All the cool kids are working on FROST Today, I demo'ed a toy Cashu wallet to some Nostr/Bitcoin folks where spending from the wallet requires a signature to spend (this enforcement is standard Cashu, enforced by the mint), and the signature is made via FROST, and the signers coordinate via Nostr to exchange the FROST data I'll share some links later. It's not really usable, but I just found it fun to connect Frost+Cashu+Nostr View quoted note →

Hacking feels good this evening, mostly because I'm using it as a distraction from hunger 😀 I decided two days ago to cut back drastically on carbs, so I need the dopamine from seeing tests passing to substitute I'm suddenly craving chocolate

Noob question about @White Noise: Are DMs simply a full MLS group, but just limited by the app to two users? So the Marmot protocol would allow a third user (or a second device for one user) to be added? (That might be bad UX though)

bitchat

View quoted note →

Look after your knees! I've just recovered from a knee injury and I'm loving hiking again With the benefit of hindsight, I now realise that the relative lack of sport was affecting my mood and productivity for the last few months And I'm enjoying the gym again Bear markets are for (body) builders!

If you already know something about has Cashu Spilman Channels work, then you might find this useful if you want some details to help you integrate it into your server I have an 'integration kit' for each of four languages (Rust, Go, Python, TypeScript) where you can start receiving payments by just specifying a config yaml file and a few dozen lines of code Still more to do to tidy it up, especially on the client side, but the server side is more complete now (pending real-world feedback which could change everything of course 😀) https://github.com/SatsAndSports/demo_of_spillman_cashu_channel/blob/spilman.channel/ARCHITECTURE.md

The only correct instruction from the clanker in this video. Stop using VS Code and instead use ...:

🤯 I don't actually know anything about zero knowledge proofs, so no real idea how serious this is https://xcancel.com/osec_io/status/2028901596965552572

"Rebasing 442 commits on top of 132 new ones is a significant operation." No excuses, Claude. No mistakes, I've got a movie to watch while you're working ==== Seriously though, I've had a very mentally stimulating two days, lots of discussion of bitcoin and nostr tech, and a couple of more weeks of this ahead! So I'm finally watching The Hangover (2009), which I'd never watched before. It's excellent (Actually, using OpenCode instead of Claude)

Anyone working on a file storage system based on Merkle trees, paid via bitcoin? So I pay your (blossom?) server a monthly fee to store my data, where I know the Merkle root of the data. [I'll describe mutable data below] Every day, I ping your server for one randomly-selected (portion of a) file. You return the requested data, along with the Merkle proof that it's the correct data If any data is missing, I can then publicize this fact, and challenge you publicly (The data is encrypted of course, with a key that I never release to anyone, hence I don't mind publicly proving that the server has lost the ciphertext) I pay multiple servers to host the same data, and my client will automatically copy data to new servers (paying them, of course) when one server fails If I want to make a modification to one part of the data, then I simply send the new data to the storage server(s) and we update the Merkle root. The server signs the new root, committing to store it, and I sign a message which releases the old root [When I say "Merkle root", I probably really mean using something like git's tree of file hashes] #AskNostr ( @hzrd149 )

I told my agent to write some code "without side effects" and it thought that it was neither allowed to write nor read from the database. It felt funny to remind it that "no side effects" means it has permission to read, just not write. The point I'm making is that coding models can still be pretty dumb sometimes, even stuff that I would consider extremely obvious. GPT-5.2 Codex (via OpenCode)