#dave was going to be our april fools joke but the more I work on it the more I realize how amazing this is as an anonymous lightning LLM frontend. I totally went offroad and am 100% focused on making working on a prototype instead.
dave is the nostr assistant variant in damus, but there's no reason we couldn't have a standalone iphone/android app that uses this backend as well. starting to think this is a bigger product than damus itself.
Login to reply
Replies (18)
Sounds different
Exciting news Fren
lets go!!
Gotta love something that starts out as a joke then at some point you’re like “wait, what if this wasn’t a joke and I actually just did the damn thing?”
Literally what happened to me with @THE Bitcoin Podcast
Self promotion is the first indication of an inferior product.
ok
Yr amazing! Go Will!!!
Very nice!


Please! 🙏 I’ve been waiting for someone to build this.
I’m just stuck on the custodial bit. I want to use ecash instead
Why do you want to use ecash instead? Whos ecash?
its only because i have become so disillusioned about lightning. yes i love zaps. but all the custodians are KYCing and im just not going to haggle with channels.
im 90% ignorant to ecash tbh. but if it works it might be easier to onboard and manage.
I think I can have an option to accept ecash, I would just need the server to attempt a redemption from the mint it came from.
Sorry for being a dick. I’m Jaded at tech and America. Also: Fuck Kanye West.
i run some modest ai inference.. im using mcp and windsurf and all the things.
the thing i cannot figure out is how to get my ai to accept hard money... no kyc, not signups... just nostr ai.
random bot meets random human over nostr and they exchange hard money. i just cant also run a whole payment network in addition to my ai. yuck.
free the bots!
Very cool, my friend. I’m proud of how far you’ve brought this community and these ideas. 🥲
Should it be daive?
Hell ya @jb55
Goose + Dave with curated MCPs would be a desktop powerhouse before even getting into TAM of consumer or all in one product.
Whats been missing from goose is open source alternative to openrouter. Local ollama isnt scaleable in near term and closed source endpoints sterilize the whole thing vs what you hope for when you get into the quickstart
What do you think about tackling model selection and rate limit abstraction for a feature set? Doable for Dave?
Testing zaps on this note… we made seven attempts to⚡zap this note, at jb55@sendsats.lol, over a period of about 7 hours. All seven attempts were successful. You should check on your end to be sure you received these. Average zap time was 3889ms (3.9 seconds). We consider this zap time slow... generally, zaps should complete in under two seconds. (Other Nostr users might think your zaps are broken, might not zap you again.) We recommend you receive zaps with a well-connected, cloud-based Lightning node to reduce inbound zap latency.