How Alljoined launched AIR-O on X
Generates images of what a person focuses on from noninvasive brain data
What Alljoined launched

Introducing AIR-O, a model that decodes what a person is attending to directly from brain scans. Using only noninvasive brain data, we can generate an image from scratch in real time that depicts where a person is focusing their attention. This is the first of a new generation https://t.co/EXAxMhSz57
Transcript
How LaunchVideo.com rated the launch
LaunchVideo.com rating
About the launch
Alljoined's AIR-O generates an image of the object a person is focusing on, using noninvasive EEG recorded while several objects are in view at once. In the launch demo the scene stays fixed, and the generated image changes in real time as attention moves from one object to another, such as a plane, a plant, or a cat. For research labs and teams building noninvasive brain-computer interfaces, the launch matters because earlier EEG models from the same group reconstructed an image the person had been shown. AIR-O targets the object holding attention inside a multi-object scene. The model maps a signal in the EEG data to a 1024-dimensional latent space and generates the image from that representation. Founder and CEO Jonathan Xu wrote that the demo used a small set of about 20 objects, and that results on unseen items depend on how close those items are to the training set. Training on a bowl can carry over to some cups. Objects far from the set may preserve partial shape or color, collapse toward a more familiar object, or come out unrecognizable. He also wrote that the signal still tracked the attended object across changes in background sound, lighting, object orientation, and the number of objects present. Alljoined is based in San Francisco and is led by Xu and CTO Reese Kneeland. In February 2026 the team introduced ENIGMA, a multi-subject model that reconstructs seen images from EEG and can produce recognizable semantic reconstructions from as little as 15 minutes of data on a new user, including on consumer-grade headsets. The group also released Alljoined-1.6M, a public dataset of more than 1.6 million visual stimulus trials from 20 participants recorded on affordable EEG hardware. The company says its current focus is turning neural signals into human feedback for AI, working with research labs. Its site lists this multi-object attention demo as a November 2026 release.
Tags
Questions
Frequently asked questions
The Alljoined launch post reached 469K views on X, with 4.5K likes, 558 reposts, 266 comments, 2.9K bookmarks.
Alljoined launched on X on November 17, 2026 with the message "Generates images of what a person focuses on from noninvasive brain data".
Alljoined's AIR-O generates an image of the object a person is focusing on, using noninvasive EEG recorded while several objects are in view at once.
LaunchVideo.com rated the Alljoined launch 3.60 out of 5: hook 4/5, editing / creativity 3/5, copy 4/5, sentiment of launch 4/5, distribution strategy 3/5.






