This is a scripted walkthrough. All people, belongings and exchanges are synthetic. It shows the experience from a person's side and the layers behind it, and it does not build the network.
YIMBY (Yes In My Backyard) is a concept for a system where resources can be shared between households by way of agents. Illustrated as a scripted walkthrough of neighbors borrowing and lending through household agents, YIMBY explores how agents might work at the household and inter-household level as resource managers. By opting in, an agent learns what the household owns and is willing to share. When a neighbor needs something, the agents negotiate the loan, the time and the place. Each human makes one decision: yes, no, or ask me later. The humans then show up for the handoff and the return. All people, belongings and exchanges in the demo are synthetic.
Asking a neighbor to borrow something feels like imposing, and declining feels rude, so most borrowing and lending between neighbors never happens.
Households that own things they would lend, and neighbors who need to borrow them, have this problem. Sharing between neighbors takes effort, it carries social awkwardness, obligations, and potential risks, and depends on existing social ties. As such, sharing resources does not happen as much as it could for ordinary and emergent reasons.
Part of the issue is households that own things don't always want to lend items for risk of not getting them back or not wanting them to be damaged. For those households that are willing to lend out items, knowing about the request requires existing social connections and coordination that folks can find burdensome.
Today's neighborhood networks are owned by platforms, which hold the graph, the data and the attention. These presume sharing is a communication problem; YIMBY presumes sharing is an information (who has what, who needs what) and a coordination problem (how to manage the actual hand off and retrieval).
This problem matters because shared resources and neighbor connections make a neighborhood more resilient, and household agents are becoming the layer that manages a family's errands and logistics without a good use for it.
YIMBY treats sharing as an information and coordination problem, not a communication problem. A household opts in, and its agent learns what the household owns and is willing to share. When a neighbor needs something, the agents negotiate the loan, the time and the place, so neither household has to ask or decline in person. The human makes one decision: yes, no, or ask me later. The humans still show up for the handoff and the return.
The demo is a mobile-first, scripted walkthrough with four scenes and toggles that show the bot chatter, who runs the network and a privacy view. An architecture tab maps each layer of the real system to what the prototype fakes.
If an agent removes the effort and awkwardness from requesting and sharing resources, neighbors will share more, and the exchanges will bring people together and make the neighborhood more resilient. If that holds, neighborly sharing needs a protocol, not a platform. Agents that speak open standards can coordinate between neighbors while households keep their data and any community can run its own node.
In terms of consumer use cases for AI and agents, YIMBY's broader hypothesis is that the unit of design for consumer AI is not just individual but the group (here the household) where tasks that fall through the cracks or lack clear responsibility can be accomplished with no or low effort.