Notes on the Looks Protocol
Completing the first meters.
At present our ability to connect online to others is incomplete. We can talk instantly to another person around the world, but someone sitting a few tables away is out of reach unless we already have their contact details.
The Looks protocol aims to bridge this gap, allowing anyone we can see to be contactable if they choose to be. The protocol uses the recipient's superficial appearance as a temporary addressing surface in a local context. When a user provides their current look to the protocol, others can use their view of that same look to initiate contact. No other user information is needed. Looks sits at the ends of a long chain of existing communication infrastructure, extending it to include ad-hoc p2p connections.
Looks is meant to live in smart devices, where local context is available and updating can be routine. Providing "my look" could be as easy as instructing a smart doorbell to update it whenever you leave the house. And adding "their look" would not require taking an explicit photo. A helpful assistant could remain aware of who is within reach and connect you based on a minimal cue. No ambient recording is required and anyone not using the protocol can be ignored.
Then directing a few words at someone within viewing range could get you an instant response, just like how we approach people within talking range. They might be a few tables away, at the other end of the street, or on a passing bus; a minimal view is enough to uniquely address them.
Privacy by design.
The need to explicitly opt-in and update your appearance in the protocol preserves the ability to remain private whenever you want. When you do use the protocol, it does not expose an online profile for everyone to see. Being present in real life is enough, your look is the profile.
This profile is inherently local: it only works where you can be seen. Any appearance you provide naturally expires as soon as you change your outfit, so it can no longer connect anyone to you. Looks connections are created for the moment without relying on existing accounts or profiles, so you can drop them without trace.
Rules of the protocol.
The Looks protocol is designed to see, never show. Therefore protocol implementations should follow these core practices: (i) The look images used to route the connection between sender and recipient are private and should only be used by the system to establish contact. (ii) The protocol should allow a user to silently refuse a connection attempt, without revealing whether they are using the system or not. (iii) Connections made through the protocol should be disposable; disconnecting should not leave a user traceable.
Some use cases.
The Looks protocol is a new primitive. Many applications can be built on top of it.
Chat. In a mature implementation, protocol operations disappear into the background. You simply speak to someone within sight and, if they have opted in, you get a response, or you do not. No different from initiating conversation within earshot. And you can still walk away cleanly, without leaving lasting ties.
Commerce. A vendor reachable through the Looks protocol could be engaged from a distance, no one having to walk more than necessary. Decentralised payment and contracting methods are a natural fit for the protocol's decentralised user addressing, minimising disclosed identifiers as much as possible.
Coordination. Interaction with and between automated systems like robots and kiosks could start off using the Looks protocol. When there are as many robots as humans, they can't all be called R2-D2. Differentiating them visually makes it possible to locally address them using the Looks protocol.
Using vision rather than physical proximity to limit reach means that any way of extending vision can extend reach as well. A zoom lens can resolve an appearance from further away. When a camera view is tied to its coordinates, a screen can become an interactive portal into that place.
← Back to slides