Two things are happening at once, and they are usually covered as separate stories. Screens are learning to show depth again, in phones, laptops and monitors that need no glasses and no headset. And machines that have learned language and images are running into the part of the world flat pictures do not teach: how far, how solid, what is behind what, and what stays put when you move. This site is about the place those two meet.
The human half came first here and it still comes first. A photograph with depth in it holds a memory differently. A call where you can meet someone’s eyes is a different call. The reason to want any of this is not that a model somewhere needs training data. It is that presence, memory and shared work are better when space survives the recording.
The machine half is the harder argument. A generated world can look right from the camera and have nothing behind the facade, the way a film set does. Appearance can be generated. Geometry has to be measured. So the later essays ask where measured observation comes from, how you would grade a model that builds worlds, and what stays scarce as reconstruction gets cheap.
Three trades cover this and they rarely read each other. The AI press covers the generating. The robotics press covers the systems that fail in real rooms. The display and XR press covers the hardware people will actually look through, and rarely mentions the models. The argument here is that it is one story. The same depth that makes a memory worth keeping is the thing a machine has to get right before it can act in your kitchen.
Who writes this
David Fattal. Founder and CTO of Leia Inc., which makes glasses-free 3D displays, and founder of the DisplayXR project, a vendor-neutral effort that Leia supports as an early member and does not own. Written here in a personal capacity. Neither organisation speaks on this site, and nothing here should be read as a position of either.
Displays come up often in the essays, with a narrower role than you might expect. A display creates no geometric truth; cameras and calibration do. What a display changes is behaviour. Depth you can see is a reason to capture depth in the first place, which is the whole reason the consumer story and the data story are connected at all.
Signal
Signal is a machine-collected news river covering the same ground: what shipped, what was claimed, and what nobody has checked yet. It is automated on purpose and labelled as such, so a quiet week looks like a quiet week. Sources are chosen, not scraped indiscriminately: the market-research mills that mail-merge invented CAGRs across unrelated categories are excluded by policy, not by accident.