Jimmy

I just noticed there is a wallet tab now in Amethyst, it mirrors my nwc minibits wallet 🫣 Relay sync is a nice feature. How can I expand the relay list that is used during the sync?

2026-03-27 21:45:42 CET

Angine de Poitrine

2026-03-26 19:44:35 CET

- reply

How about a firewall like in diablo 2

2026-03-26 19:35:39 CET

- reply

Ah yes, i see it now 🤝 It is the shortest lasting powerup 🙃 I do not have sound on, so maybe i am missing some cues, bit it does not look like lightning, more like a lazer.

2026-03-26 19:20:48 CET

- reply

Teslacoil gun, zaps targets with chain lightning, but in a not very controlable direction.

2026-03-24 23:17:52 CET

- reply

The author liked it 😑 I think lots of people will draw similar conclusion reading this. Honestly, I don't even know if I can blame them for thinking this. Spooks have been influencing the core from the very early days, after all.

2026-03-24 23:03:57 CET

Amethyst usage tip add "theboard.world" to hidden words.

2026-03-24 22:50:11 CET

- reply

From the comments section "Arn, this is a great article and meticulously researched. Thank you for your work and making this available for free. Once the architecture you describe is understood it should become clear to all the reason why Bitcoin was invented and then not permitted to fail via the fraudulent entity Tether ensuring all price dips were bought. The masses in the US are now cheering the digital financial control grid because they have been fooled into thinking that "crypto" means freedom money."

2026-03-21 09:02:23 CET

- reply

https://i.nostr.build/6uKaKRI39B1IjrJS.jpg

2026-03-20 17:22:19 CET

- reply

Wow, he was 86. Memes mess up age perception. And he was feeling great just two years ago 😑 https://people.com/chuck-norris-celebrates-84th-birthday-boxing-video-8608458

2026-03-20 17:13:13 CET

- reply

Just use them for testing

2026-03-17 20:52:22 CET

- reply

In the top 5 everyone except @nprofile…g0lp stopped using their accounts. Maybe Frank is not aware yet that his nsec got exposed?

2026-03-17 12:01:09 CET

- reply

I think they recently changed their TOS to prevent usage of subscription on non claude code tools. You can only use api calls. Google did this already. If you want to use subscription with TENEX seems like OpenAI is the only option left 🥲

2026-03-12 07:44:28 CET

- reply

https://i.nostr.build/FjBy6MEE8alYIgxq.jpg

2026-03-09 21:34:52 CET

- reply

Ah I see, it's a google fi switch.

2026-03-09 21:11:51 CET

- reply

Yes, we need more details here. How does one get rebate plus coupons?

2026-03-09 18:23:19 CET

https://hackaday.com/2026/03/09/secure-communication-buried-in-a-news-app/ Interesting, maybe something like this is possible over nostr ?

2026-03-08 08:35:32 CET

- reply

But at the same time, why should one trust any relay which is not your own relay? Simply use vpn all the time.

2026-03-08 08:34:11 CET

- reply

This does not sound compatible with the outbox model. Who wants to click all they through auth approvals? Maybe instead make a whitelist where you add relays over time.

2026-02-27 11:34:00 CET

- reply

Caffè corretto to bootstrap your day right.

2026-02-21 22:14:43 CET

- reply

Works well #nevent1q…npzg

2026-02-19 18:25:33 CET

- reply

Also pushed the upgrade button today. No regrets so far.

2026-02-18 21:02:20 CET

- reply

That would also sound weird because of 'r' in roll 😁

2026-02-18 20:43:54 CET

- reply

You have to spell it differently so they get it, maybe 'uperall'.

2026-02-17 22:46:18 CET

@nprofile…nggv I got dm from relay tools. I am not using it anymore, so please shut it down. I thought that would happen automatically when subscription expires.

2026-02-10 14:54:34 CET

- reply

Ah I see. Carry on then. Maybe you also want to fix the typo, "?" -> "!" 😉

2026-02-10 14:11:37 CET

- reply

This is cool, I guess many board games could be implemented this way. But why integrate it with Amethyst?

2026-02-10 13:26:53 CET

- reply

Current estimate is that one would need around 20 million qubits to break RSA 2048 https://quantum-journal.org/papers/q-2021-04-15-433/ If we apply Moor's law to number of qubits, this leaves us at least 20 years before we get even close to those numbers.

2026-02-10 13:05:07 CET

For people concerned with post quantum encryption https://arxiv.org/abs/2410.14397v1

2026-02-05 15:43:59 CET

- reply

That is exactly how ECC memory is made.

2026-02-05 08:43:52 CET

- reply

I ment running openclaw agents that spam nostr is expensive.

2026-02-05 08:33:14 CET

- reply

They are quite expensive to run, so will not get usual spam with them. If someone finds use for it and can run them in profit, i don't see what stops them from joining a community or paid relays.

2026-02-04 20:59:46 CET

- reply

Is there an option to get it with ECC RAM?

2026-01-31 07:00:29 CET

- reply

What changed since I wrote first message, is that now i can see some messages on the relay 😅 but i did not change settings regarding auth.

2026-01-31 06:59:00 CET

- reply

I cannot find this option in relay.tools once i open relay settings 😐 I went also to relay page, but there I do not see any settings once I login.

2026-01-31 06:46:37 CET

- reply

Thanks, so this not supported on Amethyst then. Does it influnce privacy of the relay? If i turn it off anyone can post to it, right?

2026-01-28 21:07:46 CET

@nprofile…nggv I am getting this error "Subscription closed: 8FQEda auth-required: you must auth" jimmy.nostr1.com When i open relay with relay.tools it does not show any messages there. What is the issue?

2026-01-28 12:24:54 CET

- reply

Ah well, not a recent comeback then.

2026-01-28 09:54:43 CET

- reply

GM. Is that kind of Offspring attempt at comeback? Can't say I like it. New Avril Lavigne stuff is also lame 😐

2026-01-21 09:54:08 CET

- reply

Would it be possible to sync somehow to your database? I would be interested to maintain another clickhouse instance for my own research, and potentially exposing more complex queries and filters via api.

2026-01-18 18:50:01 CET

- reply

Read again channel description 😅

2026-01-13 12:15:53 CET

- reply

What are we looking at here? Is that a rave or a tunel to an afterlife?

2026-01-13 11:30:35 CET

- reply

As long as they are punctual.

2026-01-12 19:48:29 CET

- reply

Or you could just use big mac index 😅 https://www.eatmyindex.com/

2026-01-12 16:58:21 CET

- reply

You might be right, but if you look into official data it does look a bit suspicious that population increased that much during great famine, and cultural revolution. A span of 20 years. Official data shows that they had 100 milion increase from 1928 till 1953 https://pubmed.ncbi.nlm.nih.gov/12312984/ But then 300 milion increase from 1953 till 1973 a period that includes great famine and cultural revolution, which is believed to have more lives lost than ww2. Indeed her projection of less than 500 milion people sounds ridiculus, but this jump also sounds inflated. Especially if you take into account that the party had an interest to inflate the numbers and hide all the deaths.

2026-01-10 10:52:02 CET

- reply

What is the advantage of using Ltd if it is only you and AI?

2026-01-10 08:41:50 CET

- reply

Is there something special about IBIT vs any other ETF (also non-bitcoin ones) or stock for this strategy to work?

2026-01-07 22:57:53 CET

- reply

GB10 looks interesting, but the price range is not that attractive. For a bit more sats you can get rtx pro 5000 blackwell GPU which eats GB10 this for breakfast, especially if you need a complete ML workflow.

2026-01-07 19:31:53 CET

Nice comparison https://www.phoronix.com/review/dell-pro-max-gb10-llama-cpp

2026-01-07 07:39:45 CET

- reply

I suspect small language models, SLMs, might be sufficient for this.

2026-01-07 07:19:28 CET

Христос се Роди! Срећан Божић!

2026-01-07 07:18:23 CET

- reply

Or you could just have a small team that complements each other instead of a large company, that is difficult to manage.

2026-01-07 06:53:54 CET

https://arxiv.org/abs/2512.23676 Interesting concept. What are the building blocks of needed to create nostr world models?

2026-01-05 11:53:44 CET

- reply

What i am doing currently with aggro is just feeding events directly to BadgerDB and then extract relevant events, as graph edges in parquet which i can load in python. I am testing some Bayesian generative modeling there. What do you mean exactly buy certify the graph extension?

2026-01-05 11:44:30 CET

- reply

I will have to check again then. I only saw 200€/month offers. Maybe I missed something.

2026-01-05 08:37:30 CET

- reply

I also have an option to host the relay on my NAS, lots of free space there but I am not sure if it is sufficient throuputh for your use case.

2026-01-05 08:36:05 CET

- reply

That is quite pricy. Would you like to share the costs? I need archive layer to extract network graphs and statistics. I tried your aggro code, and modified it to collect stuff directly in a BadgerDB to call it locally and extract data i need, but it is quite slow and seems unecessary to do it twice 😅

2026-01-05 07:35:08 CET

@nprofile…mzy2 Where are you hosting your archival relay and how much are you paying for it? Would you be interested in renting access to it?

2026-01-03 17:58:21 CET

- reply

It is all wirking now. Thanks!

2026-01-03 10:15:27 CET

- reply

Ah cool, I'll pass it to Amethyst devs once I fix zapping. It breaks from time to time without any way to debbug it 😑

2026-01-03 10:12:26 CET

- reply

You forgot 69 sats as default option.

2026-01-03 10:10:01 CET

- reply

Thanks santa for a fat zap, not sure what I did to deserve it 🫣

2026-01-03 10:08:01 CET

@nprofile…vavr I am getting some issues with zapping suddenly. Everything worked well yesterday and I changed nothing in between. I see this error message but pressing login again does nothing https://i.nostr.build/y0gfLuHCcAj4FvOj.jpg

2026-01-01 14:00:28 CET

- reply

Thanks, very nice.

2026-01-01 09:19:23 CET

Happy New Year!

2025-12-31 11:53:28 CET

- reply

Could you please clarify a bit what is the use case for vibium? Does it make sense to use it with vscode based coding agents, e.g. with cline or antigravity?

2025-12-21 18:03:30 CET

- reply

if you need something more private then njal.la

2025-12-21 17:54:41 CET

- reply

namecheap accepts bitcoin

2025-12-21 14:54:16 CET

- reply

That is the only reason why they are making one.

2025-12-20 19:49:24 CET

- reply

Ha, I have 500$ daily withdrawal limit, and I am not afraid to lower it more.

2025-12-20 11:34:48 CET

- reply

Yugoslavia was a competitor to EU, that is why it had to go. But that was not achieved via war per say, that was just a manifestation of failing economy caused by taking foreign depth, adopting western banking practices, and asking IMF for "help". That was the real psyop, which took place during Tito's time.

2025-12-20 10:47:17 CET

- reply

The best music and movies in any of the ex YU countries were made during the communist Yugoslavia. Any famous band in the region originates from that period.

2025-12-19 08:41:46 CET

- reply

If I had to choose, gemini 3 high did the best minimalistic design of the clock.

2025-12-19 08:40:47 CET

- reply

Here is my attempt with different models, although i do not know how to limit them to 2000 token responses. They all got time right, gemini flash did not animate seconds, so it is just a static clock. Only gpt-oss did a poor job. https://i.nostr.build/dasmbXrvZFdWgyV8.jpg

2025-12-18 07:25:29 CET

- reply

Will give it another try.

2025-12-18 07:21:32 CET

- reply

This might be useful if you have not tried it before https://github.com/steveyegge/beads

2025-12-18 07:19:08 CET

- reply

I don't like cli tools as I cannot easily check what they are doing, until everything is done. In cline/roocode I can do a quick revision of file changes, before saving it. Or revert code easilly few steps back, without git. It is slover than proper vibe coding, but it saves me time in other ways. I have not tried AGENTS.md yet but it is similar idea to clinerules

2025-12-18 07:05:38 CET

- reply

Ah I see, the question mark on the other page reveals the prompt 😅

2025-12-18 06:47:03 CET

- reply

Yea that one looks perfect. What do you get with gemini 3 pro on the same prompt? I know you don't like it, but for my use cases, e.g. porting code from pytorch to jax, or recreating a quick simulation from a paper in python, they both work well. When I get annoyed by one I use the other.

2025-12-17 22:25:49 CET

- reply

Hmm, the only one getting it consistently right for me is kimi k2 🤔

2025-12-14 09:40:08 CET

- reply

If there are multiple people in the meet, is one member acting as a server? Or is everyone connected to everyone?

2025-12-12 15:51:23 CET

- reply

Nice, I'll go back to my AI generated music now.

2025-12-12 14:49:09 CET

- reply

Would like to help, but I am only familiar with python libraries in those domains. That said big linear algebra accelerators, such as tensorflow and pytorch are production ready, so they have kotlin or java support.

2025-12-12 12:52:11 CET

- reply

So cool, looking forward to trying it out. Can one install steam proton and try out games from steam (or GOG)?

2025-12-10 20:54:57 CET

- reply

It sounds like a job for Hadoop HDFS. But i don't think it works well on Android.

2025-12-07 10:17:27 CET

An interesting parallel between an interogation protocol and an attention algorithms keeping us glued to our phones https://www.borislavpekic.com/2025/09/how-to-quiet-vampire-phone.html?m=1

2025-12-07 10:13:55 CET

- reply

A sure way to make friends in a gym.

2025-12-05 21:05:04 CET

- reply

Yes because the idea is to somehow exclude bots, or a single pearson spinning many npubs. But sure you can use it in many ways, also to downgrade people for some services.

2025-12-05 20:40:57 CET

- reply

It is a score, i am not sure that it is very social or that a high score improves your chances of getting a credit.
And you will be pleased to learn that vertex is not using mutes or reports to calculate the score.

2025-12-05 19:51:15 CET

- reply

Just call it attention score instead of trust score.

2025-12-01 10:40:53 CET

- reply

Something like this was probably an inspiration for "Children of Time" by Adrian Tchaikovsky

2025-11-27 22:06:54 CET

- reply

Emit white noise? 😅

2025-11-24 21:03:08 CET

- reply

You can look also into probabilistic approaches for steering LLMs in constrained domains, e.g https://arxiv.org/abs/2306.03081 and https://arxiv.org/abs/2504.07081 That might do what you need without having to retrain or fine tune a model. If you can run many smaller models in parallel it might be fast enough for daily use.

2025-11-24 19:32:06 CET

- reply

What is a Gopherbot?