Every major spatial-computing product of the last decade has put the display on the user’s head. The Lightbulb Computer, a design project by Guillaume Ardaud, puts it in the lighting circuit.
The form factor is the whole idea
The device combines a projector with computer vision, in the shape of a large light bulb, so it screws into any Edison socket in a wall or ceiling — or sits in a portable lamp base.
That single decision resolves several problems at once, and it is worth being explicit about which:
- Power. A projector needs mains power. Every room already has mains power routed to the ceiling, terminating in a standardised mechanical and electrical connector.
- Position. Projection wants height and a clear line to surfaces. Ceilings are where lighting already is, for the same optical reason.
- Installation. There is no mounting, drilling, cable run or bracket. The installed base of compatible fittings is essentially every building.
Nothing about that requires a new standard, which is the usual tax on ambient-computing proposals.
What it does
The device responds to voice commands, sees where a person points, analyses what it sees, and projects content onto real surfaces.
The demonstrations are deliberately domestic rather than spectacular:
- Recipe widgets and timers pinned to kitchen counters, where your hands are wet and a phone is the wrong object to touch.
- A wall-sized map that two people can pan and scroll together — a size and a mode no headset offers, because a headset’s display is for one skull.
- Photos sent over from a phone and thrown onto a wall.
The argument it is really making
A head-mounted display is private by construction. That is a feature for a cinema and a defect for a kitchen, a studio or a living room, where the interesting computing tasks are frequently things two or more people are doing together: cooking, planning, looking at a map, choosing a photograph.
Projection inverts every property of a headset. It is shared rather than private, unworn rather than worn, fixed rather than egocentric, and it degrades gracefully — a projected interface nobody is looking at is just light on a wall, where a headset nobody is wearing is a device switched off.
The costs are equally structural, and the project does not get to dodge them. Projection cannot show you anything in a sunlit room, cannot render anything in front of a surface, has no stereo depth, and is occluded by the hand reaching for it. A ceiling-mounted camera watching a room continuously is a surveillance object in the place people are least willing to accept one. And the kitchen-counter demo requires the projector to know where the counter is, which is a geometry and calibration problem that the smart-projector category has repeatedly failed to make reliable.
What it is, and is not
This is a speculative design project with a concept video and a website, not a product with a price, a shipping date or a developer kit. Treat the demos as renders of an argument.
The argument is the valuable part, and it is cleanly stated: the industry standardised on the one display surface that cannot be shared, and in doing so skipped the form factor that was already wired into every room.
Sources
- The Lightbulb Computer
- The Lightbulb Computer: Reimagining Spatial & Ambient Computing — Guillaume Ardaud (YouTube)
- The Lightbulb Computer: reimagining spatial & ambient computing — Adafruit Blog