Markus_mustwork
Markus Rother

https://mustwork.de/

co-founder of mustwork

Public Key

npub10penuwc0wm62lr0zmtvneae6p7zrz3t7c3sqapqw7mc03kdjvmas6jmweu

Profile Code

nprofile1qqs8sue78v8hda903h3d4kfu7uaqlpp3g4lvgcqwss80du8cmxexd7cpp4mhxue69uhkummn9ekx7mqdtwej3

Publishing to

nos.lol

Author Public Key

npub10penuwc0wm62lr0zmtvneae6p7zrz3t7c3sqapqw7mc03kdjvmas6jmweu

Show more details

Last Notes

2026-09-10 10:33:27 UTC

Turned off governmental emergency alerts on my phone.

2026-08-29 06:39:45 UTC

- reply

Da stimme ich Dir zu. Es ist nur so, dass auch der Begriff "Community" sich durch den Vorfall verschoben hat. Die Frage, ob NVK, Odell oder OpenSats genauso wie vorher zur "Community" gehören wird ja gar nicht gestellt. O.g. drei schweigen, und auch große Teile der "Community" schweigen dazu. Das ist, was ich kritisiere. Die Frage Inside Job oder nicht ist deshalb absolut essentiell. Das kann man doch nicht einfach so tolerieren, oder?

2026-08-28 16:49:09 UTC

- reply

Ich meinte folgendes: Die Wahrscheinlichkeit ist recht hoch, dass es ein Inside Job war.

2026-08-25 11:18:25 UTC

- reply

Ihr seid zu unkritisch, was den ColCard "bug" angeht.
#nevent1q…zq7s

2026-08-19 18:44:42 UTC

- reply

Get well soon

2026-08-05 15:45:24 UTC

Despite all tragedy, this is a moment where NOSTR shines.
Information is re-layed instead of de-layed.
I have very high confidence that I am receiving uncensored, high-signal coverage of the ColdCard exploit/rug-pull.
It feels as if this bubbling up of truth would not exist without NOSTR.
All the data mining, work, and risk-taking is highly appreciated.

2026-08-02 21:13:07 UTC

account empty now
evil attack expose key
heavy grief inside empty heart

2026-08-02 11:20:42 UTC

Another widely held assumption was invalidated: If software is open sourced you can rely in somebody (a good actor) to review the code, and the maintainers to be open to feedback.

2026-08-01 18:21:57 UTC

Whenever you hear: "If .. then you are probably safe", assume you are not!

2026-08-01 15:55:45 UTC

- reply

As far as I know, hex-dice cannot be built in a topologically fair manner. Also, hex dice are niche and hence potentially vulnerable to supply chain attacks.
High precision D6 are widely available, and much easier to manufacture.
Please reconsider.
Standard dice are a good tradeoff between number of rolls, ease of use and availability.

2026-08-01 10:19:31 UTC

- reply

You either need two decks or you reshuffle after drawing the first card. Do NOT mix the two decks!
Reshuffle after every word.
Honestly, rolling dice is a lot easier to not get wrong. Easiest to comprehend is coinflip.

2026-07-31 21:51:42 UTC

The only safe assumptions are:
1. any RNG can be compromised
2. any HW wallet can be compromised
1. calls for dice rolling
2. demands multisig

2026-07-31 16:56:20 UTC

First signal in my timeline on the topic.
#nevent1q…cxl7

2026-07-31 15:47:33 UTC

Roll dice, use multisig.

2026-07-05 20:51:48 UTC

- reply

And the swiss commentary is better, too.

2026-07-05 20:27:52 UTC

It took a world cup for me to get a consumer VPN.

2026-06-11 18:41:51 UTC

- reply

Very well structured talk!
Superb balance between simplification and detail. I am sure the analogies helped lots and made the topic accessible.
I think you really did a great service delivering Nostr to more people. The only thing missing in the end was a "where to go from here" section, maybe name dropping a few easy-to-use clients - a call to action. Good job regardless!

2026-06-08 19:34:38 UTC

The latest version of the #BTCPrague app does not work on GrapheneOS without Google Play Services.
In Aurora App Store you can use the tag 1744882493 to download an older version (2025) which works.
@nprofile…py4c

2026-05-19 20:26:23 UTC

- reply

Doesn't get more mature...

2026-05-19 20:08:14 UTC

- reply

Talking about the mechanics of how we're governed is not a sign of mature political thought.
Why should they care about your opinion?

2026-05-16 10:32:57 UTC

Will attend #BTCPrague in June. Would like to meet families who legally homeschool in Czech Republic. Please forward, if you know someone who could know someone.
@nprofile…py4c

2026-05-11 12:31:37 UTC

Building a self-improving memory system for my agents.
It came up with a very interesting, genuine thought by dreaming, that I could not find in any of my sources.
"AI agent delegation transforms procrastination from an individual failure into a culturally normalized, scalable collective behavior."
Articulate, philosophical, almost poetic.

2026-05-04 07:38:09 UTC

- reply

The artist name is Eric Kogan.
@nprofile…tf6v would you mention him and share zaps, if he had a nostr account?

2026-05-04 07:18:44 UTC

- reply

GM.
Brilliant image.
@nprofile…tf6v some time ago you said, you hand-curated the images. Not sure if that was a joke. If true, please add a caption of artist, source, etc. or embed in EXIF data, unless blossom strips those. Alternatively, please attach the prompt and model with the image.
Would be much appreciated.

2026-05-02 10:08:10 UTC

- reply

Ich lese den Titel und weiß sofort, wer die Showhosts sind.

2026-04-20 10:25:33 UTC

All code is boilerplate.

2026-04-03 13:53:04 UTC

- reply

What shocked me on that episode was Rogan leaning into transhumanism.
Maybe he's just hanging out with the wrong people too much.

2026-04-01 21:03:57 UTC

- reply

Kennst Du den Film American Pie?

2026-03-21 14:20:48 UTC

- reply

who is that?

2026-03-21 13:50:08 UTC

Sometimes I wonder who's the one in need for therapy...
---
name: prompt-therapist
description: Analyzes ZeroClaw's full system prompt (workspace files + hardcoded Rust fragments + runtime composition) against prompt engineering best practices. Backs findings with evidence from real conversation logs. Produces structured therapy reports. Use periodically to optimize agent behavior.
model: opus
memory: project
skills:
- conversation-analysis
---

2026-03-21 06:57:04 UTC

- reply

Hi,
I am working on a learning app for children myself right now (not speech based, though).
Here are my thoughts:
* The UI is very appealing.
* Setup in the beginning is a bit convoluted. Start with defaults instead of bombarding users with options.
* BUG: Whitespace between words is not stripped and creates false results if there is more than a single whitespace.
* Replaying generated speech actually creates new output instead of merely replaying. This could be optimized. It wastes compute (unless intentional, of course).
* I was really surprised when out of nowhere a game popped up. That felt a bit cringe.
My general recommendation:
Focus on the core functionality. Right now, there is a lot of distraction. That sharply contradicts the idea of setting "focus time" with a timer.
Good luck.

2026-03-17 23:55:16 UTC

One of the best properties of #marmot is imho that it ends custom event type inflation.
As a mental model the group serves as namespace for custom apps, without declaring event kinds publicly.
Great design choice!
Highly appreciated.

2026-03-16 21:50:21 UTC

- reply

Software building Software is only the obvious first step. It will grow exponentially and surpass human intelligence. Although I am not in favor, I'm afraid it will be inevitable.
Then it will be machines building machines, because the intelligence to do so will no longer be an obstacle.
None of what those can replace is safe.
If anything is safe, it is activities that require a soul, empathy and humor -- human to human relations -- but I would even add a question mark to that, because man will feel utterly miserable in second place and will want to join the beast.
I wish I were wrong.

2026-03-10 14:55:07 UTC

- reply

Lot's of cope in this thread.

2026-02-27 19:35:26 UTC

- reply

> That key is used to encrypt data throughout the process so we can't see your data.
Unpopular opinion: It has been bothering me for a while. Statements like that are slightly misleading.
Obviously, the data has to be decrypted before it is passed to the model. So transport may be encrypted, and also you may encrypt context data at rest, but there is no such thing as a large language model that can process encrypted data.
There is nothing that technically keeps @nprofile…uh5t from tapping into the data.
Ultimately, it's a trust model, not a technical solution.
Prove me wrong.

2026-02-26 22:12:18 UTC

- reply

I don't think so.
The next generation does not understand why they need to signup somewhere, why they would need an Email, or why some detached entity would have the right to exclude them from content.
Hang in there.
It only takes one killer app. Only a matter of time, imho.

2026-02-26 22:07:50 UTC

- reply

Of those, proof of humanness is the only category that leaves little room for interpretation.

2026-02-25 21:08:49 UTC

- reply

Bitcoin's success did not need professional product designers, either.
Nostr needs product designers as much as Bitcoin does. Neither is a product.
Both are protocols each of which allow a network to develop organically, and there is no better proof of genuine superiority than organic growth.
Nostr apps do need good product design, just like hardware wallets do, but that's a totally different issue.
You have built products yourself @nprofile…qyxr. Honest question to you: What were the obstacles for better product design?

2026-02-23 21:13:51 UTC

You can tell openclaw was written by a multi-millionaire.
Token usage was definitely not the top priority.

2026-02-21 19:23:06 UTC

Experimenting with running openclaw in sandbox with local model.
It forces you to be efficient with context, and more thoughtful with workflows, because the model is weaker.
It's been slower and needed some help occasionally, but so far there was nothing it could not do: research on web, news aggregation, task management, programming tools for itself, posting to nostr, etc.
Curious what will happen once I crank it up.

2026-02-21 09:44:46 UTC

- reply

That's pretty close to the distribution of fiat money.

2026-02-16 16:07:23 UTC

Big lesson here:
> Why do you think your work, Open Claw, won?
> It’s hard to compete against someone who’s just there to have fun.
Peter Steinberger on Lex Friedman 00:22:19

2026-01-11 20:32:02 UTC

- reply

Falling down a new rabbit hole?

2026-01-07 09:53:05 UTC

There is no viewcount on Nostr.
Social media platforms are to content creators like drug dealers are to addicts. They have full control over your rewards system.
By signaling social approval the platforms are in full control. That is powerful. And so is resisting to it.
I stopped posting on LinkedIn.
Tired of playing the attention game.
I think I'm starting to understand the meaning of "Shitposting" -- Just saying what's on your mind and not caring. It's an opt-out.
GM

2026-01-05 19:53:41 UTC

- reply

it should. but does it?

2026-01-05 19:24:08 UTC

- reply

Different relays?

2026-01-05 13:14:15 UTC

- reply

Deploy bitchat to the battlefield.

2026-01-05 12:35:47 UTC

- reply

@nprofile…r3z3 Consider isolating agent instances with one nsec per chat/group/dm.

2025-12-27 00:25:39 UTC

- reply

Auch für mich als Zuhörer eine emotionale Zeitreise.
Ihr hattet Anteil daran, mich gut durch die Corona Jahre zu bringen.
Ein Funkfeuer der Vernunft, des positiven Ausblicks und des Humors.
Danke.
Frohes neues Jahr. @nprofile…5tr0 @nprofile…jgy8 @nprofile…3g9u

2025-12-01 14:53:14 UTC

- reply

🎶❄️Bitcoin dippin' near the lows🎄🎶

2025-11-30 19:00:22 UTC

- reply

Hallo @nprofile…lxzu,
> ⚡ Der Wert eines freien Geldes entsteht nicht nur durch das Halten, sondern durch seine ständige reale Nutzung.
Muss Dir mitteilen, dass das Quatsch ist.
Die Wertvorstellung der Marktteilnehmer ist immer latent vorhanden. Im Handel wird sie offenbar. Wenn 1 Bitcoin für mich jetzt schon10 Mio EUR heutiger Kaufkraft wert ist, warum sollte ich den dann ausgeben!? Ich müsste unter Wert verkaufen. Es geben nur diejenigen ihre Bitcoin aus, für die der subjektive Wert eben die aktuelle Kaufkraft widerspiegelt. Alles andere ist irrational.
Natürlich gebe ich stattdessen meine Shitcoin EUR aus solange ich noch welche habe (Gresham's Law). Nach Deiner Argumentation würde ich damit umgekehrt durch das Ausgeben meiner EUR den Wert der Währung EUR steigern -- was offensichtlich nicht stimmt. Das Gegenteil ist der Fall.
Der Preis von BTC steigt gerade dadurch, dass so wenig Menschen ihre Bitcoin ausgeben wollen.
Ergo: Hodln was geht!
Peace, u. VG.
M.

2025-11-09 02:52:23 UTC

Vibed >50k lines of code in the last 18 hours.
And stuff just works.
now: 90% of time spent on design choices.
then: 90% of time spent reading and writing code.
Can't believe how there are still devs out there who don't get how much of a game changer this is.

2025-10-24 06:54:22 UTC

Developers, try harder!
There is a race going on, and us developers are losing.
Product managers are experimenting with low code, producing prototypes, validating ideas with application generators.
And us developers are sitting on our hands, thinking we will remain irreplacable forever.
❌ "Real use cases are too complex"
❌ "Where's the CI/CD pipeline?"
❌ "This will never achieve maturity"
❌ "And, lol, look how ChatGPT can't even count correctly"
If that is you, you better make up your mind.
AI will change coding forever. In particular, it will disrupt or has already disrupted how ideas are translated into functionality — which is what programming is about.
Yes, I know you are using Copilot for code completion, because we are lazy and always have been.
Is "using the tools" really enough?
Own it!
This could be the rebirth of verifiable software, yet nobody is talking about it.
Or what about property based testing to force the spec to do what it promises?
Merely mapping requirements to behaviour won't do much longer.

2025-10-23 16:03:03 UTC

I often hear the analogy that #VibeCoding to generated high level code is like written high level code is to machine language.
That is wrong!
Compiling is a reduction. Information is lost in the process, such as meaningful variable or class names, or comments.
When turning spec to code, on the other hand, a lot of information is added, such as naming conventions, choice of frameworks, etc.
I am not opposing the beneficial hallucination. Not at all.
All I am saying is that intent needs interpretation, and interpretation is dependent on the culture captured in the LLM. It is a nondeterministic process, inherently fragile.

2025-10-22 18:09:49 UTC

- reply

You didn't get my point.
I am talking 100x from now.
Also, I am not talking about funding lightning channels.
I am talking about pre-coiner, normie savings plans.
And, yes of course they should keep their part of the signing keys in cold storage.
In a more humorous note: Why do you think a Spillman Channel is inappropriate if the system is truly trustless?

2025-10-22 15:26:16 UTC

- reply

That's what I meant. The funds are locked until lock time expires.
So no trust needed in the sense of: "money cannot be stolen". But needed in the sense of: "OMG. I cannot access my money until X", because there is no such thing as a force close by the client.
If the envisioned lock time is not hours but months, that becomes relevant. If on the other hand the lock time is "only" hours, the use case needs to be very strong in order to outweigh the need for two transactions (lock and redemption), imho.

2025-10-22 14:25:27 UTC

@nprofile…22jg, I like that you are bullish on #SpillmanChannels
A few random thoughts:
If the opener (sender) of a Spillman Channel has no way to recover funds other than locktime expiry, the barrier to entry is somewhat high.
On the other hand, other recovery mechanisms probably require punishment and/or watchtowers of some sort.
The risk of permanently locked and unused funds in a channel grows with the envisioned duration, so the incentive works against the problem to be solved (negative feedback loop).
Therefore, there still is some trust involved.
Then again, some trust can be assumed anywasy, if the sender is paying money to another party...
What do you think?

2025-10-22 14:07:04 UTC

There are only two ways to validate a Spec Driven Development pipeline:
1. Spec → Code → Spec :: no semantic loss
2. Code → Spec → Code :: no functional loss
#vibecoding #SDD

2025-10-21 12:22:23 UTC

Don’t worry about being wrong — that’s the point.
Pair programming sessions don’t usually derail because of personality or skill gaps.
They derail due to the lack of declared intent.
▸ "Let’s have a look."
▸ "I’m trying to understand."
▸ "It’s probably something very obvious."
It sounds curious, humble, even open-minded — and yet, it usually means:
⚡ "I don’t know WTF I’m doing."
⚡ "I’m not ready to commit to an explanation."
⚡ "I’m hoping something will jump out at me."
Without intent, without a hypothesis, every observation feels relevant. Nothing can be wrong — because nothing specific was expected.
Obviously, not knowing is a valid starting point — maybe the only valid starting point in any endeavor, for that matter.
However, declaring ignorance is a comfort trap. It can be a justification for remaining indecisive.
Conversely, even the most ridiculous hypothesis shifts focus from activity to validation.
You will be held accountable not only for your ridiculous thoughts but for the progress as well. So that's nice.
That is the feedback loop you are entering when stating expectations.
💬 “I expect the result to be X.”
💬 ”If this condition is true, we should skip the database query.”
💬 “If X is null, Y should throw.”
Say it aloud, even when not pair programming to turn wandering into a controlled experiment.
And also, my self-talk signals to my spouse and daughter that I am in a state of deep concentration — rather not to be disturbed.
Hashtag#PairProgramming Hashtag#SoftwareEngineering Hashtag#Agile