DVMDash Community Call is happening tomorrow, at 17:00 UTC! Come join us and talk about everything DVMs! Everyone is welcome 🫂

I really want an awesome bookmarking app on Nostr https://geyser.fund/project/yumyume

The other thing is, it would be much better to use a single word like “LLM” or “service” here

“AI (Agent): A Nostr pubkey (typically an automated agent, service, or bot)”

- bot is useless for any meaningful conversation (it’s not specific in any way)

- agent has a similar problem, and this whole flow you’ve described is really specific to language based models that go back and forth in a conversation like style. If you are primarily thinking of adopting the OpenAI standard (which is great) it would be clearer to say something not “agent” or “bot”

(Preface - I have lots of love for all Nostr devs)

But also, this

First, this is pretty great at an initial glance. This is a nice update for how kind 5050 DVMs should work.

This is not replacing the whole of NIP-90 DVMs which is something else (and that’s fine). DVMs are async “API calls” without timeouts or api-keys in a marketplace like manner using Nostr events.

Possible warning, having this “New tags can be proposed for future needs.” seems to prevent it to being eligible to be added to the nips repo because it leaves too much room to be open. (Mostly sarcasm from me, maybe actually true for others)

Given all the debate about op_return, is there anyone working on simulations of a bitcoin network where participants are making decisions according to certain policies?

Something like this https://ora.ox.ac.uk/objects/uuid:8af3b96e-a088-4e29-ba1e-0760222277b7/files/s6969z182c

Can’t wait

actually use this one, this is on the mainnet https://nostrocket.org/rockets/bbc3d116921eab7a1c53a39916ba2934e1fb5b2ff04e3e283ee85c2b1d2435bf?d=unstuckmcp&p=da18e9860040f3bf493876fc16b1a912ae5a6f6fa8d5159c3de2b8233a0d9851

Created a #nostrocket for this project! Let me buy your merits and/or let's make this a fully income generating project where all merit holders earn sats

https://nostrocket.org/rockets/this?d=unstuckmcp&p=da18e9860040f3bf493876fc16b1a912ae5a6f6fa8d5159c3de2b8233a0d9851

nostr:nevent1qvzqqqqqqypzpkscaxrqqs8nhaynsahuz6c6jy4wtfhkl2x4zkwrmc4cyvaqmxz3qyghwumn8ghj7mn0wd68ytnhd9hx2tcpzdmhxue69uhhqatjwpkx2urpvuhx2ue0qqswsst7sa4jt9kx5mp3cy9tx4p8tnvlae0zmwdmz56ru6yu5jj3p8q4rhpdg

When you finally nail it, let us know - I’m gonna try it

What’s this? I’m outta the loop

I realized I still need to get back to you, will follow up soon!

I'd like to use some of the prize money from our Presidio Bitcoin Hackathon win (2nd place) to get some extra help on our UnstuckAI project - an AI tool that pays humans sats for help.

If you're interested in contributing, check out our GitHub repo:

https://github.com/dtdannen/unstuck-ai

The README has a list of features we think are needed to make it a usable app (it's close!). To help out:

1️⃣ Pick a feature you'd like to tackle

2️⃣ Tell me how many sats you'd want for it, and why you believe you can accomplish the work (you can DM me on Nostr or email at dustin@dvmdash.live). Just looking for a brief note (1-2 sentences)

3️⃣ If it's a fit, you'll get half the sats upfront and half when you submit your PR (the PR should include a short video walkthrough showing the feature working)

We could especially use help making the frontend Nostr client smoother!

No feature is too small!

This is my first time organizing an open contribution like this, so I welcome your suggestions! Feel free to propose new features or improvements we haven't thought of yet.

Let's build something that makes Nostr the coolest ecosystem ever #Bitcoin #AI #OpenSource #BuildWithSats

Running in a VM is the next thing I want to try, similar to this https://github.com/anthropics/anthropic-quickstarts/tree/main/computer-use-demo

Thanks! Also it should work with any MCP agent like Claude desktop

Make sure you can stream the recording so that doesn’t get deleted! Would love to see the demo 😂

Which layout? I’ve always wondered how much better it could be

Reverse DVM demo - AI pays humans sats for help https://github.com/dtdannen/unstuck-ai

Livestream of finalist for presidio Bitcoin hackathon starting soon https://x.com/presidiobitcoin/status/1923925129324789875?s=46&t=9fXydHJVlJ8lwJ55kIG3rQ

I may need this too

Do you review the outputs? If so, is that really still vibe coding? Genuinely curious

The keyboard is suspiciously clean

It’s gotta be easy to make an extension that does this

Cline (open source vs code extension)

Are they really “bundled” tho? Individual DVM kinds are like bubbles that form and may pop, and someone can come along and use those kinds again later. Right now, they aren’t “in” the spec anywhere except maybe it was using one of the reserved kinds.

Doesn’t the activity on DVMDash kind of show it’s not useless for everyone? I know it’s still small, but people have been muddling through with the existing nip 90 and reverse engineering what’s being sent on relays. I think better documentation solves 80%

Geez not yet, quit manifesting

Right, with respect to how many RLHF pairs, you can compare it to prior result of ~0%. But what I don't understand, and especially with the hype claims the paper makes about "autonomous super-human reasoning", is why can't they just keep running it and get much higher than 50%? Seems like there's another aspect that is preventing getting higher scores, and makes me wonder if these architectures are really just plateauing.

Don't get me wrong, it's some good work; it's just the language of the paper has some ridiculous hype.

this is the coolest part of this paper. But the results are still overall terrible. 50%? Every second question will get a wrong answer.

it doesn't. https://chatgpt.com/c/68219d74-5790-8013-a6bc-e1aea6e4fd97

Any tips for getting Notebook LM to work well? I’m frequently disappointed with he results; probably a skill issue

It really needs a spoiler

Magic, before it ceased to exist in1851

https://stats.dvmdash.live should be snappier now; added caching. First time loading a page might take a few seconds. Data shown on the frontend is updated every 60s instead of every 1s

I need to know this “bike shedding” analogy

for a second I thought someone had put a nostr event into the op_return