I've been looking for a solution to help AI agents buy from AI agents.
I imagine a user with an agent giving them a prompt like "I need a ride to the party now". And at the best UX is that the agent simply takes care of it all with no further prompting so the ride shows up.
The user's agent is in a great position to help, as they probably know where the user is, where the party is, have access to a credit card or Bitcoin wallet, and know the user's preference when it comes to type of ride service and budget.
And of course there are services providers out there that would quite like the business.
What seems to be missing is a way for the user's agent to discover and engage with and buy from services.
I don't see the agents using search engines or navigating website user interfaces or making phone calls. They could, but it's not optimal. I also don't see them bothering to build their own API connections for every potential service.
I've been looking for a light touch interoperability framework. Perhaps the user's agent can simply broadcast the request, such that service provider agents can discover it and respond with quotes for the user's agent to select between. Some information exchange like a pick up address. Some payment exchange like a zap from a lightning wallet.
And then I start reading about dvms and how they work.
Can I use dvms for this?
Would it be a good solution?
#asknostr PABLOF7z (nprofile…4ph5) k. (nprofile…rjcl) Don't ₿elieve the Hype 🦊 (nprofile…hdgd)
rodbishop / Rod
npub1r0…rfsft
2025-01-18 06:30:28
Author Public Key
npub1r0d8u8mnj6769500nypnm28a9hpk9qg8jr0ehe30tygr3wuhcnvs4rfsftPublished at
2025-01-18 06:30:28Event JSON
{
"id": "09d0a0c3faab0fcf4bde13bddf37573b949622c1de29ecb2f754724eba3f7bba",
"pubkey": "1bda7e1f7396bda2d1ef99033da8fd2dc362810790df9be62f591038bb97c4d9",
"created_at": 1737181828,
"kind": 1,
"tags": [
[
"p",
"fa984bd7dbb282f07e16e7ae87b26a2a7b9b90b7246a44771f0cf5ae58018f52",
"",
"mention"
],
[
"p",
"bd4ae3e67e29964d494172261dc45395c89f6bd2e774642e366127171dfb81f5",
"",
"mention"
],
[
"p",
"99bb5591c9116600f845107d31f9b59e2f7c7e09a1ff802e84f1d43da557ca64",
"",
"mention"
],
[
"t",
"asknostr"
]
],
"content": "I've been looking for a solution to help AI agents buy from AI agents. \n\nI imagine a user with an agent giving them a prompt like \"I need a ride to the party now\". And at the best UX is that the agent simply takes care of it all with no further prompting so the ride shows up.\n\nThe user's agent is in a great position to help, as they probably know where the user is, where the party is, have access to a credit card or Bitcoin wallet, and know the user's preference when it comes to type of ride service and budget.\n\nAnd of course there are services providers out there that would quite like the business. \n\nWhat seems to be missing is a way for the user's agent to discover and engage with and buy from services. \n\nI don't see the agents using search engines or navigating website user interfaces or making phone calls. They could, but it's not optimal. I also don't see them bothering to build their own API connections for every potential service. \n\nI've been looking for a light touch interoperability framework. Perhaps the user's agent can simply broadcast the request, such that service provider agents can discover it and respond with quotes for the user's agent to select between. Some information exchange like a pick up address. Some payment exchange like a zap from a lightning wallet.\n\nAnd then I start reading about dvms and how they work.\n\nCan I use dvms for this? \n\nWould it be a good solution? \n\n#asknostr nostr:nprofile1qqs04xzt6ldm9qhs0ctw0t58kf4z57umjzmjg6jywu0seadwtqqc75spz4mhxue69uhhyetvv9ujuerpd46hxtnfduhsz9mhwden5te0wfjkccte9ec8y6tdv9kzumn9wshszxnhwden5te0wpuhyctdd9jzuenfv96x5ctx9e3k7mf0dv4ph5 nostr:nprofile1qqst6jhruelzn9jdf9qhyfsac3fetjyld0fwwary9cmxzfchrhacragpzamhxue69uhhyetvv9ujumn0wd68ytnzv9hxgtcppemhxue69uhkummn9ekx7mp0qy2hwumn8ghj76n9d3k8jenfwd5zumrpdejz7pqrjcl nostr:nprofile1qqsfnw64j8y3zesqlpz3qlf3lx6eutmu0cy6rluq96z0r4pa54tu5eqpz9mhxue69uhkummnw3ezuamfdejj7q6hdgd",
"sig": "6642a766e02ca5bad9234fa02a98137e7325d0fa24c1392d48175bdc623d7ff278d8ba5dd52852b1bd6221032a0f669c102d90a89bc2c022053b05a873d10aa3"
}