Sign up to try The Blind Computer. Join the early access list.

The Blind Computer

The Best Computer in the World Wasn't Built For Everyone

The tech industry is racing to build the ultimate AI computer. Almost nobody is asking what a computer should be for someone who can't see the screen.

Every few months, the same scene plays out. A famous chip company and a famous software company share a stage. They unveil the most powerful personal computer ever made. They call it an AI PC. It runs smart assistants right on the device. It watches your screen, reads your apps, and guesses your next move before you make it.

The demos are dazzling. The hardware is real.

And every time, I think the same thing. This is the best computer in the world, and my mother still can't use it.

She's sharp, and she's not afraid of technology. She's blind. And every one of these amazing machines was designed, from the chip all the way up to the apps, by people who quietly assumed you'd be looking at the screen.

Why the screen is the whole problem

Here's something most sighted people never have to think about. A modern computer is, at its heart, a picture.

Think about what's actually on a screen. Icons sitting in a grid. Windows stacked on top of each other. Menus that drop down when you click. All of it assumes two things: eyes to scan the picture, and a hand to steer a pointer toward whatever those eyes find. Every app you've ever used is built on top of that picture. The picture isn't decoration. For these computers, the picture is the entire point.

So how does a blind person use one? Through a clever workaround called a screen reader.

A screen reader is software that crawls across the screen and reads it out loud, one piece at a time. Button. Checkbox. Link. Heading. It's an impressive invention, and the people who build and depend on these tools have earned real respect.

But look closely at what's really happening here. We built a computer for eyes. Then we strapped on a tool that describes that eye-computer, out loud, to someone who can't see it. The blind person isn't really using the computer. They're listening to a running description of a computer that was never meant for them. And they pay for that translation with friction on every single task.

Picture it. A button has no label, so the screen reader just says "button," and she has no idea what it does. Her banking app pushes an update, the layout shifts, and overnight a bill she's paid for years is suddenly out of reach. A website gets a flashy redesign, and the description falls apart. She lives one careless design choice away from being locked out of ordinary things, every day.

The new AI PCs don't fix any of this. They make the computer more powerful for people who can see. The blind user is left exactly where they were. A faster picture is still a picture.

Nobody has ever built the other kind of computer

Here's the part that should bother more people than it does. In the entire history of personal computing, no one has shipped an operating system built from the ground up for people who can't see.

Sit with how strange that is. We have computers tuned for gamers. For giant company servers. For cars, for watches, for store checkout screens. We have decades of accessibility add-ons, many of them genuinely good, won by advocates who fought for every inch.

But an add-on is something you bolt onto a finished house. What has never existed is a house designed, from the foundation up, for someone who gets around by sound and touch. A computer where you talk and it talks back, where the main way in is a conversation instead of a screen. A computer that always knows where you are and what you can do next, and just tells you, because it was built to know.

It's never reached the market. The reasons are familiar. It's hard. The audience was written off as too small to bother with. And the lazy path of patching the visual computer and bolting on a screen reader was always sitting right there. So that's the road the industry took, over and over, for forty years.

Ten years ago, when I first became a programmer, I told my mother what I actually wanted to build. Not an app. Not a slightly better screen reader. A computer that assumes, from its very first line of code, that the person using it cannot see, and that feels genuinely good to use on those terms. That promise has been my north star ever since. Everything I've built since is a step toward keeping it.

Why we build on Linux

If you're going to build something this fundamental, the first question is what you build it on. The answer is Linux, and the reasons are practical. Here are four, in plain terms.

You can change everything. Linux is open all the way down. No company stands between you and the machine saying "that part is ours, hands off." When your whole goal is to tear out the screen-first assumptions and put a voice-first computer in their place, you need a foundation where nothing is locked. Windows and macOS keep their most important layers sealed shut. Linux hands you the keys to all of them.

It doesn't force a screen on you. This is the one most people don't realize. On a normal computer, the visual desktop is basically mandatory. On Linux, the visual desktop is just one optional program you can leave out. You can run a full, powerful computer with no screen at all, and the system honestly doesn't care whether a monitor is even plugged in. For a blind-first computer, that's everything. We don't have to fight to remove the screen. It was never required in the first place.

You can swap any part. Linux is built like a set of building blocks instead of one sealed box. Don't like how a piece works? Replace just that piece. That's exactly the freedom you need when you're inventing ways of using a computer that have never existed, and you have to try things, throw them out, and try again.

It runs on anything. The same core software that powers the world's biggest data centers also runs on a board the size of a credit card. That range matters enormously for where this is headed, which I'll get to in a second.

Put those together and you get a foundation that is open, flexible, screen-optional, and able to scale from a tiny gadget up to a serious machine. Nothing else on Earth gives you all four.

What we've actually built, told straight

I'll be honest about where things stand, because the space between a big vision and a working product is where most accessibility promises quietly die.

What we run today is a hybrid system. Several computers work together, each handling the part it does best. You talk to a voice layer. Behind it, a reasoning system figures out what you mean. Behind that, real services get the actual work done. You speak. The system understands, acts, and speaks back. No screen. No screen reader describing a picture you can't see. The conversation is the computer.

It's a hybrid for a reason. The heaviest thinking happens on powerful machines in the cloud, because that's where the strongest AI lives right now. Other parts run closer to you. The pieces stay in step with each other behind the scenes.

And it's real. It works. It already handles real phone calls and gets real tasks done. It's early, and I won't pretend the seams are invisible or that we've arrived, because we haven't. We're at the honest beginning of something that already works, and it gets better every week.

The hybrid is the right call for right now. It lets us learn what a voice-first computer truly needs by putting a working one in real hands, instead of hiding in a lab for five years and shipping something nobody wants.

The real destination is a device you can hold

The hybrid system is a bridge. The destination sits on the other side of it.

Capable AI is moving off distant servers and onto devices you can hold, faster than almost anyone predicted. As that happens, the goal comes into focus. One device. A single piece of hardware you can hold, running a computer built from the ground up around sound, speech, and touch. No screen to translate. No far-off systems that have to stay in sync. No depending on whether some app maker remembered to label a button. You hold it, you talk to it, it talks back. Whole and complete. Made for your hands and your ears from the silicon up.

This is why the foundation matters so much. The same open, screen-optional core that runs our hybrid system today is the same core that will run on a device the size of a deck of cards tomorrow. Getting from here to there isn't a do-over. It's a steady walk along a foundation we chose, on purpose, to make that walk possible.

That's the machine I described to my mother ten years ago. A computer that's hers from the very first line of code.

The best computer in the world

The industry will keep building more powerful PCs. The chips will get faster, the assistants smarter, the demos shinier. Good. I'm genuinely glad it's happening.

But "the best computer in the world" should mean more than the most powerful one. It should mean the one that finally welcomes in the people every earlier generation of computing left standing at the door. For the millions of people who live without sight, the best computer isn't the one with the biggest chip. It's the first one actually built for them.

We're building it. Out in the open. One honest step at a time.

If that's a future you want to see, whether you're blind, you love someone who is, or you just believe a computer should belong to everyone who reaches for it, get on the list. We're opening early access to the people who want in from the start, and they'll be the first to put their hands on it.

Ten years ago this was a promise to my mother. We intend to keep it.