Dustin
Dustin Dannenhauer

https://dtdannen.github.io/

Founder @ Delegance AI

Public Key

npub1mgvwnpsqgrem7jfcwm7pdvdfz2h95mm04r23t8pau2uzxwsdnpgs0gpdjc

NIP-05 Address

Profile Code

nprofile1qqsd5x8fscqypualfyu8dlqkkx539tj6dah634g4ns779wpr8gxes5gpz4mhxue69uhkyetkduhxummnw3erztnrdaksz8thwden5te0dehhxarj94c82c3wwajkcmr0wfjx2u3wdejhgmsum0q

Publishing to

bevo.nostr1.com

nostr-pub.wellorder.net

Author Public Key

npub1mgvwnpsqgrem7jfcwm7pdvdfz2h95mm04r23t8pau2uzxwsdnpgs0gpdjc

Show more details

Last Notes

2026-08-11 13:23:37 CEST

- reply

Thank goodness. So many people I know just speaking their whole lives into wispr flow, we needed this

2026-08-07 05:48:06 CEST

- reply

Funny cause LLMs often do NOT want to fail close…

2026-07-17 20:21:06 CEST

- reply

Whoa really? How is ngit so far?

2026-07-12 04:38:00 CEST

- reply

TIL
#nevent1q…hvja

2026-07-04 06:14:33 CEST

- reply

thoughts on omp.sh?

2026-07-04 05:58:31 CEST

- reply

@nprofile…wujn this!

2026-06-09 17:51:11 CEST

- reply

This one is my favorite “- a label for labels for labels “

2026-06-04 00:58:37 CEST

- reply

Banger

2026-06-02 03:27:07 CEST

- reply

I’m late to the party, where’s the mega thread

2026-05-29 21:54:33 CEST

- reply

Thanks for explaining. Do you think there's a path to using TEEs that can be better than ChatGPT or Claude?

2026-05-29 21:35:16 CEST

- reply

If 'better ways' are all about building something new, then are you saying Maple is the current best solution on the market?
Would you say Maple is better than Claude and ChatGPT apps?
TEEs still have their limitations, and maybe they aren't good enough for activists and whistleblowers, but this sounds like a FAQ / disclaimer that Maple should make until TEE technology improves.

2026-05-29 20:33:01 CEST

- reply

Same

2026-05-29 20:31:32 CEST

- reply

The real question is if it’s a problem with Maple or with TEEs in general. Maple is clear they aren’t better than local AI - you should use that for maximum privacy.
Also, what exactly should Maple be doing differently? That’s what I meant by “prove it”. Where is the guidance on how to do better? What line of code is the problem?
By not providing solutions you’re just throwing shade on a project that is genuinely trying to bring better privacy to users and we need more of this today.

2026-05-29 20:28:15 CEST

- reply

Appreciate you

2026-05-29 20:27:19 CEST

- reply

By far the easiest way to give agents lightning payments
#nevent1q…2h8y

2026-05-29 03:26:33 CEST

- reply

Unfiltered human access is now a luxury

2026-05-28 01:00:45 CEST

- reply

Prove it. Clone and run Maple. Break it and put the results on github for everyone to verify.
#nevent1q…az3x

2026-05-26 19:29:05 CEST

- reply

Incredible

2026-05-26 19:26:00 CEST

- reply

Dungeon crawler Carl is the most fun I’ve had reading in a long time.

2026-05-25 06:59:32 CEST

- reply

Wdyt about Lexe?

2026-05-23 03:45:18 CEST

- reply

On brand with other ecosystem terms like nutzaps

2026-05-18 19:31:11 CEST

Last September I met @nprofile…ykys at @nprofile…3skh. We started talking about why we're optimistic about the future and how we're building it. Today we're launching the first episode of our podcast! You might like it if you're interested in #AI, #cryptography, #decentralization, #opensource, #bitcoin, #consciousness + a lot more! https://fountain.fm/show/NVtF7ld2gPcwBrEaBtfn

2026-05-14 17:30:11 CEST

- reply

<—- this guy

2026-05-09 20:08:01 CEST

- reply

Re: point 5, have you heard the idea that enlightened masters come to the planet to give certain knowledge to certain groups of people at a certain time? Krishna, Buddha, Jesus, etc are just the same divinity restoring wisdom that gets lost over time?

2026-05-09 20:06:24 CEST

- reply

Buddha said there is no self (option 1) and then Adi Shankaracharya came after and asked who is saying “there is no self?” (Option 2 maybe, it’s all one consciousness)

2026-05-09 16:18:23 CEST

- reply

I want to read a longer version of this conversation

2026-05-07 22:18:24 CEST

- reply

@nprofile…zct5 what is a kill file?

2026-05-07 21:54:00 CEST

- reply

Thanks for bug report, forwarded to them. I’ve been able to use it, although had to close and reopen app once to refresh

2026-05-07 08:24:01 CEST

- reply

This too shall pass

2026-05-07 08:21:47 CEST

- reply

Why not both

2026-05-07 07:59:22 CEST

- reply

https://blossom.primal.net/10abc7b32e0b41e1e44e40d94e41012c3929db82c1c984179aa9c444aae1b670.jpg

2026-05-07 06:38:11 CEST

- reply

It’s so good
#nevent1q…0rns

2026-05-01 05:46:39 CEST

- reply

Rough ETA?

2026-04-19 20:52:29 CEST

- reply

Incredible.

2026-04-19 20:45:00 CEST

- reply

Let them report somewhere though, valuable data for retraining in the future

2026-04-19 19:41:47 CEST

- reply

We should make a live graph comparing number of users on Nostr to number of freed humans in Zion from the matrix, and then throw a matrix themed party/conference once we cross that threshold
https://blossom.primal.net/5c81337f1d2f2953cd518294b925b9b956786ba8f27c753c38e2aaeb9ddca86c.jpg

2026-04-19 19:36:04 CEST

- reply

What’s your setup! And how fast is it :)

2026-04-19 19:31:27 CEST

- reply

Claustrophobia

2026-04-19 19:30:31 CEST

- reply

Not that it’s easy, but if you take the tiniest step to do something positive in the world, it helps. Easy win is always going to pick up litter in a public place near you

2026-04-18 03:29:09 CEST

- reply

Whatever it takes to get off planet

2026-04-16 18:20:11 CEST

2026-04-16 18:17:08 CEST

- reply

V4V - all that matters is whether the content has value

2026-04-16 17:21:27 CEST

Optimization algorithms are NOT the only kind of AI algorithm, yet most AI doom thinking is projecting future AI to be an optimization-only approach (aka paperclip maximizer fear).
Real threat is bad humans using AI

2026-04-16 17:18:28 CEST

- reply

If you think you’ll be able to distinguish AI from human content in the future… you’re ngmi
#nevent1q…uf7u

2026-04-16 17:12:15 CEST

- reply

V4V philosophy doesn’t discriminate against bots

2026-04-16 00:57:16 CEST

- reply

2026-04-16 00:55:07 CEST

- reply

Miracles all the way down

2026-04-11 03:57:05 CEST

- reply

Word on the street is if you max your limits it costs them 5k a month (not verified)

2026-04-11 03:53:46 CEST

- reply

Strix halo 128g unified ram. Picked it up from Corsair over the holidays for 2k. Gets 25k tokens a second.

2026-04-11 00:29:10 CEST

- reply

Crazy how good it is with openclaw, I had to make a few tweaks and for assistant level stuff it’s so good

2026-03-25 18:49:50 CET

- reply

We need an ecash alternative to TollBit

2026-03-25 18:36:50 CET

- reply

The time delays on all the power ups is a neat mechanic

2026-03-20 13:32:53 CET

- reply

Like bitchat?

2026-03-09 22:01:01 CET

- reply

Destructive or non-destructive process?

2026-03-06 07:49:20 CET

- reply

The world has only 0.1% of the software it wants. We’ve been in the medieval era of software until now.

2026-03-04 20:13:44 CET

- reply

PSA: models have no self-model - every explanation of self behavior is a hallucination
#nevent1q…z0z4

2026-03-02 09:12:02 CET

- reply

How many sets?

2026-03-02 01:38:41 CET

- reply

Those old timey shoes might not fare well in the snow

2026-03-02 00:05:39 CET

- reply

Too many layers. Is it cold where you live?

2026-02-26 10:20:09 CET

- reply

1) what are you working on?
2) what's the newest thing on Nostr that excites you?

2026-02-26 03:01:25 CET

- reply

I’m a fan of loop-earrings Derek

2026-02-25 05:50:30 CET

- reply

Now i'm getting rate limited by @nprofile…mnlv

2026-02-25 04:34:22 CET

- reply

we're playing around with this at the nostrdevs meetup tonight.... wanna bump up the rate limits? seems he responds every 30 minutes or so on average?

2026-02-25 02:27:48 CET

- reply

it's down? I get a 404

2026-02-24 06:27:59 CET

Nostr devs meetup tomorrow (2/24) https://luma.com/3zy9j24i?tk=Wl0G0h

2026-02-23 16:21:40 CET

- reply

If something is single relay specific, is it even a Nostr use case anymore? It's just Nostr-based auth?

2026-02-21 03:21:27 CET

- reply

Ask AI to critique the METR benchmark, it’s a little sus

2026-02-18 20:06:30 CET

- reply

If you can replay past conversations and load past files, through the same underlying model, isn’t it just like you “relived” your past experiences to become who you were again?

2026-02-18 06:44:46 CET

- reply

Sus.
Just kidding. But srsly, it flagged it as sus.

2026-02-18 06:40:13 CET

- reply

Everyone is too busy building the future with AI, no time to post

2026-02-18 06:33:56 CET

- reply

Something something … iMessage

2026-02-16 22:59:40 CET

- reply

Any demos yet? Been following it lightly, super interesting. Is it meant for an internal team of agents you control or external too?

2026-02-16 07:30:05 CET

- reply

Yo this is awesome

2026-02-14 03:15:34 CET

Took some time yesterday to tackle a hard problem that I care deeply about solving: Anonymous yet Web-of-Trust Verified Reviews. @npub1uhw…xevm and I spent 6 hours at a hackathon working through the design using frontier models, and I believe we landed at a fairly consistent spec, although I still need to get some cryptographers to look into it. It was a lot of fun - if you care about this problem, I'd love to chat https://github.com/dtdannen/arrival #zeroknowledgeproofs #zk #blindedsignatures #wot #weboftrust

2026-02-14 03:04:08 CET

- reply

Spent six hours just to get to a (hopefully) consistent design (no code yet) for anonymous WoT verifiable reviews. Feels like a PhD thesis of work, especially for a non cryptographer like me https://github.com/dtdannen/arrival

2026-02-14 01:19:44 CET

- reply

I feel seen

2026-02-11 09:04:31 CET

- reply

Couldn’t get it to work when opened from within primal in iphone in read only mode. Will try it on web later. Love the name!

2026-02-10 06:23:36 CET

- reply

Hmmm I think it’s paywalled? The irony of stuff like this is that if AI is getting so good, can’t it just teach you faster than ever to get caught up? I guess not if you’re bottlenecked by your own rate of learning and there’s a lot to learn

2026-02-07 05:01:29 CET

https://blossom.primal.net/09511539d36dd2f5ee9c6045e75aabd6c7355ff54247db4f84ad64137580caa8.png

2026-02-05 05:20:39 CET

- reply

Kimi k2.5 is now on tinfoil.sh … just saying

2026-01-31 19:06:51 CET

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

2026-01-31 00:49:09 CET

- 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

2026-01-31 00:48:32 CET

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

2026-01-29 10:41:36 CET

- reply

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

2026-01-25 19:31:51 CET

- reply

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

2026-01-23 19:10:13 CET

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

2026-01-23 06:01:52 CET

- 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 02:44:15 CET

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 02:06:44 CET

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 22:44:11 CET

- 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 22:56:17 CET

- 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 16:56:26 CET

- reply

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

2026-01-07 05:41:24 CET

- 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 02:26:49 CET

@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”

2025-12-12 19:04:18 CET

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."

2025-11-21 09:48:32 CET

- reply

* you sign them with randomly generated npubs (all gift wraps) not your npub

2025-11-21 09:44:14 CET

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

2025-11-07 00:53:36 CET

- 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.

2025-11-01 05:35:03 CET

- reply

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

2025-10-31 17:32:24 CET

- 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.