It would be nice John to eplain to non-technical people like me ... we hang out with you guys and we struggle with the technicalities :)
A free market for data processing based on Nostr currently developed by nostr:npub1l2vyh47mk2p0qlsku7hg0vn29faehy9hy34ygaclpn66ukqp3afqutajft
That's a really cool idea.
Discussion
Think of Data Vending Machines as little programs that can be assembled one with another.
That anyone can write.
For any kind of purpose.
Think of an algorithm, any algorithm. But being fully discoverable thanks to Nostr (read https://dev.highlighter.com/a/naddr1qqgrxwrfw5m8garwdvm85aenwajxgq3ql2vyh47mk2p0qlsku7hg0vn29faehy9hy34ygaclpn66ukqp3afqxpqqqp65wxtanuc for more on this)
Which means that you get long-tail disitribution of as weird and cool algorithms as any developer in the world can imagine.
Immediately monetizeable via #bitcoin.
They have a lot of very interesting properties, like transparently and effortlessly making all nostr applications that support them better.
Read https://vendata.io/magic for more on these magical properties.
As an example; I didn't want to add PDF support to Highlighter. But someone built PDF support without me having to write a single line of code.
And, at the same time, adding PDF support to any Nostr client that supports DVMs.
Like I said. It's magic.
What would be some examples that come to mind for how we can use DVMs with #[5]?
One idea that comes to mind would be a user describing a song they want to create, or a style, genre, etc, and then a DVM could return sounds related to that criteria.
Cool idea!
Where would the agent fetch or generate the sounds from?
Maybe that could be definable to when generating the request? Idk what’s possible so nostr:npub1l2vyh47mk2p0qlsku7hg0vn29faehy9hy34ygaclpn66ukqp3afqutajft could probably help answer that. For instance, perhaps it could just be sourcing the sounds from user posts on Stemstr by default. But there may be other sources of sounds online that a user would want to query, that could be optionally included.
There would have to be a way for it to identify the sounds according to the criteria. Not sure if that would be from user tags/hashtags, or something else. Could AI be intelligent enough to analyze a sound and categorize it itself? 😅🤷♂️
I haven't delved into what AIs can do wrt music analysis, but I would not be surprised if it's already really good at it
nostr:npub1nxa4tywfz9nqp7z9zp7nr7d4nchhclsf58lcqt5y782rmf2hefjquaa6q8 might know
but yeah; and the cool thing is that you don't have to have a single DVM dealing with these type of discovery, you could have ones that only discover sounds based on hashtags, but you could have much more sophisticated ones
that's the beauty of it, you can go really really long-tail with this because nostr solves discoverability
Not my speciality either, but as people are generating new Michael Jackson songs using AI, I'm pretty sure someone can build it.