Seed Future
A camera takes a picture. That isn’t enough. We teach vehicles and drones to see the way people do — by putting sight, shape, and motion together into one sense of the world.
Get early accessPeople don’t walk through a room using only their eyes. We hear, we feel distance, we notice movement. Machines shouldn’t either.
Color, faces, signs, the look of a place. Wonderful in good light. Lost in fog, glare, and night.
Shape and distance. A quiet sketch of the world in three dimensions, even when the picture goes dark.
Motion through rain and fog. It doesn’t take a pretty picture. It finds what’s coming anyway.
Each one has a blind spot. Together they fill the gaps — so the machine knows what’s there, what’s empty, and what might move.
Fog
On a morning like this, a photograph is mostly gray. A person still knows a car is up ahead — the lights, the shape in the mist, a sense of motion.
We keep the picture, and we keep the senses that still work when the picture fails. The machine isn’t asked to guess from fog. It is asked to use everything it has.
From the air
A farm from above is beautiful. For a drone it has to be more than that: where the ground is, where people are, where the open sky actually is.
We turn those mixed senses into a living map — occupied, empty, safe to pass. The kind of map you’d want if you were the one in the air.
The street
A crowded corner is not one object. It’s a lot of small truths at once: who is moving, what is parked, where the pavement is actually open.
That’s the job. Name what’s there. Leave the empty space empty. Don’t invent what you didn’t see.
Early access
We’re teaching this on real roads and real flights. Sign up with Google and we’ll write when there’s something you can see for yourself.
We only receive your name and email. No posts, no contacts, nothing else.