I seem to remember this sort of tech has been discussed before on Psience Quest, but I've been unable to dig up where.
In any case, here's a very recent article. I haven't followed the link to let alone read the technical paper, nor even followed the link to the original Meta announcement, so I don't know how accurately the article represents what's happened, but at face value it's very interesting. It could have gone in the AI megathread but I thought it deserved a thread of its own in this forum given its implications for mind, brain, and their relationship.
Meta’s AI Is Getting Better at Reading Your Thoughts—Without Cracking Open Your Skull
June 30, 2026
Webb Wright
Gizmodo
In any case, here's a very recent article. I haven't followed the link to let alone read the technical paper, nor even followed the link to the original Meta announcement, so I don't know how accurately the article represents what's happened, but at face value it's very interesting. It could have gone in the AI megathread but I thought it deserved a thread of its own in this forum given its implications for mind, brain, and their relationship.
Meta’s AI Is Getting Better at Reading Your Thoughts—Without Cracking Open Your Skull
June 30, 2026
Webb Wright
Gizmodo
Quote:The training for the new model, which was conducted at the Basque Center on Cognition, Brain, and Language in San Sebastián, Spain, involved nine healthy volunteers between the ages of 25 and 56 who were asked to type more than 2,500 sentences over the course of ten sessions. Throughout these sessions, their brain activity was monitored via magnetoencephelography (MEG), which measures the minuscule electric fields produced by neuronal activity. All of those typed sentences and brainscans then served as the raw training data that was fed into Brain2Qwerty.
In its most successful experiment, Brain2Qwerty v2 achieved a word accuracy—meaning more than half of the sentences that were decoded from brain activity contained no more than one word error—of 78%. In contrast, Brain2Qwerty v1 (which was released last year) achieved a score of 48% in its most successful case.
Mastodon