Telekinesis for Typists: Glasses

Telekinesis for Typists: Glasses That Read Your Silent Mouthing

Ah yes, another genius idea. Seriously. I mean, this is practically... telekinesis. The ability to talk silently, to communicate with a colleague sitting across the way without making a noise.

Ah, the good old days of tossing your homie a crumpled-up piece of paper in class, only for the teacher to instantly catch you and read the note out loud: "You dweeb, why'd you wear the same shirt as me?" Mega embarrassing as an adolescent. Those days are no more. Not that I ever did any of that, of course. Mine would have been to the cute girl sitting next to me, asking her out, which is that much more embarrassing.

Here's the short that started all this:

Then the thought has to come around to: what else? Can a technology like this be used for surveillance? I think it's too "specific." A sensor on a pole, I'd say, isn't practical. We aren't all walking around muttering to ourselves. Well, not all of us. Some, but not enough. Unless someone thinks there are enough. I can see it now: the reports from a public surveillance system based on this tech, and they read "why I oughtta ruffin ma back geez oh man what are they thinking huh errr." Just jibber jabber.

I saw some XREAL glasses yesterday on Linus Tech Tips, and that technology has come a LONG way. For a moment I was tempted to splurge on the Apple Vision Pro. Wow, those are cool. You can place yourself or a friend into a scene, teleport, be so immersed. The price tag is mighty immersive too. Ex-pen-sive, with a capital EX. And that's all strapped to your face, which isn't something you walk around with while augmenting prices at the grocery store or showing street names and directions on a bus. The XREAL glasses are something else. They look like a tiny bit thicker "They Live" specs, and the compute is external, on a thing that looks like a power brick. The design is a win, and the product demo on Linus was amazing. Like I said, this tech has really taken off, augmenting imagery like that in such a small form factor. So cool. And not multiple thousands of dollars. OK, maybe one or two of those thousands. It's getting there.

I mention XREAL because not long ago smart glasses were thick-framed, easy to spot, and everyone figured you were a creep for wearing them. "Keep your eyes on the road, pal." Keep your AI out of my eye. Cameras are something most people consider TMI. With the newer versions you could get away with not being profiled as a creepo. Still, keep them out of the change rooms, y'all, and out of sensitive areas. Why are you lurking around the playground anyway!?

The glasses in the video must pack some serious tech, and it's amazing they can do this. Radar? Local AI? Reading not lips but very subtle muscle movements, face twitches? I don't know, but still, amazing. Props to these folks for thinking outside the box. It's just something, you know? Or don't. Different.

Anyway, enough role play. Still interesting. Great... interesting, my great idea.

I'm sure an idea like this will either get built into things like XREAL's tech or get bought by some tech company, either to silence the idea or fold it into their own. I'm sure we'll see something like this soon in smart glasses and such. I'm sure teachers will be excited.

So what is this thing, actually?

Turns out the radar guess wasn't far off. The glasses are from Subvocal (The Subvocal Company), founded by Aarush Yadav and Arnav Kumar. The company says the glasses send radar signals across your face, and when you silently form words, the tiny movements in your speech muscles change how those signals bounce back. An AI model turns that into text. There's no camera, and it isn't reading your lips in the usual sense. You still have to deliberately mouth the words, just as subtly as you like.

The headline claim is 200+ words per minute, which is faster than most people type. On October 30, 2026 they announced a $1.3M pre-seed round led by Afore Capital with Y Combinator participating, and opened $499 Early Bird preorders, with first shipments estimated for Q1 2027. The intended use is writing messages, taking notes, and talking to AI assistants without making a sound. The company also says it only listens when you choose to articulate something, rather than constantly transcribing. All of that is the company's own claim for now. I haven't seen any independent testing.

This isn't the first silent-speech glasses idea, either. Cornell's SciFi Lab built EchoSpeech in 2023, which uses tiny speakers and microphones on the frame to bounce inaudible sound off your face, a sonar approach. It recognized up to 31 silent commands with roughly 95 percent accuracy in testing, and the researchers pointed out it could help people who can't vocalize. Subvocal's radar approach looks like the next step toward free-flowing text instead of a short command list.

Resources

• The original short from ObaNewsNow
• Subvocal, official site
• Digital Trends: Subvocal's glasses promise to let you talk to your computer in complete silence
• FourWeekMBA: Subvocal raises $1.3M for silent-speech glasses at $499
• Cornell Chronicle: AI-equipped eyeglasses can read silent speech (EchoSpeech)
• Neuroscience News: AI-equipped eyeglasses read silent speech