One day I was thinking about the future, as I usually do. I thought to my self that apps will probably become obsolete. We will probably each have this small device that contains an LLM installed locally. Businesses and services will all have their own AI agent. So I put together this concept for a decentralized P2P network that enables discovery using pubsub, and allowing agents to broadcast their capabilities.
It's currently at a proof of concept stage, but already has concepts that handle identification for agents, capability negotiations and even consensus strategies.
I'm really curious peoples thoughts, but I think this could be really powerful and a fundamental step towards ASI. I'm not trying to necessarily "promote" a product but instead promote an idea to facilitate how we may use agents in the near future.
Hmm maybe I stick with my current setup in using Vagrant with big projects, and add something like CodePen or JSBin to mix for quick features. Then when I'm happy append the code inside the vagrant box. Gah.
Hey guys, I've been working on this for a few months as I've been really interested in building mobile apps with Meteor. Through out the process of building this I had the intention of licensing the software and make some money off this project. But I've come to realize with the direction everything is going with mobile and shopping that releasing this as an open source project would have huge benefits for everyone. I plan on working on this consistently, but now hopefully have some great minds contribute as well.
I'm currently setting up a demo for people to see, but if any one needs help deploying this to their server and building the app it self, let me know as I'd be happy to help!
It's currently at a proof of concept stage, but already has concepts that handle identification for agents, capability negotiations and even consensus strategies.
I'm calling it Ecco. It's open source and can be found here https://github.com/dileet/ecco
I'm really curious peoples thoughts, but I think this could be really powerful and a fundamental step towards ASI. I'm not trying to necessarily "promote" a product but instead promote an idea to facilitate how we may use agents in the near future.