Holy shit, being able to run is a super power… I had to bring my car in for a service. The shop is in the next town. I simply ran back. Turns out you can just do things

I guess I should put this on @Umbrel ☂️ and have it claim directly to the LND/CLN client, essentially giving your umbrel a Lightning Address with zap support with close to no setup… View quoted note →

Need a privacy preserving wallet with a Lightning Address? Coco and npub.cash have your back! 34 lines of code:

Finally merged melt sagas in Coco 🥳

GitHub

Add MeltOperation (saga) with support for generic melt methods by Egge21M · Pull Request #54 · cashubtc/coco

Adds MeltOperationService to orchestrate melt operations end‑to‑end (init → prepare → execute → pending/finalized/rolled_back), with oper...

Freaking degens...

Everyone at home is sick :( Gotta bring out the big guns!

People liked my last air-fryer post, so here is my favourite recipe: Egg Muffins! 450kcal, 34g fat, 36g protein, 5g carbs 1. Whisk 3 eggs, add 90g of feta cheese 2. Add some spinach (fresh or frozen) and half an onion with a bit of olive oil to a pan and cook briefly 3. Mix everything together, add some pepper (feta adds enough salt) 4. Fill in muffin cups and air-fry at 160 degrees Celsius for 15 minutes

I know I said GLM 4.7 is my new favorite, but using the OpenAI sub inside opencode with Codex 5.2 is pretty impressive as well. I am on the 20 USD plan and have not hit any limits so far. Codex 5.2 high is MUCH faster than GLM 4.7 is.

npub.cash v3 has a massive amount of code quality changes, but I think the biggest impact on UX is going to be the responsiveness of the LNURL server. V3 is BLANZINGLY fast 🔥 (ps: v3 is a massive WIP and not yet usable)

My vibe odyssey might just have found an end. Recap: Just when I finally got to my happy place vibe setup (Claude + opencode) Anthropic decided to rip my heart out by banning any third party use of the Claude Code Pro and Max plans. Bye bye power tag-team opencode Opus. I tried a few things since then: CC API Key (expeeeensive), Zen (actually reaaaally awesome but Opus is expensive as well), Codex (good, but even though benchmarks indicate otherwise it feels way worse than Opus). This is when I started to use GLM 4.7 via Zen and I have to say, I am really impressed! It is currently free in Zen, but only for limited time. Zed AI got a great offer for it right now so I pulled the trigger (40 bucks for 3 months is a steal TBH). If you decide to go for it as well, consider using my ref code lol

GLM Coding Plan — AI Coding Powered by GLM-5.1 & GLM-5 for Agents & IDEs

Use GLM models like GLM-5.1, GLM-5 & GLM-5-Turbo for AI coding in Claude Code, Cursor, Kilo Code, Cline, OpenCode, Droid, OpenClaw and more. Plans ...

What are you missing in npub.cash? V3 is going to make many thing right, but let me know so I don’t miss anything