Guillaume Ardaud pitches Lightbulb Computer as a smart glasses alternative

In a piece published in September 2026, Guillaume Ardaud presents the Lightbulb Computer, a speculative device that combines a projector with computer vision in the shape of a large lightbulb. He frames it as a series of design prototypes and as an alternative to what he calls modern visions of human-computer interaction that are excessively focused on smart glasses and headsets, devices he describes as constraining and socially awkward. His idea is to project information into the real world to augment it, rather than keeping it on small rectangular screens or piping it through devices worn on the face.

The lightbulb form factor is the practical hook. According to the author, it can be mounted in a small portable lamp-like base and carried around the house, or screwed into any wall or ceiling Edison socket for a permanent room-scale setup. The device reacts to voice commands, sees where the user points, can analyze what it sees, and can project content and highlight things in the real world.

The piece walks through use cases. In the kitchen, recipes, timers and mise en place overlays can be pinned to surfaces, which the author says is easier with wet or flour-covered hands than unlocking a phone. When studying or reading, you can ask questions about the content and get the answer right there. For trip planning, a giant projected map replaces huddling around a phone, with basic gestures to pan and scroll and toggles for transit overlays. Photos can be sent from a phone and projected on any surface, and the author suggests board games that are quick to set up with nothing to tidy afterwards. Smart home devices can be controlled by pointing, since the device sees what you point at and you do not have to remember their names. Other ideas include an ambient display by the bed showing the time, alarm, weather or next-day calendar, and a shared family board in the hallway where items are checked off by tapping. The author also says modern vision models can read information in the world in milliseconds, and that he would need at least a few minutes to go through all the books on his shelves by hand, whereas he can simply ask the device. A 3D map of the area around his house can be overlaid with weather or time-specific information such as conditions at ski stations, and made interactive, for instance by asking it to highlight a mountain village he forgot the location of.

The argument against eyewear-based computing, which he groups with VR, MR and AR, runs on several points. Putting something on your face might interfere with hair, makeup, jewelry, or glasses and contact lenses. The form factor must be very small and light, which limits power, computation and thermals. Most important to him, it feels fundamentally antisocial, because everyone needs their own device and those devices must be compatible with one another for any shared experience. He says this does not address the deeper problems of screen-based computing, which has been criticised as addictive, antisocial and reductive.

He acknowledges a long history of projector-based computing: research projects such as I/O Bulb and LuminAR, products like the Humane Pin, community projects like Dynamicland and Folk Computer, and installations in museums and public spaces. In his view, none of those cover personal computing, the kind of things done today with laptops, phones and tablets at home or at work, and that is where he sees room for reinvention.

On status, he says all the videos shown come from a real but bulky prototype setup, with no video editing, special effects or AI generation. He does not think the technology is quite ready for a consumer version today, but does not think it is far off, citing tremendous recent improvement in computer vision and bright, high-resolution, small-footprint projectors now being a reality. Privacy would have to be a first-order consideration; he says existing best practices can address it, such as on-device processing and hardware guardrails: a camera that only turns on when the user activates it, with a hard-wired indicator light and a physical cover. He concludes that, given the alternatives the industry is pushing, this is a path for the future of computing worth pursuing, and offers more demos and behind-the-scenes detail on the prototypes if there is interest.

Key facts

The Lightbulb Computer is a speculative device by Guillaume Ardaud: a projector combined with computer vision, in the shape of a large lightbulb, shown as a series of design prototypes.
It can sit in a portable lamp-like base or screw into a wall or ceiling Edison socket, and it responds to voice, sees where you point, and projects and highlights content in the real world.
Use cases include kitchen widgets, a shared projected map for trip planning, pointing to control smart home devices, an ambient bedside display and a family board in the hallway.
The author argues eyewear computing is constrained in power, computation and thermals and feels fundamentally antisocial, and that earlier projector efforts such as the Humane Pin and Dynamicland do not cover personal computing.
Videos come from a real but bulky prototype with no editing or AI generation; the author says the technology is not quite there yet for a consumer version today, but he does not think it is far off.

Why it matters

The piece is a counter-proposal to the smart glasses and headset direction that the author says the industry is pushing. Its core claim is that computing could be shared and ambient, projected onto the room, instead of confined to small screens or worn on the face. The author argues that earlier projector-based work (I/O Bulb, LuminAR, the Humane Pin, Dynamicland, Folk Computer) did not reach personal computing, the laptop, phone and tablet kind of tasks, and that a lightbulb-shaped device is a way to get there. It is one designer's vision, but it names the specific objection to eyewear: everybody needs their own compatible device for any shared experience.

Who it affects

Nobody is affected yet; this is a design concept. The use cases in the piece are aimed at people at home: someone cooking with wet or floury hands, people planning a trip together over a map, families sharing a hallway board or looking at photos, and anyone with more smart home devices than they can remember names for. For people following wearable and spatial computing, it offers an argument from outside the eyewear camp.

How to use it

You cannot. No price, release date or commercial availability is given, and the author says the technology is not quite there yet for a consumer version today. What is available is the written piece with its demonstration videos, and a video version on YouTube. The author says there are more demos he has not shown and that he can share them, along with behind-the-scenes detail on how the prototypes were built, if there is interest.

How solid is it

This is a personal design exploration by a single author, with no company, funding or product plans named. The author says the videos are from a real prototype setup with no editing, special effects or AI generation, which is the strongest evidence in the piece. No hardware specifications of the prototype are given, no user study, benchmark or measured accuracy of the vision or voice recognition is reported, and the source does not name which vision models or software the prototype uses. The privacy measures are described as what can be done, not as implemented in the prototype.

Risks and caveats

The main claims are the author's opinion: that eyewear is antisocial and that this is a path worth pursuing. The author himself says the technology is not quite there for a consumer version, and gives no timescale beyond not being too far off. A device with a camera that watches a room makes privacy a first-order concern, and the safeguards he lists (on-device processing, an activation-only camera with a hard-wired indicator light, a physical cover) are proposals rather than shipped features. The source does not say how the prototype handles multiple users or lighting conditions in a bright room.

“I think in light of the alternatives that the industry is pushing right now, it's a path for the future of computing really worth pursuing.”

— Guillaume Ardaud, The Lightbulb Computer