics emerges. Overall, entropy increases for all these atoms over time. That means that if you take a film of all these atoms, you can tell which way is forward in time and which way is backward. You see an egg break and you know that is forward in time. See an egg come together, and the film is running in reverse. Zoom into the individual atoms of the egg, and you can't see it.
The second law of thermodynamics is an emergent phenomena. You look at deterministic laws of motion of atoms and you don't see it. Put a bunch of deterministic atoms together and it appears.
It is possible that "free will" is like that. That even though the atoms bump around deterministically, free will appears as an emergent phenomena.

2026-07-16 13:03:47 CEST

- reply

https://grokipedia.com/page/Schrdingers_cat

2026-07-16 13:02:01 CEST

"So you are telling me, that we actually live in a multiverse, and that all of 'this' is just one universe and there are many parallel universes with copies of me in them? That's insane. How can you believe such a thing?"
I get it. I totally get it. It is kind of insane. But consider for a second the alternative.
Picture young Laan in Quantum Physics I discussing Schrodinger's Cat https://grokipedia.com/page/Schr%C3%B6dinger's_cat.
[professor] So the cat is neither alive or dead in the box, but is in a superposition of both alive and dead states.
[young laan] What does superposition mean?
[professor] It means the cat is neither in an alive state or a dead state, but in a superposition of each.
[young laan thinks prof is just repeating himself] I still have no idea what a superposition of a cat means. I mean the atoms of the cat are still in the box correct?
[professor] Yes.
[young laan] So the heart atoms of the cat are still there, they are either pulsing or they aren't, right?
[professor] They are in a superposition of both.
[young laan] And you're saying that if I peek into the box, I can't see this superposition of the cat, because somehow as soon as I peek, every atom knows I am peeking and this superposition goes away, and the cat becomes either dead or alive.
[professor] Correct.
[young laan] Well how do we know that this superposition even exists if we can't ever see it.
[professor] {crickets}
[young laan] If a frog looks, does the cat collapse into dead or alive states? A mosquito? A virus? If we put a camera in there, and have it take a picture, do the atoms of the film show a dead or alive cat.
[professor] If you put a camera in there, then it becomes "engangled" with the cat and it also becomes a superposition of states.
[young laan] And so there is no way to observe a "superposition", or even ask questions about it. It is fundamentally unknowable?
[professor] Correct.
[young laan] [[thinks this is bullshit]]
So imagine my relief when I discovered the "many worlds" explanation. In some universes the radioactive atom decays, and in those universes the cat is dead, and in the other universes the cat is alive.
Yeah, that is extremely unintuitive, but at least we are allowed to think again. I can understand a universe with a dead cat, and a universe with an alive cat. Things aren't changing because I look their way. There are not "forbidden questions" in this explanation. It's not religion, it is science again.
In science we don't (or at least shouldn't) judge explanations by themselves. We are always comparing them. Which is the better of the explanations that we have? We go with the best one, and "There are many universes out there" is the best explanation we have.
P.S. Einstein intuited that superposition was bullshit as well.
#nevent1q…lxvx

2026-07-16 12:04:11 CEST

GM.
https://blossom.laantungir.net/7d1cb73f714ac892fd062eae357a5331c1ef5d3674921639d3b91f7cc3d94175.jpg

2026-07-16 04:38:56 CEST

- reply

❤

2026-07-16 04:32:55 CEST

- reply

Nice project. It would be fun to build something along those lines. I have been buying my RNGs, but in some universe I have enough time to make my own.

2026-07-16 04:27:15 CEST

- reply

My thought was that you could batch the events and anchor their Merkle root on Bitcoin once an hour or once a day which is pretty cheap given tx costs currently.

2026-07-16 03:17:47 CEST

- reply

It is the solution to this video as well.
#nevent1q…vfsy
You are not changing the past by viewing something.
In some universes, the light took one path, and in some universes it took the other path. Until the astronomer measures the light, he doesn't know which universe he is in.

2026-07-16 03:00:55 CEST

- reply

Yeah, I just found out about this an hour ago ... crazy.

2026-07-16 02:57:54 CEST

Randomness doesn't exist. @npub1v0l…qj49 is correct.
#nevent1q…vm2m
Every possible outcome of a quantum mechanical state "evolves continuously and deterministically" in some universe, so every universe is deterministic.
Instead of thinking about random events, it helps me to think about some events as being inherently unpredictable. I replace the word "random" with "unpredictable".
Every possible thing happens to me in some universe, but I can't know which universe I am in and where I am going.
From my perspective some things seem inherently random, but it is just me not being able to know which universe I am in.
If Schrodinger's Cat ever seemed mysterious or unsatisfying to you, this is the solution. The cat isn't in some weird "superposition" of dead and alive at the same time in the box. The cat is actually either dead or alive in there, but until you open the box, you don't and can't know if you are in one of the "cat is dead" universes or you are in one of the "cat is alive" universes.
If you want to understand this better, drop this paper in your favorite LLM, and have a discussion about it.
https://blossom.laantungir.net/a7767c90b39398794b0d9815bd39df1c31fc01fa45280912d90800694ff9774c.pdf

2026-07-16 01:58:59 CEST

- reply

https://blossom.laantungir.net/99ec0103680d4aff299d798794dd7c8af15314590f0ce6d411bb190e84cb9f19.png
Having to sign in makes Satoshi cry.

2026-07-16 01:51:56 CEST

- reply

2026-07-15 22:50:30 CEST

- reply

Game got good

2026-07-15 20:12:25 CEST

Just when you think Kevin Kelley is to old to have interesting insights into technology...
https://kevinkelly.substack.com/p/latent-space-as-a-new-medium

2026-07-15 18:14:03 CEST

- reply

I see a fault.

2026-07-15 18:12:28 CEST

https://blossom.laantungir.net/3ad3647c22176e5bef7cd6275643bf3927df8c9639f8dfd7221f4880e2e8e117.jpg

2026-07-15 12:52:17 CEST

- reply

Opentimestamps is working for me. I take that back.

2026-07-15 11:56:58 CEST

- reply

This is what building codes took from us :-(

2026-07-15 11:56:11 CEST

GM
https://blossom.laantungir.net/8bd75b0bb4ae2eda7a6a98904c82ed5a4bf80bc868df91e2b5dbaa638461c965.jpg

2026-07-15 11:48:02 CEST

- reply

I'm curious what you came up with. Do you have a repo I could look at?
Also, is opentimestamps even working? Last time I tried to get it working it failed, and I tried yesterday to to use the website, and it has failed today to even return a proof after almost a full day. Maybe it is me.

2026-07-14 23:17:15 CEST

https://blossom.laantungir.net/ec83c83432a0c5d588839c1ed9f0a57794f0fbb41ce8797afd2bba560200465e.jpg

2026-07-14 17:47:03 CEST

- reply

Live the buzz life Alex!

2026-07-14 17:30:56 CEST

Has anyone built a relay that timestamps all the events on it using OpenTimestamps?
https://blossom.laantungir.net/10e269c1472a5af2fac9bd46ed355186827961bd9bf8aa6b3c8de57736567b65.jpg

2026-07-14 17:07:48 CEST

- reply

I'm getting the green check. Repo not up yet ... soon.

2026-07-14 13:30:31 CEST

- reply

I use @npub16g4…kv4h and @npub130m…l29s
Pay with sats.
I have found that I can use GLM 5.2 for almost everything unless it gets stuck. Will use other models to double check work, particularly cryptography where I want several opinions.

2026-07-14 12:46:32 CEST

NIP-06 is unrecommended???? What did I miss? :-(
https://blossom.laantungir.net/e98f8e16b624774f4134db7aab028cd217b3773c67a5a012bb8ef7790a1baee0.jpg

2026-07-14 11:51:34 CEST

Hmmmmmm. Well that is interesting.
https://blossom.laantungir.net/e5465585824ea2cd2b0ceb4dbe1a56a5f2d40776072529c580cfcc36cd60ab15.png

2026-07-14 11:40:34 CEST

- reply

🥰

2026-07-14 11:38:38 CEST

GM
https://blossom.laantungir.net/a096fcb78a4c9542187ff5b5aa65202668d5c069dc6389e69b5f91998ec1a1cf.webp

2026-07-13 18:51:32 CEST

- reply

That is a hot take right there.

2026-07-13 18:02:51 CEST

https://blossom.laantungir.net/36e76f60cae78205ae9944f540edd76d5df361b7953cdfd1239fe57540bbaec6.jpg

2026-07-13 18:00:17 CEST

- reply

Yuck.

2026-07-13 11:51:21 CEST

GM.
https://blossom.laantungir.net/a7db44b46215df41f4b831144dc9b5e3e74e6a7d0885566462d33172e6d706cf.jpg

2026-07-12 17:17:23 CEST

- reply

@npub180c…h6w6 what is your best reference on Drivechains, and what it would offer?

2026-07-12 14:59:57 CEST

https://blossom.laantungir.net/c31879e627e46114c28c9e3661fe5ddd5cffbdb2a3eef86c98978cdeaba2cc87.webp

2026-07-12 14:23:45 CEST

- reply

I ❤ GLM 5.2.

2026-07-12 14:21:01 CEST

- reply

Which model? Sol?

2026-07-12 12:10:35 CEST

GM
https://blossom.laantungir.net/005d17014a5955c99c5ff2ccdeb61190df6a4cd7b47e9a08287dc137d6dce4a0.webp

2026-07-12 11:44:38 CEST

- reply

Often doing the moral thing is worse for you in some way.

2026-07-12 04:36:33 CEST

Argentine mafia strikes again.

2026-07-12 03:40:57 CEST

https://blossom.laantungir.net/7c27fb18f7a6e8c2b7efc36ce10aa392c02e6962fd6e6048ba29a5526585dfd2.jpg

2026-07-11 23:37:26 CEST

Wooooooo! Go Norway!

2026-07-11 23:09:49 CEST

Let's go Norway!
#worldcup

2026-07-11 23:05:59 CEST

- reply

🖖

2026-07-11 21:25:34 CEST

- reply

I feel seen.

2026-07-11 21:24:48 CEST

- reply

In the future, we are all developers.

2026-07-11 20:16:02 CEST

- reply

My agent's analysis. FYI, I was previously calling this "reckless browser"
Tauri Evaluation: Rejected — does not fit the reckless browser
I evaluated Tauri against the reckless browser's requirements and documented the findings in a new "Considered and rejected" section in plans/options.md.
What Tauri is
A framework for building desktop apps with a web frontend (HTML/CSS/JS in the system webview) and a Rust backend, communicating via Tauri's IPC layer. On Linux it uses WebKitGTK under the hood via the wry crate.
Why it doesn't fit (fundamental mismatches, not fixable)
Language mismatch — Rust vs. C99. The entire architecture is C99: the host, the didactyl integration (C, in-process), and the n_signer integration (C API via nostr_core_lib, direct link). Tauri forces a Rust host, requiring FFI shims at every integration point for zero benefit.
Security model is the opposite of what we need. Tauri is designed to be secure — CSP enforcement, IPC isolation, capability-scoped commands. The reckless browser's premise is stripping web security (CORS, same-origin, TLS cert enforcement, cookie sandboxing). Tauri doesn't expose the low-level WebKitGTK settings needed to do this.
Tauri abstracts away the exact APIs the project depends on. The POC and plans require direct access to webkit_web_context_register_uri_scheme() (FIPS routing), WebKitUserContentManager + JSC (window.nostr injection), webkit_web_context_allow_tls_certificate_for_host() (cert stripping), shared WebKitWebContext (cross-origin identity), and webkit_web_view_load_html() (the browser_render agent tool). Tauri's wry crate hides these — you'd have to fork/patch wry, at which point you're just using WebKitGTK with extra steps.
Tauri builds apps, not browsers. In Tauri, the webview renders your app's bundled frontend. In the reckless browser, the webview loads arbitrary URLs from the internet — it is the browser content area.
Agent architecture is incompatible. The didactyl integration requires multiple webviews managed directly by the C host, window.agent.tools.* injected as native C functions, and browser_render loading agent-authored HTML/JS into named windows. Tauri's model (one webview, Rust backend, IPC commands) has no concept of this.
The POC already proved the right path. WebKitGTK + C99: 17 KB binary, 130 lines of C, real site loaded in under an hour. Tauri would add a Rust framework layer on top of the same WebKitGTK, removing control and adding complexity.
Bottom line
Tauri is the wrong tool — it's a framework for building secure desktop apps with web frontends, not a framework for building a browser that strips security and needs low-level webview control. If a Rust host were ever desired, Servo (Option C) is the purpose-built alternative. Tauri is neither a browser engine nor a C99-friendly embedding path. The decision to use WebKitGTK + C99 (Path A) stands.

2026-07-11 20:03:01 CEST

- reply

Didn't even know about it until you mentioned it.

2026-07-11 17:40:08 CEST

I've been holding off on writing my own browser, but I suppose it is time.
https://blossom.laantungir.net/608ca12f2424a59c8f13582cd5c1df295787435be7b752caec38d08659ea07d2.jpg
I mean, why not have 20 different projects in various stages of completion, instead of finishing one at a time?
"They all work together" I say to myself.
# sovereign_browser
A Linux x86 web browser built in C99 on WebKitGTK, designed around **sovereign identity** instead of the traditional web's permissioned infrastructure.
The thesis: the browser's security model (DNS/domains, TLS/CAs, same-origin policy, CORS, cookie sandboxing) is what forces you to rely on permissioned domains, certificate authorities, and centralized account systems. By deprecating that model and replacing it with **Nostr identity** and **FIPS
mesh transport**, you get a browser that's more capable, not less safe — because trust moves to the layer where it belongs: your keys.

2026-07-11 17:26:01 CEST

- reply

I boosted, so it is now. 😁

2026-07-11 14:39:40 CEST

https://blossom.laantungir.net/52d47be1af976d7f8c299106c0966b9608ec665661bb54a7e254cd3e4b94c99e.jpg
#naddr1qv…hllk

2026-07-11 14:07:14 CEST

- reply

WE ARE ALL WITCHES!
I guess then witches aren't that special. I would look for a more exclusive coven.

2026-07-11 14:02:34 CEST

- reply

I find myself rooting for the rocks.

2026-07-11 13:36:32 CEST

GM morning fellow creators and defenders of FOSS. Rise up, there is work to be done!
https://blossom.laantungir.net/b1161c3b31e2c7a8c1601ea21fdab44c3dbfed14d31a17b9306dd874f57767af.webp
There is no reason we should have a soundtrack to our lives. I'm live now.
#naddr1qv…hllk

2026-07-10 20:39:06 CEST

- reply

🖖

2026-07-10 20:37:59 CEST

GM
https://blossom.laantungir.net/4937690495e518a6b72ba4c7ed1aa5034b5f3125eef6189bcb796dead12c2f35.jpg

2026-07-10 17:35:41 CEST

- reply

Instant follow

2026-07-10 08:27:50 CEST

- reply

2026-07-10 00:42:22 CEST

- reply

The Loudest: Kailh BOX Crystal Jade
The Kailh BOX Crystal Jade takes the crown as the loudest production mechanical switch available, measuring approximately 91 dBA at 10cm. It features:
A fully transparent polycarbonate (PC) housing, making it excellent for RGB builds
A thick metal clickbar mechanism that produces a sharp, piercing click
A 55g actuation force with a 75g bottom-out force
A POM stem with factory lubrication for smooth travel
For context, 91 dBA is comparable to a lawn mower or a loud conversation — your coworkers or housemates will definitely hear every keystroke.

2026-07-09 23:20:47 CEST

- reply

Thanks! Looking into it.

2026-07-09 20:02:10 CEST

- reply

Which link are you trying to use?

2026-07-09 16:48:51 CEST

- reply

If by protocol you mean TCP/IP then yes. ;-)

2026-07-09 16:26:46 CEST

I came across the following picture, and started wondering "What was it like to program a UNIVAC 490, and what could the thing do?
https://blossom.laantungir.net/9bd75ba199288edb9eaad8a8cfe1493cca4c34a3a8d0a87acc732e8a446ab1ef.jpg
Well through the magic of LLMs, I could easily find out. The following web page is a simulation we created of programming a UNIVAC 490.
[clearnet] https://laantungir.net/univac-490
[tor] http://kn2jam4kyz6s5wacyozo3do3d2zsvh4to45uh7xliulwdust6zjuvnad.onion/univac-490
[fips] http://npub1crpldvy49ef8z34wlacwujnfudy4nd7k96aqdx5wgn6ckztz7z8q9t59ud.fips/univac-490
There are 6 sample programs you can run and I have included the instruction set in the reference tab, if you wish to try coding the 490 yourself.
The "Add two numbers" program should be queued up. It is a good one to step through and watch the registers to give you a general idea of what is going on.
The "Calculate Pi" program is a good one to let run, because the page can simulate the actual speed that the UNIVAC 490 processed at and give you a feel for what this puppy had under the hood.
Stay sovereign
[Laan]
*****************************
The UNIVAC 490 was a UNIVAC computer with 16K or 32K words of magnetic-core memory. The words had 30 bits and the cycle time was 4.8 microseconds. It was a commercial derivative of the instruction set that had been developed for the AN/USQ-17 by Seymour Cray for the United States Navy. This was the last machine that Cray designed before leaving UNIVAC to join the early Control Data Corporation.
At least 47 of these machines were made (serial numbers run from 101 to 147). Six were installed at NASA and played important roles in Gemini and the Apollo missions. The U490 had complete control of most or all of the data readout screens in Houston Mission Control. The USAF had two installed, as did Lockheed.
The instruction word format:
f – Function code designator (6 bits)
j – Branch condition designator (3 bits)
k – Operand-interpretation designator (3 bits)
b – Operand address modification designator (3 bits)
y – Operand designator (15 bits)
Numbers were represented in ones' complement.
The machine provided the programmer with the following registers:
Seven B-registers (Address modifying index registers) 15 bits each
One A-register or accumulator 30 bits
One Q-register and auxiliary arithmetic register 30 bits

2026-07-09 13:02:55 CEST

GM
https://blossom.laantungir.net/38c4d2c641f36512d5856c466bc2f110e204844f7a0f597f696d0a0a2230a0b5.png

2026-07-08 23:09:56 CEST

- reply

Does it need to be valid?

2026-07-08 16:22:32 CEST

https://blossom.laantungir.net/121fce0f4ca95767d964c550cd157c0fafe59ab6d6b7a0ae984de2d14d79521e.jpg

2026-07-08 16:13:10 CEST

- reply

Nice car ;-)

2026-07-08 16:12:00 CEST

- reply

I would say truth is objective, but we can never be sure we have the truth.

2026-07-08 15:39:24 CEST

- reply

It's honest work.

2026-07-08 15:03:23 CEST

https://blossom.laantungir.net/bc67dce9e891ded46981d06170576fb9786e096618aced8b8abe4d94e1e47688.jpg

2026-07-08 11:53:47 CEST

- reply

Certainly more so now with LLMs.

2026-07-08 11:51:09 CEST

GM
https://blossom.laantungir.net/ce341a945cb265e0e3e7b3c165da944b87ca6fd8110a01462afccf3272b4d1da.png
https://blossom.laantungir.net/08ddbc5ea99952777579a20cd08f446c62e1ed5d31b9688e53dc2d1239ed8e68.png

2026-07-08 04:36:48 CEST

- reply

That last one is great.

2026-07-07 23:43:25 CEST

- reply

That could be anywhere 😁

2026-07-07 23:40:14 CEST

Yawn.
#nevent1q…5ldm

2026-07-07 22:51:24 CEST

- reply

I was expecting at least one goal by the half.

2026-07-07 22:22:49 CEST

This game is great right from the start.
#nevent1q…d9cw

2026-07-07 22:02:51 CEST

If Switzerland still had private banking, I would have to root for them to win this game, but as it is I'm a free agent.
https://blossom.laantungir.net/9c812c97715784d28ed3c651de94080209b4ca094eb21350fa20305fbc07fa7d.jpg
#worldcup

2026-07-07 21:27:41 CEST

- reply

Looks cool ... zap.store? I don't see it.