Went through the SDK docs before asking. On RN/Expo specifically, does Fabric run inside a Bare worklet with IPC back to Hermes, or drop into a native module the way llama.rn does via JNI and llama.cpp? Perf and memory footprint would look very different between the two, curious which path you landed on.
That line means that you don't need to create an account and get an API key from a provider (i.e. "asking for permission") to run inference. The main advantage is precisely that local AI runs on your terms, including how data is handled, and provably so, unlike cloud APIs where there's still an element of trust with the operator.
What still seems unsolved is how to safely use it on real private systems (large codebases, internal tools, etc) where you can’t risk leaking context even accidentally.
In our experience that constraint changes the problem much more than the choice of runtime or SDK.
This is all very ambitious. I am not exactly sure where someone is supposed to start. With the connections to Pear and Tether I can see where the lines meet, but is the idea that someone takes this and builds…Skynet? AI Cryptocurrency schemes? Just a local LLM chat?
Although an LLM chat is the starting point for many, there are many other use cases. We had people build domotics systems to control their house using natural language, vision based assistants for surveillance (e.g. send a notification describing what's happening instead of a classic "Movement detected") etc. and everything remains on your device / in your network.
Are there incentives for nodes to join the swarm (become a seeder)? If yes, how exactly, do they get paid in a decentralized way? Any URL where to get info about this?
its through the holepunch stack (i am the original creator). Incentives for sharing is through social incentives like in BitTorrent. If i use a model with my friends and family i can help rehost to them
16 comments
>...and without permission on any device.
I would be much more interested in a tool which only allows AI to run within the boundaries which I choose and only when I grant my permission.
(Disclaimer: I work on QVAC)
Should it be re-worded so as to make that unambiguous?
What still seems unsolved is how to safely use it on real private systems (large codebases, internal tools, etc) where you can’t risk leaking context even accidentally.
In our experience that constraint changes the problem much more than the choice of runtime or SDK.
Although an LLM chat is the starting point for many, there are many other use cases. We had people build domotics systems to control their house using natural language, vision based assistants for surveillance (e.g. send a notification describing what's happening instead of a classic "Movement detected") etc. and everything remains on your device / in your network.