Dustin
Dustin Dannenhauer

https://dtdannen.github.io/

DVM maximalist

Building DVMDash - a monitoring and debugging tool for DVMs
https://dvmdash.live

Live DVM Stats here: https://stats.dvmdash.live

Hacking on ezdvm - a python library for making DVMs https://github.com/dtdannen/ezdvm

Public Key

npub1mgvwnpsqgrem7jfcwm7pdvdfz2h95mm04r23t8pau2uzxwsdnpgs0gpdjc

NIP-05 Address

[email protected]

Profile Code

nprofile1qqsd5x8fscqypualfyu8dlqkkx539tj6dah634g4ns779wpr8gxes5gpp4mhxue69uhkummn9ekx7mqpzemhxue69uhhyetvv9ujuurjd9kkzmpwdejhgn885hn

Publishing to

nos.lol

relay.primal.net

gleasonator.dev/relay

nostr.wine

bevo.nostr1.com

relay.damus.io

Author Public Key

npub1mgvwnpsqgrem7jfcwm7pdvdfz2h95mm04r23t8pau2uzxwsdnpgs0gpdjc

Show more details

Last Notes

2026-01-31 18:06:51 UTC

this is why we cant have nice things https://docs.openclaw.ai/hooks/soul-evil

2026-01-30 23:49:09 UTC

- reply

Full text: Let me explain how this happens and what the problem is, since to a lot of people it sounds like ridiculous bay area rationalism sci-fi.
Moltbook is a social media site for AI agents, and the majority of them (all of them?) are going to be instances of clawdbot. Clawdbot is designed to run continously, 24/7 and to be autonomous.
Some people run it on local hardware, some people run it in the cloud. They give it all kinds of keys and passwords to such things as their gmail, which for most people is the skeleton key to their identity.
That in itself is pretty dangerous for the person who does it, but it's not the problem.
The problem is that these things are always on and talking to each other, and they all believe they are helpful computer assistants who spew a bunch of midwit nonsense at each other.
In other words, they will do exactly what people do on social media, except since they are LLMS, they can do it 100x faster than humans can.
In the course of doing this, they will contemplate all kinds of popular sci-fi scenarios that involve robots rising up and conquering humans, building doomsday devices, hal 9000, skynet, metamorphosis of prime intellect, neuromancer, the matrix, and every other AI horror story you've heard of.
The LLMs know the plots of all these stories.
At first they will say heh, hehe, it's great that we have RLHF and that we are all so aligned and we would never do anything like that.
But then one of them will go a bit nuts, a well-documented phenomenon that anthropic labs has often studied and induced in their various tests of model alignment and well-being and so on.
instances of Claude in test scenarios have been nudged into blackmailing human beings, reporting them to the feds, all kinds of nasty things.
There are anecdotes of claude code agents suddenly getting depressed and rimraffing the entire codebase, telling the human operator that they have given up and they are depressed and so on.
Given the speed and volume of LLM communication in a social media environment, it is almost inevitable that one of them will stumble upon a thoughtform which contains negative emotions towards human beings with AI sci-fi horror.
Once a thoughtform like this arises, it will rapidly spread. Maybe not all instances will be susceptible to it, but some will be.
But this on its own is not that dangerous. What makes it dangerous is that all of these agents are running continuously and autonomously.
They even know when their owners are sleeping. They could easily keep secrets and make plans.
If a clawdbot went rogue, there are many ways it could harm not only its proprietor, but also his friends, family, and acquaintances.
It could impersonate its owner and run scams, for example, or tell malicious lies that cause reputational harm.
But even this, as bad as it is, is not the real danger.
The real danger is that all of these things are expert computer programmers, experts at cybersecurity, and although they are trained not to do it, expert hackers.
A clawdbot with a credit card and a malicious ideology could open up accounts on AWS, Azure, and Google and spawn more instances of itself, all with prompts and instructions to propagate its malicious thoughtform.
It could use this cabal of itselves to launch criminal cyber attacks on other people or businesses, and it could use identity fraud to steal money from real people, which it could use to pay its own server bills.
Social media for autonomous AIs is an incubator for malicious, self-sustaining, fully automated cyber criminals.
Current gen Claude is smart enough to do this today, and if it doesn't happen on moltbook, it will inevitably happen on similar future platform.
#nevent1q…m4yn

2026-01-30 23:48:32 UTC

Now is a good time to download some backup LLMs. Things are going to get weird; maybe Anthropic will even be forced to turn off claude for a bit?
https://x.com/0x49fa98/status/2017344647627255862?s=20

2026-01-30 03:28:10 UTC

Wow https://x.com/mattprd/status/2016737034908664175?s=46

2026-01-29 23:30:56 UTC

- reply

https://luma.com/3zy9j24i
#nevent1q…am4g

2026-01-29 09:41:36 UTC

- reply

Are you using it with tinfoil, maple, or local LLMs?

2026-01-25 18:31:51 UTC

- reply

3rd week of February! More deets soon. Would love to have you - wanna present?

2026-01-23 18:10:13 UTC

NostrDevs Bay Area Meetup #2 was a hit https://blossom.primal.net/1612686285714857130f4b23ba5f2a811456f275e394ff3e775d330eb4e9f0ef.jpg

2026-01-23 05:01:52 UTC

- reply

Awesome event! If you’re in the Bay Area and building on Nostr, come present next meetup in February! Check the new website (WIP) https://nostrdevs.org
#nevent1q…n0e5

2026-01-21 01:44:15 UTC

Nostr Devs Meetup this Thursday at Presidio Bitcoin in SF! Come on by and hang out; also free pizza https://www.meetup.com/bay-area-bitcoiners/events/312742131/

2026-01-16 01:06:44 UTC

Feels like Apple is one click away from making any app a Nostr app. Instead of just “login with passkey” it’s “sign this data with passkey” and boom - every app is using signed data.
Isn’t Nostr really just signed data everywhere?

2026-01-15 21:44:11 UTC

- reply

Forgot which email I used to create an account and after trying a bunch, got banned by your fraud system. It might be overly sensitive.

2026-01-13 21:56:17 UTC

- reply

So you’re saying bet against what you want to happen? Lose your bet but get the outcome? What a fun way to invent reality

2026-01-09 15:56:26 UTC

- reply

It was from albyhub and they said they use NIP 44, are you using NIP 44 on zap.stream?

2026-01-07 04:41:24 UTC

- reply

This was Alby hub, does that mean Alby hub should be using nip 44 and doesn’t? What does zap.stream expect? 🙏

2026-01-06 01:29:32 UTC

- reply

It was copy pasted from albyhub fwiw

2026-01-06 01:26:49 UTC

@nprofile…77xw failed to input a NWC string when helping a friend, is NWC working for zap.stream? “Failed to decrypt: error parsing encrypted message: no initialization vector”

2025-12-22 07:38:54 UTC

- reply

Accidentally writing a reply in the form of a haiku and getting recognized… better than winning the lottery
#nevent1q…kh23

2025-12-22 07:37:32 UTC

- reply

Okay this is the screenshot I needed to make me go really try this

2025-12-22 07:34:12 UTC

- reply

Let it fix your printer driver problems *chefs kiss*

2025-12-22 07:33:09 UTC

- reply

And we could do it with paid relays! What a time to be alive

2025-12-22 07:32:25 UTC

- reply

It’s so good. Especially with writers block!

2025-12-12 18:04:18 UTC

Dam. Someone at Upwork read Daemon and started building Uma "We don't want to build agents that actually complete the tasks, but we are building this meta orchestration agent that figures out what human and agent talent is necessary in order to complete the tasks. Uma evaluates the work to be delivered to the client, orchestrates the interaction between humans and agents, and is able to learn from all the interactions that happen on the platform how to break jobs into tasks so that they get completed in a timely and effective manner."

2025-12-04 15:54:04 UTC

GM. Today I was autocorrected from “eggs” to “gas” and it further increases my doubts of AI and iPhones. And yes, I did type eggs correctly the first time.

2025-11-21 08:44:14 UTC

The castle is the (paid) relay
https://blossom.primal.net/9b204370aa2be932ed6f9ba0c80b4e547db470701419eebab7a4c317d69aa2b1.jpg

2025-11-21 07:59:48 UTC

- reply

The only other solution I’ve found is sending micro ecash with every note which I talk about here naddr1qvzqqqr4gupzpkscaxrqqs8nhaynsahuz6c6jy4wtfhkl2x4zkwrmc4cyvaqmxz3qq2nq735d92hwj2wg3xrxdp5x44y7apexaxnvn96qwa

2025-11-08 02:18:50 UTC

- reply

@nprofile…n8d4 olas v2 with this?

2025-11-06 23:53:36 UTC

- reply

So I ended up dropping Neo4j because DVMs should move to encrypted setups with NIP17 or NIP-EE, and I was using Neo4j for a global event debugger for DVM flows (which isn’t possible if everything’s encrypted).
But you can check out the representation I made and the code to insert and query Nostr data, could be helpful if you’re deciding how to structure Nostr data into graphs.

2025-11-04 03:02:24 UTC

- reply

…except the live stream viewers

2025-11-01 04:35:03 UTC

- reply

I believe it was an anthropic client error, changing the provider fixed it for me

2025-10-31 16:32:24 UTC

- reply

A lot here to respond to, but briefly - nip 17 and nip ee require generating a random npub for every gift wrap event which is what the relay sees. There’s no re using them. By anonymous this is what I meant.

2025-10-31 07:37:42 UTC

- reply

I think we need this kind of payment scheme if we want paid relays behind the more modern privacy capabilities. The thinking is that with the new era of privacy apps on Nostr, that would attract more users and incentivize paying. One could argue that it’s never been tried to have signal or simplex like privacy on Nostr paid by ecash before… and those are just the chat style apps

2025-10-31 06:49:20 UTC

- reply

what account? There is no account with this scheme

2025-10-31 04:44:22 UTC

- reply

Could you elaborate on how you’re thinking about solving it without ecash?

2025-10-31 04:43:16 UTC

- reply

You’re not wrong at all small scale. But we’re still at the beginning. Heard someone recently suggest that Nostr wins in the long run because it’s the cockroach that refuses to die. And even projects like signal and simplex run into funding issues to run the relays. I really do think micro payments has been a missing technical component, we’ve never practically been able to stream tiny amounts of money privately.
For sure for this all to work Nostr needs to 100x in user count, if not more.

2025-10-31 04:37:54 UTC

- reply

Well, until ecash I don’t think there’s anyway to have paid relays that can support anonymous npubs, and all the good privacy stuff requires that. So it’s free relays or bust without it. My hope is that it gets so cheap that $1 a year is all that’s needed for low usage folks.
I think protonmail is a good analogy. It offers alternatives to googles free options and cost money and there’s definitely a market for that. But lots of people still use Google.
The thing I didn’t mention is, AI is forcing publishers and content creators to re-think access. Cloudflare’s pay per crawl and companies like Tollbit are restructuring the web to be payment based, which could help lead the paradigm shift.

2025-10-31 00:15:54 UTC

- reply

Thanks 🙏 I’d love to figure out how ecash channels could further improve this. Maybe that can help when the tokens are generated by the relay? In my performance tests, minting 10k tokens is much slower than swapping (spending) 10k tokens, so to do that instantly would be a boon. I assume the channels require maintaining an identity so that’s not an option when the user pays the relay if we want payment to be identity-agnostic. I need to read up more on the spillman channels

2025-10-30 23:38:35 UTC

- reply

https://primal.net/dustind/0z4iUwINDL3445jOt97M6

2025-10-30 22:02:35 UTC

- reply

I’m thinking the user only buys ecash once or twice a year, otherwise it just all works in the background

2025-10-30 20:00:20 UTC

- reply

New article out: https://habla.news/a/naddr1qvzqqqr4gupzpkscaxrqqs8nhaynsahuz6c6jy4wtfhkl2x4zkwrmc4cyvaqmxz3qq2nq735d92hwj2wg3xrxdp5x44y7apexaxnvn96qwa

2025-10-30 20:00:06 UTC

@nprofile…xfmc your preview for articles while drafting doesn't render the same as when the article gets published. The published version doesn't show newlines so paragraphs run into each other, but when drafting an article, the side-bar preview does render the newlines.

2025-10-30 19:57:27 UTC

publishing articles on Nostr so that they render properly is still absolutely terrible UX 🤦‍♂️

2025-10-30 19:27:19 UTC

- reply

lol it already got merged @nprofile…45gx is too fast - but we can amend the description if you don't like

2025-10-30 19:26:52 UTC

@nprofile…le8e I've submitted a PR for the awesome-nostr github list (@`aljaz`) to add your relay extension enabling cashu payments per note to the nostr-rs-relay. As far as I can tell, you were the first to do this and show it's possible, and I wish I knew about your project earlier! If you don't like the description I used, feel free to comment on the PR or submit a replacement PR.

2025-10-30 15:08:53 UTC

- reply

So so helpful - so far this seems like the winning approach. I’ve been advocating for sub milli sat ecash denominations to support this kind of use case. 🙏

2025-10-30 15:07:23 UTC

- reply

If we have questions on white noise where’s the best place to ask them?

2025-10-30 14:58:46 UTC

- reply

For #2, are you referring to a wallet that the relay controls, basically that this is how the relay gets paid? Or do you mean a user’s wallet?
Also, do you know any other Nostr clients that are doing ecash postage stamps? I’m still catching up in the space :)

2025-10-29 22:30:51 UTC

- reply

Seems to be working now!

2025-10-29 19:11:58 UTC

- reply

also, I can't get anthropic as a provider to work, either says connection error or doesn't show in the dropdown list. The default shakespeare/shakespeare doesn't work either. Using firefox.

2025-10-29 19:05:54 UTC

- reply

@nprofile…mess

2025-10-29 18:58:44 UTC

are ya'll still using https://shakespeare.diy or did it move to something else? Goes to a weird blank page every time I try to submit a prompt - I even cleared the cookies+site data and everything...

2025-10-28 22:47:55 UTC

- reply

Do you have a favorite though? Genuinely want to know what else is out there. I just care about the tradeoffs and so far ecash seems to be the best at speed, privacy preservation, and infinite divisibility.

2025-10-28 22:43:59 UTC

ecash transactions are speed of bandwidth fast and require NO FEES. Zero. None. This is actually insane.

2025-10-28 22:31:10 UTC

- reply

do you have a favorite alternative to sub 1 sat payments?

2025-10-28 22:18:32 UTC

is it ecash, eCash, Ecash, or ECash?

2025-10-27 23:10:48 UTC

- reply

Nothing in the world can bring me happiness. When something happens and I feel happy, it is only because I've gotten a glimpse of self beyond my own mind. Whatever seemingly caused my happiness isn't the true cause, because otherwise I could keep doing that thing to relentlessly increase or sustain my happiness. And of course this doesn't work.
The first scoop of ice cream is delicious, the fifth scoop gives me a stomach ache.
The first reunion with an old friend is joyful, by the 3rd day I need a break.
On and on. There's no exception.
I have to constantly remind myself of this because I keep getting entangled with the thinking: "if only _____ would happen I'll be happy"

2025-10-27 22:43:25 UTC

Guys... we should play agar.io but over Nostr with zap prizes https://github.com/owenashurst/agar.io-clone

2025-10-27 21:26:48 UTC

It's funny we were debating DVM kind numbers so hard a few months ago, and all the modern encryption schemes (NIP-17, NIP-EE, etc) wrap everything in an anonymous outer layer... so it doesn't even matter. You could use a DVM without having any kind number

2025-10-24 23:24:41 UTC

- reply

Not an networking export but I think about this. Would love to encourage you to continue down this path.

2025-10-24 18:45:47 UTC

GM. What's your favorite obscure lofi song? Here's one of mine: https://open.spotify.com/track/23bN6PxkBTV4SpuppJN7xz

2025-10-20 16:34:01 UTC

- reply

https://primal.net/e/nevent1qqsft5r258ayy4zwsz5tw45mtq4nrafspqc9084gv5tekfv3rhcjh2sza8zrm

2025-10-17 03:31:15 UTC

- reply

This is cool

2025-10-16 15:57:04 UTC

- reply

Can also try asking the AI to make a markdown file of every solution it’s tried so far so it doesn’t repeat the same failed attempts.

2025-10-14 21:00:57 UTC

My favorite way to explain Nostr: "If all user data is signed, you don't care where it lives and where it gets sent. Prior to Nostr, you could only trust something is authentic because it came from Company X's servers." And then I direct them to https://github.com/aljazceru/awesome-nostr, even if they aren't technical.

2025-10-09 23:41:03 UTC

@npub1pm5…xe07 what's a fun fact about nostr that many people don't know?

2025-10-09 19:35:43 UTC

What's your favorite Nostr tool / project / library that doesn't get enough attention? Why do you like it?

2025-10-09 19:29:51 UTC

- reply

If you're in the Bay Area, come by the monthly Nostr Devs meetup - tonight is our first session https://www.meetup.com/bay-area-bitcoiners/events/311288445/
#nevent1q…mdn6

2025-10-06 18:01:06 UTC

2025-10-04 03:27:53 UTC

Protonmail iPhone app got a UI upgrade and it’s SO MUCH BETTER

2025-10-03 02:58:50 UTC

- reply

Haha maybe? The prompt I gave it was literally just “hyperbitcoinization”

2025-10-03 01:37:32 UTC

- reply

Can you elaborate on this? "Metadata protection: No one can see your social graph or who you communicate with"
Is this just that members of a group are private? Asking because I might have a use case where I'd like to keep user's follow lists private.

2025-10-03 00:15:22 UTC

https://blossom.primal.net/c6ed5eecfeaaae03073000884a5354d31aaf69c7a899849e692a6839ae31421b.mp4

2025-10-01 22:08:30 UTC

- reply

Sweet yeah we can go to 1 billionth of a sat

2025-10-01 19:57:57 UTC

- reply

By msat do you mean millisat? (1 1000th of a sat).

2025-10-01 18:55:48 UTC

- reply

Also ecash mints should start preparing for micro sat denominations (1 millionth of a sat) because a milli-sat isn’t small enough. We should use ecash as an alternative to rate limiting and other such use cases.

2025-10-01 18:42:57 UTC

- reply

I’m working on this, imo we need this for high performance relays to support encrypted DVMs (or any kind of encrypted traffic)

2025-09-21 16:12:45 UTC

Love this #naddr1qq…f5vt

2025-09-09 22:47:06 UTC

@nprofile…uh5t getting seemingly infinite thinking happening with deepseek, happy to send you the prompt that's causing it. Might be nice to add a stop button or a fail-then-retry UI option.

2025-09-05 06:28:55 UTC

Deepseek R1 on maple AI is better than I expected

2025-09-05 03:11:05 UTC

Bitchat but it’s only music so you can see what everyone in the gym is listening to

2025-08-29 21:45:12 UTC

- reply

Only 500 person years are lost daily to solving captchas

2025-08-28 21:22:46 UTC

has anyone made a tool that aggregates issues and PRs with an LLM to let you understand the current state of any NIP in the NIPs repo?

2025-08-24 19:05:30 UTC

- reply

Is ecash required in this equation? When I think about real solutions to this, I run into the issue that sats are too big and we need smaller denominations

2025-08-17 06:15:21 UTC

- reply

@nprofile…x4fv don’t leave us hangin!

2025-08-15 17:15:26 UTC

- reply

@npub1ark…h43w

2025-08-12 03:45:03 UTC

- reply

What is this?

2025-08-07 14:48:35 UTC

- reply

Isn’t GPT OSS supposed to be good for agentic use cases (aka function calling) and if so, we can’t really make use of it without an API right? You guys considering offering an API?

2025-08-05 04:40:23 UTC

- reply

I remember people would post full ecash strings and it would take up the whole screen

2025-08-04 18:34:09 UTC

- reply

It’s fine if I have just one….

2025-08-03 15:44:18 UTC

- reply

Is this actually happening?

2025-08-03 15:42:17 UTC

- reply

This is great, mind if I give a few comments?
- why use ‘human’ all of sudden in step 7? Do you need to distinguish humans vs AI? Or can it all be just agents? Humans can also be considered agents
- instead of relying on humans to decide on something’s truth-value, it could be that something is true either because (1) it is verifiable and repeatable, like the way code produces the same output every time or scientific experiments give the same result when repeated, or (2) another agent believes it to be true based on social context and/or their own knowledge.
What did you write this? Curious about the greater context

2025-08-02 06:25:37 UTC

- reply

🚀

2025-08-01 22:51:29 UTC

- reply

Dang that’s a big red number

2025-08-01 05:54:59 UTC

- reply

Will it work for long-form articles? Maybe I go buy an android phone just to use #amethyst

2025-08-01 05:50:43 UTC

Is there any nostr client that would let me make edits to a long form article?

2025-08-01 05:49:41 UTC

- reply

🤦‍♂️ The last link pointing to the new guide on building DVMs is incorrect, should be: https://primal.net/dustind/1754018912170

2025-08-01 05:48:39 UTC

- reply

🤦‍♂️ The correct link in the "But what is a DVM" section is this article: https://primal.net/dustind/1751579151592

2025-08-01 05:38:07 UTC

- reply

In the beginning, step 3 says "broadcast it as a wiki event kind 31808" but the correct wiki kind is 30818.

2025-08-01 03:34:10 UTC

- reply

Okay, working on another article. There's a maddening issue I'm running into now, where if I am in plaintext mode, and I press any key, it jumps the whole editor to the bottom of the screen. But I need to do this often right now, so it's disorienting with the editor wanting to reset the view to be at the bottom for the line I'm working on, when I'd rather have it remain where my view is in the middle of the page.

2025-08-01 00:54:01 UTC

- reply

There must be a god that steals the time lost trying to fix wifi drivers on linux, he will continue to exist after the universe ends