The Future of I/O

As computers get faster—and AI grows more capable—we’re hitting an unexpected bottleneck: ourselves.

Or more specifically, our interfaces.

Keyboards, mice, and even touchscreens were designed for slower, more deliberate interactions. They’re great for control, but terrible for speed. If the future of AI is real-time, context-rich, and deeply personalized, then the future of input/output (I/O) needs to be just as fluid—ideally moving at the speed of thought.

Neuralink and other brain-computer interfaces are aiming for that ideal. But we’re probably years away from widespread adoption. In the meantime, we’re already seeing early signs of that shift: voice assistants, wearable sensors, smart environments, and context-aware tools. We’re becoming just a little more cybernetic—fusing brain and machine, one small upgrade at a time.

One of the most fascinating examples of this in-between state is AlterEgo, a project from MIT’s Media Lab. It’s a wearable device that rests on your jaw and picks up subvocalizations—the tiny neuromuscular signals sent to your vocal cords when you internally verbalize words, even if you don’t speak them aloud. It’s not reading your thoughts, but it comes close to turning silent speech into action. No voice. No typing. Just intention translated into interface. It’s a glimpse into a future where we can communicate with machines almost invisibly—and with far less friction.

Even on a smaller scale, I’ve been noticing how my own preferences shift depending on context:

I speak instead of type when I want speed.

I read instead of listen when I want comprehension and retention.

But when I’m on the go—biking, walking, or in motion—I’ll often use something like NotebookLM or a voice interface to help me get through long reports or documents. Even though audio is slower, it’s available. And that makes it powerful.

This isn’t just about convenience. It’s also about comfort. I’ve dealt with RSI on and off for years. I’m obsessed with ergonomics now: I use a Kinesis keyboard, a vertical mouse with a thumb roller, and I constantly tweak my setup to avoid triggering issues. When it flares up, it’s debilitating. It’s not just that typing hurts—it’s that interacting with a device at all suddenly feels foreign. Disconnected.

That feeling—the loss of fluency—is part of what keeps me thinking about better I/O. I love being out in nature. I love spending time with people. But when I am working, I want the connection with my computer to be as lossless as possible. I want to think, and have the machine understand.

So maybe I/O isn’t just about speed. Maybe it’s multidimensional:

• Speed: How fast can you transmit or receive the info?
• Fidelity: How rich is the transfer—how much nuance survives?
• Availability: Can you use it now, in your current environment?

The best systems don’t just optimize one axis—they adapt across all three. And maybe more?

The future of I/O is about designing interfaces that fit us. That meet us in motion, respect how we think, and let us shift fluidly across modalities. The goal isn’t just faster machines. It’s tighter loops between thought and action.