plantimals

so far, this is the closest project I've found. I love the concept of DVMs, but I'm not excited to use them at this layer. https://github.com/OpenAgentsInc/openagents
#nevent1q…5mzq

2025-12-27 17:28:44 CET

is anyone building a nostr-based agentic communication / orchestration app?

2025-12-24 23:46:44 CET

- reply

https://static1.srcdn.com/wordpress/wp-content/uploads/2023/09/gary-oldman-in-bram-stoker-s-dracula.jpeg

2025-12-24 17:39:37 CET

I don't enjoy having to say this, but @nprofile…ys9t your emails and notifications constantly jamming offers at me is not something sovereign people tolerate much of. your service and app have been incredibly useful. I'd really not like to have to block all notifications from your app or emails from you. but that is the next step if the panhandling doesn't stop.

2025-12-22 17:38:53 CET

this is solid advice for how to think about context engineering generally, and configuring your CLAUDE.md specifically:
https://www.humanlayer.dev/blog/writing-a-good-claude-md

2025-12-15 20:58:47 CET

- reply

reminds of the old joke:
what did Scotty find in the enterprise toilet?

2025-12-15 14:38:58 CET

- reply

#nostr is better with you. we all should be celebrating

2025-12-10 13:21:40 CET

- reply

translation note for Americans (I had to look it up):
Mayfair => Boardwalk (in the context of Monopoly)

2025-12-07 23:32:45 CET

- reply

nice vasoconstriction, like a cigarette after a meal. it counters the felt effects of blood going to the gut to help with digestion.

2025-11-24 22:42:43 CET

this a sign that you should dump the context and start on the feature from the specs up.
context engineering is real
https://youtu.be/IS_y40zY-hc
#nevent1q…r9gm

2025-11-17 15:17:42 CET

self-improvising memory. it rhymes with AI
https://www.mdpi.com/1099-4300/26/6/481

2025-11-17 00:10:56 CET

- reply

same. forced to reload if I want to read notifications. has been that way for at least a month

2025-11-12 03:52:31 CET

https://video.nostr.build/184cdb7aafbbff23eac02c14174a838aa5dc73a16052d064ce1543b5d12b2436.mp4

2025-11-10 14:22:27 CET

- reply

gm 🌄☕

2025-11-06 00:47:47 CET

context engineering seems like a valid concept. I've had good results following this approach https://github.com/humanlayer/advanced-context-engineering-for-coding-agents/blob/main/ace-fca.md

2025-11-05 16:15:58 CET

from Erik Hoel:
https://drss.io/DNAl

2025-10-30 06:58:31 CET

finally got my claude code set up the way I like it
https://www.youtube.com/watch?v=VddS5IWxHd8

2025-10-28 17:17:51 CET

- reply

you can listen by going upstream to the actual podcast link, or play the audio via the imeta tag attached to the article linked above on nostr clients that support it, like drss.io: https://drss.io/yuN3

2025-10-28 17:15:31 CET

- reply

great podcast episode from @nprofile…z04z
Luke Gromen
"Print the Money or Trigger the Revolution"
demonetization of farmland by #bitcoin
#naddr1qv…249z

2025-10-27 23:54:49 CET

https://www.youtube.com/watch?v=PorrJ0wupaY

2025-10-27 13:45:10 CET

most of the insights of The Wealth of Networks have since diffused into our awareness. it sees mostly positive feedbacks and not much negative, but still worth a look for those working on snd thinking about the protocol.
https://cyber.harvard.edu/wealth_of_networks/Main_Page

2025-10-25 20:08:47 CEST

testing drss link shortener: https://drss.io/w9p7

2025-10-17 22:47:14 CEST

- reply

"The Peregrine" by J.A. Baker

2025-10-14 16:50:16 CEST

a funny aspect of getting the coding agent to write meaningful tests around #nostr functionality is the understanding that secret key -> pubkey cannot be faked. I see so many "npub123-abc" in its tests. no matter how often I explicitly instruct it otherwise. I have a rule in CLAUDE.md which points out that we have a keys.md with a markdown table with 10 secret keys, pubkeys, and npubs for use in tests, and it always forgets and just makes things up.

2025-10-12 23:10:22 CEST

- reply

a library with real books is a good metaphor for running a node. no one between you and the contents, no snoops.

2025-10-08 21:34:27 CEST

- reply

welcome!
fix the money, fix the internet.

2025-10-04 16:38:13 CEST

- reply

attention is your bandwidth for changing the universe

2025-09-24 14:05:47 CEST

- reply

snowclone overreach

2025-09-24 00:47:25 CEST

twitter is the uber of microblogging

2025-09-22 17:29:00 CEST

- reply

it's beautiful. also this is when the length of day is changing the fastest, so the angle of the sun and the colors it makes are changing the fastest as well.

2025-09-12 23:55:10 CEST

I keep coming back to making lists of lists, or more accurately, set of sets. is there a reason why it doesn't exist in the NIPS? just hasn't been a need so far, or was it discussed and discarded? I ask because I've looked and did not find anything.
thinking of using 30100 for that kind. I like the idea of a hierarchical list of lists than can link in lists owned and maintained by others. if you hang follow sets from those lists, it becomes a nice way to subscribe to a feed, if there are clients that allow it.

2025-09-09 00:01:32 CEST

- reply

this is because long form content is currently an afterthought for most clients. no blame, most don't care. but we definitely need better tools.

2025-09-07 22:07:38 CEST

https://youtu.be/iqomTAiRnVM

2025-09-07 21:16:13 CEST

- reply

so beautiful. ordovician?

2025-09-07 19:46:30 CEST

this is a common issue I think, but specifically with claude code.. I'm so exhausted by having to explain when I'm asking a question about a proposed change I am never asserting something... always just asking.
when i ask something like "why are you using Getenv() here?" I usualy get:
"You're absolutely right! I was so wrong, I'll stop that and go this other way..."
if I wanted to express that sentiment I'd do it directly. has anyone found a way to phrase a question economically that conveys that you actually just want an answer?

2025-09-04 16:48:02 CEST

This Week in Bitcoin from @nprofile…vymj
https://npub.blog/a/naddr1qvzqqqr4gupzp4hx023yj4tkmanpmnhf50qu4vl78ukwze895w9c7ckxj9vfauz8qqgrwv36942xsefd2dcxzmfd2ashy6896ln/
npub.blog translates articles with imeta tags containing 'audio/mpeg' mimetypes into podcast episodes

2025-08-31 22:49:54 CEST

- reply

they are the 1-900 numbers of our times

2025-08-31 22:41:44 CEST

- reply

the worst part about it is you know the model (assuming claude...) does not understand. worse still are apologies. if your short term memory is wiped every few minutes, an apology is pretty pointless and hollow.

2025-08-31 22:38:04 CEST

- reply

I enjoyed this alan watts podcast:
https://npub.blog/a/naddr1qvzqqqr4gupzpajgxp72cwmk78n4ydftvc65n90rspve7xm5j74yzu44mwwdw2z4qqs52upw95envt0zszfj65m9v45kueedw358ymm4va5z64rgv5k5uet5d0y6ug/

2025-08-30 21:36:36 CEST

- reply

We shall see - Jerod McDaniel is not one to play by ordinary rules

2025-08-30 19:23:52 CEST

- reply

the word "absolutely"

2025-08-30 17:53:13 CEST

- reply

nvm about lightning, must have been an issue on my side

2025-08-30 17:52:15 CEST

- reply

I see that, and believe you are correct. but is there a way to modularize that so I can use a hotkey to dump the preamble and then just type the question?
also, I would like to zap you for your help. you should add a lightning address to your profile

2025-08-30 17:35:24 CEST

I am using claude-code, and one constant struggle I have is getting it answer questions without poisoning it's concept of what we're working on. I create markdown files describing the features and such, but when I come across strange implementation and ask something like:
"is there an implementation of this in go-nostr we could use instead of writing new code?"
it takes that as a suggestion "You're absolutely right!" and runs off to edit the code, rather than just answering my question. I have various lines in my CLAUDE.md describing how not to run off and change things without being explicitly told to do so, but that all eventually falls out of context or is just not working. do any of you have suggestions on how to ask claude code questions in context without derailing things?
#ai

2025-08-29 14:09:29 CEST

- reply

npub1sm8nwsjr4hev29ntywt7yv9d05ws3tvwxtqqtckx86df5euw8t5q0p0zr9

2025-08-26 12:54:50 CEST

- reply

hey there, it's been a while. how's it going?

2025-08-26 06:37:30 CEST

- reply

$135,000

2025-08-25 02:37:58 CEST

- reply

"You're absolutely right!"

2025-08-24 21:48:08 CEST

- reply

there's an industry out there linking identities yo emails, and it's not optional.
#naddr1qq…dvnz

2025-08-24 19:46:36 CEST

https://www.youtube.com/watch?v=nL0Qagp5ku4

2025-08-24 17:20:48 CEST

- reply

this is the way

2025-08-24 14:18:25 CEST

- reply

exactly. if the climate were ever not changing it would be inert and dead.

2025-08-23 20:51:18 CEST

- reply

nostr integration
I feel like people are talking about it, and I'm trapped in the lobby, missing it. I'm using bitchat downloaded from the zapstore, on version 1.1.0. I thought the idea was that we could also use our npubs to sign messages and have some continuity of identity over this mesh.

2025-08-19 17:00:37 CEST

- reply

paging @nprofile…tuzg

2025-08-17 21:25:37 CEST

- reply

paper relays

2025-08-17 21:02:58 CEST

- reply

it's time for this. cash on the barrelhead for digital good
#nevent1q…vc0z

2025-08-13 15:38:23 CEST

- reply

WoT weaves into those two as well

2025-08-13 14:29:08 CEST

- reply

owning bricks and mortar books is like running your own culture node

2025-08-13 04:31:19 CEST

- reply

we are still so early

2025-08-11 02:11:18 CEST

"that was the time I was most frightened, waiting for my turn.
I'll never put on a lifejacket again"

2025-08-09 18:54:38 CEST

- reply

LFG

2025-08-09 18:52:58 CEST

- reply

nostr only!

2025-08-09 17:35:13 CEST

excellent paper that zooms out to question the current AI paradigm of stochastic gradient descent. it results in "fractured, entangled representations", that is, deep neural network toplogies which are not modular. this means that in order to improve a specific function, all the scattered heterogeneous loci of that function must be found and updated. the author points out an example of neural nets, though vastly smaller than modern LLMs, that were generated by a process of open ended evolution, and that they produce highly factored and modular representations.
"Questioning Representational Optimism in Deep Learning: The Fractured Entangled Representation Hypothesis" by Kenneth Stanley et al
https://arxiv.org/abs/2505.11581

2025-08-09 16:00:52 CEST

this book is still relevant, maybe more than ever
from Richard Hamming's "The Art of Doing Science and Engineering":
FORTRAN, meaning FORmula TRANslation, was proposed by Backus and friends, and again was opposed by almost all programmers. First, it was said it could not be done. Second. if it could be done, it would be too wasteful of machine time and capacity. Third, even if it did work, no respectable programmer would use it—it was only for sissies!
The use of FORTRAN, like the earlier symbolic programming, was very slow to be taken up by the professionals. And this is typical of almost all professional groups. Doctors clearly do not follow the advice they give to others, and they also have a high proportion of drug addicts. Lawyers often do not leave decent wills when they die. Almost all professionals are slow to use their own expertise for their own work. The situation is nicely summarized by the old saying, “The shoe maker’s children go without shoes”. Consider how in the future, when you are a great expert, you will avoid this typical error!
With FORTRAN available and running, I told my programmer to do the next problem in FORTRAN, get her errors out of it, let me test it to see it was doing the right problem, and then she could, if she wished, rewrite the inner loop in machine language to speed things up and save machine time. As a result we were able, with about the same amount of effort on our part, to produce almost 10 times as much as the others were doing. But to them programming in FORTRAN was not for real programmers!

2025-08-08 02:22:29 CEST

- reply

not to mention torture of ocd zero inboxers by constantly resetting it to show a nagging unread notification every day

2025-08-07 13:56:12 CEST

- reply

It should be plural unless it's very important to emphasize that it is you and never will be anyone else. you are speaking for the company, which could include others.

2025-08-07 03:06:35 CEST

@nprofile…edt4
Are you around?

2025-08-06 23:39:20 CEST

- reply

@nprofile…vymj did you ever create that npub?
#nevent1q…euw8

2025-08-06 23:34:09 CEST

- reply

@nprofile…vymj

2025-08-06 23:26:39 CEST

https://npub.blog is a webapp for reading long-form content. it pulls articles from anyone you follow, as a feed. I haven't found another site that does such a thing so I built this. I've been polishing it, but it's obviously still very raw and untested. give it a look. send me feedback if you notice something broken.
you can enter an npub or a nip05 if you want to see someone else's article feed, or just sign in with your own to read your feed.
one day nostr will transparently supplant RSS feeds as the obvious way to asynchronously distribute and track long-form content. there's just some missing components along the way we have to build.

2025-08-04 14:22:42 CEST

- reply

this is the answer. the nips are axioms, your apps are theorums. prove some useful stuff.

2025-08-03 17:17:07 CEST

- reply

chantrells for the win

2025-08-02 20:43:11 CEST

- reply

I want the privacy, autonomy, offline capability, etc.
but it may well be cheaper to run something locally than pay $200 a month or more for something they can rug you on unilaterally (thinking about the weekly caps, etc)
I think the small but smart concept is true enough. I see more capacity and quality coming to local models, especially as the bench of open source coding models gets deeper and better.

2025-08-02 20:36:38 CEST

- reply

I am able to run models locally on my rtx3060, but that is a paltry 12gb capacity. the models I can run on that are not worth coding with.

2025-08-02 20:33:35 CEST

- reply

ah no, it was me being impatient. your profile hadn't loaded in amethyst, so no lnurl. got it sorted now

2025-08-02 20:31:25 CEST

- reply

this is a very helpful answer, thankyou. wish I could zap you!

2025-08-02 20:02:09 CEST

- reply

I will #zap good answers

2025-08-02 20:01:34 CEST

what is the most cost effective way to run a #LocalLLM coding model? I'd like as much capacity as possible, for instance to run something like qwen3-coder, kimi-k2, magistral, etc in their highest fidelity instantiations.
I see three high level paths. buy an..
- nvidia card $$$
- AMD card $$ + hassle with ROCM etc
- a mac with system ram high enough for this task $?$?
- something else?
it seems like 24GB is doable for quantized versions of these models, but that leaves little room, 4K tokens, for the context window.
#asknostr #ai #llm

2025-07-29 23:22:19 CEST

- reply

yes. I tried to order something online, and the order was just silently cancelled. I called in and was told that my email wasn't associated eith my name. wtf does that even mean, I ask. no satisfactory answer is given and I am left in the lurch.

2025-07-26 16:52:27 CEST

- reply

call it SMS, a digital postcard

2025-07-24 22:07:31 CEST

- reply

autonomous drug smuggling submarines running in circles
#nevent1q…9n3x

2025-07-19 19:05:55 CEST

- reply

they might be vents, to allow air to escape the mold while it's filling, so there aren't voids among the tread detail

2025-07-19 18:40:53 CEST

- reply

librewolf

2025-07-09 15:07:25 CEST

- reply

gm ☕ 🌞

2025-06-30 00:31:58 CEST

- reply

it's great to be here with you