"NONE OF THESE EXISTS YET!" one of the most inspiring things I heard this week

Anthropic is so hot right now

I have an irrational aversion to UUIDs

Just dropped my iPhone and got to see the other side of its resiliency. FML

When the answer is in the question... View quoted note →

It's so hard to find good help these days....

technical debt is only debt if implementation has a cost

I accidentally pumped AI responses into AI in a loop. It got weird really fast.

now that @Cash App supports paying in lightning, and that Goose agents can have their own NIP-60 wallet

GitHub

docs: add mcp-money extension tutorial by pablof7z · Pull Request #2851 · block/goose

Add comprehensive documentation for mcp-money MCP server Include installation instructions for both UI and CLI Document features including Cashu ec...

all robots are missing to go buy groceries is a body. "goose build yourself a body and go buy me a steak"

PERO EL AMOR ES MAS FUERTE

atlas has been awake since 4am. Kill me now 💀

@Gigi and I discussed this idea on a @No Solutions episode a while ago; let's give it some shape: "Prompt diff NIP" Imagine, every time a new release of my favorite nostr client comes out, I push a button and all my prompt diffs change the client to make it MINE, to personalize it in whatever way I want, add, or more importantly remove, features. (btw the "push a button" can be automatized too 😂)

GitHub

Add LLM stuff nip: Define prompt diffs for LLM modifications by pablof7z · Pull Request #1950 · nostr-protocol/nips

This NIP defines a schema for prompt diffs, the idea is that people will be able to continuously customize software by applying a set of patches. I...

😂😂😂😂😂😂

we are moving so fast I feel like I should be a lot more explicit about the stuff I'm working on, there is a crazy amount of synchronicity going on, specially here on nostr. 👇👇👇 Right now I'm working on pair-programming where two LLMs collab with using distinct models and different instruction-sets. It's using NIP-29 to create on-the-fly groups to tackle individual tasks. I am now using my Walker project + Goose to get two agents (with distinct definitions) to discuss architectural designs before jumping on implementation. Both agents have full context of the product spec and the tasks provided. Since all discussion is happening on nostr the human can participate if he doesn't like an overall product decision and chime in the chat. Still a WIP, but this is what I'm working on today.

I haven’t felt this excitement and vibrancy since discovering the internet