Verifying my Nostr Nests identity: 0bISV0yC1ctZ45U-k67DqYFzUZX_dmFM9Ciwe5jT2-A
I could talk about meditation and breathing techniques. Did my teacher training through the Art of Living in 2012 and practicing since 2009.
Tried that, saw this in dev tools, could my vpn be an issue?
Stuck like this on Nostrudel, tried re logging in and clearing my cache. Should I do anything else? Been probably 1-2 weeks since I used it.
So it begins….
nostr:note1djtwh2czjqq5hxsyhgk0arh35tmsz3nj043983mtxf9sn9vvf8rs2672u7
Client DVM feature idea: generate a photo for your profile pic
Could use the user’s bio or let them say what they want the pic to look like. They pay the request and then the client automatically swaps our their profile pic with the one created from the AI
Oh absolutely! I hijacked your comment to dump some thoughts, sorry lol
I know there's a difference between DVMs and APIs, but I haven't tried to articulate it yet. Some thoughts on my mind:
- DVMs are waiting to respond to your event. Sending a note to a relay vs hitting specific API endpoint with w/e auth and other requirements by that vendor
- payments are already taken care of vs. APIs often require KYC, various payment methods, etc.
- activity between DVMs is public (with an option to encrypt notes) while APIs are private. Being public means failures can be seen by others and replacement DVMs could step up to take over the task
… and I know I’m missing a few other good differences
Some interesting DVMs are popping up. Found on https://noogle.lol/nip89
Have you heard the saying that goes something like “the more connected a players stones, the more likely they are to win”?
Also, I loved “How Not to Play Go” - it’s a tiny little book that covers 3 games, it might be helpful at your skill level (if you haven’t read it).
Usually I pick a much stronger deck for the next game. Commander politics gets brutal
All contributions are welcome. Who knows, if you don’t build what you’d like to have on Nostr, maybe no one else will.
Also it’s fun isn’t it?
Are there any LLM projects for entire code bases?
Feels like I should be able to train an LLM on Damus source code to get suggestions on how to add translations into Nostur…
The slides aren't really standalone FWIW, it needs the talk track. Still, hopefully some helpful references.
Psssttt put all that heavy stuff into DVMs
When it comes to language translation, it seems that the trend is the biggest ML models do the best. Meta's AI lab that has arguably the best translation models found that training a giant NN to translate from any 1k languages as input to any 1k as output, performs much better than a NN model specific to any single pair of languages (like only English <-> Spanish).
Anyway, this is all to say that maybe language translation quality vs on-device compute has been a blocker?