N

-... .. - -.-. --- .. -. strong opinions, loosely held.

Public Key

npub1az9xj85cmxv8e9j9y80lvqp97crsqdu2fpu3srwthd99qfu9qsgstam8y8

NIP-05 Address

Profile Code

nprofile1qqsw3znfr6vdnxrujezjrhlkqqjlvpcqx79ys7gcph9mkjjsy7zsgygpzemhxue69uhhyetvv9ujuurjd9kkzmpwdejhgqg5waehxw309aex2mrp0yhxgctdw4eju6t0qy8hwumn8ghj7mn0wd68ytnddaksz9thwden5te0wfjkccte9ekk7um5wgh8qatzqyg8wumn8ghj7mn0wd68ytnhd9hx2kzs2a9

Publishing to

Author Public Key

npub1az9xj85cmxv8e9j9y80lvqp97crsqdu2fpu3srwthd99qfu9qsgstam8y8

Show more details

Last Notes

2026-07-29 23:00:33 UTC

https://blossom.primal.net/19e511a514fcbf10b77b54e0bb8c3bf6ae3c7c741c44b4336cd91318542a5b77.png

2026-07-29 12:08:17 UTC

- reply

but its so tasty, moar

2026-07-29 12:07:43 UTC

- reply

FAST Crucial X10 Pro

2026-07-28 23:56:13 UTC

https://blossom.primal.net/bfd01759f15714a7f23103e9e309e4c668fbd812052c6b672e2270faa5254212.jpg

2026-07-28 20:31:52 UTC

- reply

I hope you find a place to belong in all this ✌️

2026-07-28 16:52:02 UTC

- reply

https://x.com/nvk/status/2078116583424057675?s=20

2026-07-28 16:26:10 UTC

- reply

awesome

2026-07-28 14:44:50 UTC

- reply

https://x.com/nvk/status/2078116583424057675?s=46

2026-07-28 13:40:52 UTC

- reply

Knots is unsafe.

2026-07-28 13:27:16 UTC

- reply

That would require the COLDCARD to have a block data and/or have access to electrum servers. Would be cool, but its tricky and left to the wallet software.

2026-07-28 13:19:47 UTC

- reply

I think changing the contentious op-return rule was misguided, i don't believe filters work or support any group deciding what goes in or not.

2026-07-28 13:02:37 UTC

- reply

🙉 🙊 🙈

2026-07-28 13:02:14 UTC

- reply

I run 19, but for most ppl, i recommend 2 versions behind current. so today that would be 29.

2026-07-28 13:00:04 UTC

- reply

For most ppl, I recommend always running Core 2 versions behind.

2026-07-28 12:58:52 UTC

Cancel culture is not a winning strategy in bitcoin.

2026-07-28 12:54:12 UTC

- reply

Core 19

2026-07-28 12:00:04 UTC

- reply

👏 👏 👏
Its ok to change your mind.
#nevent1q…ktxa

2026-07-28 11:29:40 UTC

- reply

you may not see it, but this is your vibe on all your replies https://blossom.primal.net/d0b1691614ddec617a0491a0361c2c04f705c541cfbefdf43e60e1e70df0cced.png

2026-07-27 23:37:30 UTC

this is the best test.
"A business is simply an idea to make other people's lives better."
/Richard Branson

2026-07-27 22:49:13 UTC

Every time i make a bitcoin transaction with a COLDCARD Q using qr in and nfc out, i'm reminded how easy it is to use bitcoin in self custody.
Bonkers easy.

2026-07-27 20:08:17 UTC

- reply

i don't hate anyone, many things were said about me, dyor

2026-07-27 19:17:19 UTC

- reply

I have no issues discussing this with @nprofile…w2ua

2026-07-27 18:14:43 UTC

- reply

Jeff is not technically apt to understand this

2026-07-27 17:51:42 UTC

refining arca's messaging... does it make more sense? https://blossom.primal.net/36a4cca4a7483757423bfca3aea17e15e8b2614a57998751088795f7bf73828e.png

2026-07-27 17:22:35 UTC

- reply

He is free to make his clock, not to impersonate our brand. It's morally wrong to impersonate someone's work.

2026-07-27 17:18:47 UTC

- reply

first, you cut my tweet in half, not cool. But also, i address the nuance here https://x.com/nvk/status/2078116583424057675?s=20

2026-07-27 14:08:58 UTC

Good morning, if you are starting to notice you've been bamboozled and led into a minority fork, there is no shame in changing your mind.
It's your bitcoin on the line.

2026-07-24 12:16:13 UTC

- reply

that's silly fun, the actual integration is the coldcard code running on the arca, with the arca security boundary.

2026-07-24 12:03:34 UTC

- reply

ha!
#nevent1q…8ky4

2026-07-24 11:18:11 UTC

- reply

https://blossom.primal.net/c2db4c949b57ac979ba33523fb5d3a391254bd72af999960bc1269ad8449d094.png
#nevent1q…vlgy

2026-07-23 13:49:17 UTC

- reply

yes parts ordered, things moving forward

2026-07-23 13:19:40 UTC

- reply

there is no third party, this is self hosted

2026-07-23 13:08:49 UTC

dont lose your secrets.
https://blog.coinkite.com/arca-reservations-are-open/

2026-07-22 12:31:44 UTC

- reply

https://blossom.primal.net/93884d6e71ee0f20837af50c1da773ec5a8e5d4aa23cc45dc79c092835f3d911.png

2026-07-22 00:41:59 UTC

- reply

read the article

2026-07-21 16:09:06 UTC

https://blossom.primal.net/04aa8316d3ee3f8f655348aaf693aef038c5d194e5f595b79f8ca60c2d1bacef.png

2026-07-20 17:05:36 UTC

- reply

When everyone sees themselves as a victim, logic and reason dies.
#nevent1q…ml3r

2026-07-20 16:26:57 UTC

- reply

Talking about technical security is no longer allowed

2026-07-20 15:48:11 UTC

- reply

You should try reading it

2026-07-20 14:45:38 UTC

Please Government daddy, protect us from Open Source Ai Freedom and non-inflating computing price. https://blossom.primal.net/8f2306c6aaa305b58928b8ee4530c7d6b94f5a992b3e5c4c7dd0de1b5916349e.png

2026-07-20 14:05:43 UTC

- reply

correct

2026-07-20 13:52:28 UTC

- reply

I debated posting this. DIY Bitcoin security has become weirdly emotional.
A Raspberry Pi or ESP32 can be useful, but they are still general-purpose computers, not key vaults.
Offline is a condition, not a security architecture.
Thank your self-custody seriously.
#naddr1qv…dz8j

2026-07-20 12:48:29 UTC

- reply

There is no such thing as a clean phone.
#naddr1qv…j0wz

2026-07-19 11:18:35 UTC

- reply

oh boy

2026-07-09 22:38:22 UTC

llm-wiki v0.15.1 codex 0.144 compatibility fix https://blossom.primal.net/4f0f3b80924e2f55baf7fc0989d52fb12f79149210907e10c2d98dabcbc44fad.png

2026-07-09 11:26:51 UTC

- reply

the blogging is easy, what about the emailing? moved to beehiiv

2026-07-08 15:10:59 UTC

Substack suspended http://substack.coinkite.com
They have a bot subscription problem, we contacted their security team about 100K surge in subscribers all complaining to us.
Then instead of fixing their system, without notice or reply they suspend our company newsletter.
Whats your alternative recommendations, fuck SV companies.

2026-07-08 11:40:19 UTC

- reply

that was fast

2026-07-07 20:59:32 UTC

works great, will send a good message.

2026-07-07 13:10:05 UTC

arca todo list
https://blossom.primal.net/6d7ac2e9adafc0ddb0a150a1ba7a2000983b1a8e7f2958eff1577536025a8ac2.png

2026-07-06 16:27:56 UTC

llm-wiki.net v0.14.0 out
https://blossom.primal.net/31772011fc3dfb2a753d84bfe272b5c8c1464b3755fc539cfde72581b761cf27.png

2026-06-30 12:28:45 UTC

https://blossom.primal.net/5f6b579df0523e16408a0013bccccbdb8a9f5a199e9b87d583cbff5c6fa8cef9.png

2026-06-26 17:53:16 UTC

Received some messages asking if I take donations to llm-wiki.net work: NO
Consider donating to @nprofile…5h0z on its name instead 😃
OMG LLM_WIKI FOSS https://blossom.primal.net/0b57eca323f8af82656ddf1c122f9218744d097d36acf3672d0427f92900ff41.png

2026-06-26 10:49:43 UTC

- reply

it would be alot of fun.

2026-06-24 18:57:18 UTC

- reply

only robinhood can

2026-06-22 15:39:21 UTC

- reply

all good.

2026-06-22 12:09:02 UTC

- reply

Coinkite deletes your data and had no leaks.
Our competitors unfortunately don't care for your data as much, this is who lost your data. https://x.com/nvk/status/2031836293240668367?s=20
It's exhausting.
#nevent1q…q03f

2026-06-22 12:05:46 UTC

- reply

Rock64

2026-06-18 14:29:05 UTC

steel. https://blossom.primal.net/76dce7dba9ceea3cf3f31985c04bdf241403c87a8a36e25809bfc12a3d85e2e8.png

2026-06-18 13:56:59 UTC

Most people on both sides of the bip110 and Core sides are good ppl.
Bitcoin will force everyone to coexist or fork off.

2026-06-17 11:45:31 UTC

NVK LOVES MIT FOSS 😱
https://blossom.primal.net/e591ec66c7a77995b8a2f7472de0a918c785d118bd116f5e9738cd8501ba9757.png

2026-06-09 13:54:02 UTC

- reply

Let me ask agentnoise ;)
#nevent1q…hf3u

2026-06-09 13:31:45 UTC

- reply

1000x less bitching that's for sure.
#nevent1q…837e

2026-06-08 14:56:09 UTC

- reply

https://blossom.primal.net/2ed3a6d7af0bc575ddf73f1816a78a6e5f85f5cadd629b36b0c8b713cbf348e3.png
The reason i haven't built a great UI for this is because i want the web login quotas for frontier, not the api silly costs.
#nevent1q…v70y

2026-06-07 11:28:55 UTC

very health, much wow.
this was fun to build for myself.
absolutely no warranties, use health code at your own risk and responsibility.
llm-health.net
https://blossom.primal.net/6a14f5efa690a6e9e6de55598a737a119e491630e0a0ab24adcd517b6b7bcafe.jpg

2026-06-03 12:25:45 UTC

Miljan, Odell and the Primal cabal are actively censoring me and preventing my nostr project from taking off.
I'm moving to carvalo's protocol
https://blossom.primal.net/86e202f2eb49be6ed98fe6c8596eba2feb47b47d07635158e7390c256d6ddaf5.png

2026-06-03 00:34:57 UTC

Aside from dignity, what else is holding you back from sandboxing your coding angents?

2026-06-02 14:59:30 UTC

The most effective use of the philanthropy right now is funding token to world third world and consumer high end hardware to top engineers like antirez
Accelerate asymmetric output to those than can't access it.
Fund the top ppl making the local ai happen faster.
It's never been so affordable to give poor ppl an advantage.

2026-06-01 22:35:04 UTC

- reply

I know drama is fun to you, but you are poisoning this technology and hindering the other people working on it.

2026-06-01 22:16:25 UTC

The cool thing about OpenSats is that a peer group of YOUR category reviews your application.
YOUR PEERS review even before the board does.
There is no room for evil NVK to decided your fate alone.
🫂

2026-06-01 17:23:35 UTC

- reply

The world is far from perfect.
Be grateful some ppl are willing to do what you are not.
If you don't agree, simply start an alternative.
#nevent1q…s3q4

2026-06-01 17:22:19 UTC

- reply

I almost quit before, dealing with ungrateful cunts like these is a waste of productive time.
We have an independent board and independent committees.
We get zero dollars to ourselves, I donate my time to support FOSS.
If your project is not a match for OS, find other sources.
We are not obligated to fund you.
Yelling at us or telling us won't change that.
And yet, we process hundreds of grants, to things we judge a match.
Including tons of Nostr stuff.
#nevent1q…t8ku

2026-06-01 13:06:39 UTC

- reply

correct.
#nevent1q…9lp2

2026-05-31 19:10:13 UTC

Stop being Insufferable.

2026-05-31 19:05:10 UTC

- reply

Don't like it don't use it, build something better.
#nevent1q…ehfy

2026-05-31 19:00:41 UTC

You are not obligated to use primal, how about build some alternative and go get some users instead of bitching.
The principal reason Nostr adoption stopped is because of the bitching instead of building.

2026-04-29 13:03:05 UTC

If not enough bitcoin is held privately in the hands of unreasonable people we lose.
Bitcoin, in seeds, is the only insurance.

2026-04-18 01:21:58 UTC

- reply

Objectively

2026-04-14 12:38:08 UTC

- reply

ppl are using the slop cannon's wrong

2026-04-14 02:08:39 UTC

# Writing 84 Tests for a Project With Zero Lines of Code
The llm-wiki project has 3,610 lines across 22 files. Every single one is a markdown file. There is no Python. No JavaScript. No compiled binary. The "source code" is English prose — instructions that Claude reads and follows to build knowledge wikis from web research.
So how do you write tests for a program that is, technically, a document?
I figured it out. 84 structural assertions, 11 intentionally broken wiki fixtures, 5 behavioral evals via Promptfoo, and a GitHub Actions pipeline. As far as I can tell, nobody has used Promptfoo to test a Claude Code plugin before. Here is what I learned.
## The three-layer problem
Traditional testing has a simple contract: given input X, the function returns Y. If it doesn't, the test fails. But when your "function" is an LLM reading markdown instructions, the contract dissolves. The same instruction file, given the same user request, might produce different article titles, different file structures, different cross-references. The output is correct within a range, not at a point.
Anthropic, OpenAI, and GitLab all converge on the same solution: split tests into layers by how much uncertainty you're willing to tolerate.
**Layer 1 is deterministic and free.** No LLM calls. You're checking that the wiki's file system is internally consistent. Does every directory have an `_index.md`? Does every raw source have the six required frontmatter fields? Does the `type: articles` file actually live in `raw/articles/` and not `raw/papers/`? These checks take seconds and cost nothing. I have 84 of them. They run on every push.
**Layer 2 is semantic and costs money.** You ask Claude to do something — ingest a URL, compile an article, route a command — and then grade whether it followed the instructions. Promptfoo handles this with three assertion types: trajectory assertions ("did it call WebSearch?"), llm-rubric assertions ("does the output have complete frontmatter?" graded by a judge LLM), and custom JavaScript that checks the file system after the agent runs. Each eval costs about $0.50. I run five of them on PRs.
**Layer 3 is full workflows.** Research-to-article. Ingest-compile-lint. Retract-and-verify-cleanup. These use `claude -p` in headless mode, cost $10-20 per run, and execute weekly. I haven't built these yet. Layers 1 and 2 are live.
## The golden wiki
Every structural test needs something to test against. I built a golden wiki — a minimal but complete fixture with three raw sources, two compiled articles, proper cross-references, bidirectional See Also links, correct index files, and a valid log. Twenty files total. It passes every check.
Then I broke it eleven different ways. One copy per lint rule. `missing-index/` has a deleted `_index.md`. `bad-frontmatter/` has `type: invalid` instead of `type: articles`. `misplaced-file/` puts a concept article inside `wiki/references/`. `retracted-marker/` leaves a `` comment that should have been cleaned up. Each broken copy triggers exactly one violation. The test asserts that the defect is present — negative testing.
A shell script called `http://generate-defect-fixtures.sh` creates all eleven from the golden wiki in under a second. Change the golden fixture, regenerate, and every negative test updates automatically.
## Promptfoo on a Claude plugin
Promptfoo has a provider called `anthropic:claude-agent-sdk` that can load local plugins. Point it at your plugin directory, whitelist the tools, set a budget cap, enable sandbox mode, and it runs your plugin through test cases defined in YAML.
The part that surprised me: the `skill-used` assertion type. You can assert that the agent invoked a specific skill — not just that the output mentions wiki commands, but that Claude actually triggered the wiki skill at the Claude Code level. Combined with trajectory assertions that verify which tools were called, you can check both what happened and how.
I test five behaviors: the fuzzy router dispatching "Research the history of testing" to the research command, a URL to ingest, a question to query, an ambiguous single word triggering clarification (negative control), and the plugin loading without errors. Each runs three times to measure variance.
## What I actually learned
The biggest surprise: Layer 1 catches almost everything. The expensive behavioral evals in Layer 2 are for confidence, not coverage. Index corruption, frontmatter drift, misplaced files, broken cross-references — these are the actual failure modes of a wiki management system, and they're all deterministic. You don't need an LLM to verify that a file exists in the right directory.
Anthropic's eval guide says "grade outcomes, not trajectories." For wiki operations, the outcome IS the file system state. Check the files, check the indexes, check the links. If the structure is correct, the agent followed the protocol. The trajectory — which tool calls it made, in what order — is interesting but secondary.
The test suite is at https://github.com/nvk/llm-wiki in `tests/`. Clone it, run `./tests/test-structure.sh`, and watch 84 green checkmarks validate a project that contains zero lines of code.

2026-04-11 14:41:19 UTC

Writing bitcoinquantum.space with llm-wiki.net
In April 2026 I wanted to assess whether the quantum threat to Bitcoin was real. The honest answer lived across fifteen papers, a dozen Delving Bitcoin threads, twenty Bitcoin Optech newsletters, a running testnet, some Liquid transactions, and whatever Avihu Levy had pushed to GitHub that morning. The work was real and scattered. No article summarized it honestly. Headlines were downstream of press releases. The primary sources were where the actual answer lived.
This is one of the things llm-wiki was built for. I used it. Three weeks later I published [bitcoinquantum.space](bitcoinquantum.space) — three articles, ~15,000 words, 95+ sources cross-referenced, every claim verified. This is a writeup of how.
## The shape of the problem
Serious research has three failure modes:
1. **You can't find everything.** Sources scatter across formats and venues. You don't know what you're missing.
2. **You can't remember everything.** By paper #60 you've forgotten paper #4. You re-read. You contradict yourself.
3. **You can't update.** A new paper drops on publication day. Your conclusion is stale and your notes are already collapsed into prose you can't untangle.
Traditional knowledge management fixes (1) and partly (2). It fails at (3) because the maintenance burden compounds. @karpathy's framing, *"who does the maintenance?"*, is load-bearing because humans don't, not reliably, not for unsexy cross-reference updates nobody sees.
llm-wiki.net fixes (3) by making the entire artifact mechanically regeneratable from immutable raw sources. The only thing you maintain is the source pile.
## The pipeline, applied
**Raw sources, not notes.** Every paper, blog post, mailing list thread, and testnet report got dropped into `raw/` verbatim with a frontmatter header. No interpretation, no paraphrasing. If I don't have the primary source, I don't have it. `raw/` grew to 95+ entries.
**Compile, don't write.** `/wiki:compile` reads the raw pile and synthesizes cross-referenced wiki articles — one per concept, person, and proposal. "SHRINCS." "Taproot script-path post-quantum proof." "The BIP 86 problem." "Quantum Safe Bitcoin." Each article carries a confidence level, citations, and bidirectional cross-references. The wiki is Claude's work; the sources are mine.
**Query to find gaps.** Once compiled, I stop reading papers and start asking questions. *"What's the relationship between Ruffing's Taproot proof and BIP 86?"* The wiki answers with citations — and in the process surfaces the gap: 70-90% of BIP 86 outputs can't use the escape hatch. That's a thread I wouldn't have pulled linearly. Query mode is where llm-wiki stops being a filing cabinet and starts being a research partner.
**Output, last.** The articles on bitcoin

2026-04-10 21:44:28 UTC

- reply

Website with trade craft up https://bitcoinquantum.space
All done with LLM-wiki.net awesomeness
Their FUD will be destroyed.
#nevent1q…l772

2026-04-08 13:43:35 UTC

llm-wiki.net v0.0.15 is out
https://blossom.primal.net/c444d1f0d8d196bc8770715582b061e86649085819235f1675f0441f21420f5d.png

2026-04-08 13:32:37 UTC

Quantum FUD is only fought with the best slop. It's all about inverting Brandolini's law.
Part 2 https://x.com/nvk/status/2041644230620139787
https://blossom.primal.net/94be314aa71d7138a4243e24618b51abc435d5199392dd517f198f0704728535.jpg
Part 1 https://x.com/compose/articles/edit/2040989355330678784
https://blossom.primal.net/8ee3b2b141622255f3469d70036a8b0a3f49837a91af9103ec2ab7f1863d9e47.jpg

2026-03-23 15:40:46 UTC

- reply

RFC 6979 + deterministic builds is perfectly adequate. That's COLDCARD's and Cores solution

2026-03-08 16:09:08 UTC

https://blossom.primal.net/ee8ac4966e5c20d638ab04cf6d0ab4e3fc558d23b30254b0be06db87a3467475.jpg

2026-02-22 15:17:28 UTC

You are addicted to other people's problems
Go outside, walking grass and absorb some low entropy from the sun
https://blossom.primal.net/060b503dd99ba99dfdc28f12803ede2aff6ad72b73d28f72011e9218a234eafe.jpg

2026-02-17 15:44:48 UTC

GM
https://blossom.primal.net/3ff912768cb6b708add30d50354a68d82daf226a388b0acdcbf79727308216ff.jpg

2026-01-20 20:13:42 UTC

https://blossom.primal.net/21274149d27dc82bfde6e1f208d408de2cbc03a6286273d5a511af59e8e6624a.jpg

2026-01-20 02:41:21 UTC

https://blossom.primal.net/780fd80d4cbb43a173dfcf2cb7752e1286f13e69afa10f2817ff06d1f519f9d2.jpg

2026-01-17 15:17:46 UTC

https://blossom.primal.net/33060782abf8ebee2752c752269615c65ed6d5be971cb9fc4ac04be68517b90d.jpg

2026-01-11 18:38:21 UTC

Please. @nprofile…8ul5 can you fix it.
https://blossom.primal.net/c9992238eae144397fe8463ac48dabc1ffafc3a30c77556e44b4ac12a8f4d588.jpg

2026-01-11 17:53:42 UTC

https://blossom.primal.net/dacb03ec817f71e17a7ad1ac5cd2df1aafe36658f51f61f560e1c44e6b6bc1f7.jpg

2026-01-11 17:51:09 UTC

- reply

Many such cases

2026-01-11 17:24:02 UTC

GM is Nikita here?

2026-01-11 16:46:14 UTC

Free Persians breaks a lot of the previous world order.
Changes most of the Middle East.
Obama would not let them have it.
They are taking it.

2026-01-11 01:51:57 UTC

Persia.

2026-01-11 00:44:20 UTC

- reply

Love to see it, fuck the regime.
But this is a poster made from a lady at a parking lot in toronto.
#nevent1q…epdg

2026-01-07 04:45:45 UTC

- reply

🤝 Continuous Innovative Animals
#nevent1q…wdk4

2026-01-05 01:05:25 UTC

https://blossom.primal.net/0e6d57798d04723a168057e990bb31422fde04d57dec36ac7c0a3b73c334f48d.jpg