alterego

In @Nature, our CEO @arnavkapur discusses Alterego's implications: restoring voice for those who’ve lost it, and advancing a new general-purpose interface for human-AI communication. nature.com/articles/d4158…

Telepathic voice chat with AI is going to be wild. Go Arnav and team!

In a year since I first met Arnav Kapur, his device, AlterEgo, went from a janky prototype to simulating telepathy. Arnav has been an inspiration to me since I'd first heard of him in 7yrs ago and I want to share his story.: I first heard of Arnav from his viral ~1M hits MIT Media Lab video from 2017. It harkened back to a similar talk from the same group by another Indian in 2019, Pranav Mistry, one of the major reasons I (and many others) decided to study in the US. I met him through his younger brother, who I'd gotten to know over the years (also a genius in his own right). I immediately asked him "Why didn't you commercialize this? Was this just a flashy academic demo?" and he said "I'm doing that now. With LLMs, I think it's finally time." But the first demo was janky. It took 15 mins to set up. Electrodes needed to be taped to your cheeks. It worked maybe 80% of the time. Calibration for new users took even longer. And it could only say 4000 sentences. Yet, it still felt like magic. I pointed at one of the sentences for his colleague Scott to say silently. In seconds, barely moving a muscle, I'd get a text of that sentence to my phone. I couldn't see his mouth move at all! But was silent speech really possible? Hardware is hard. There were too many sensors for it to be usable. It didn't work without constant touch to your skin (a beard was an issue). Too much head movement was an issue. The battery was external. Calibration was too much friction. And you could only say a few sentences! Fast forward a year and you have what resembles telepathy. The sensors shrank. The device resembled open-ear headphones. No 15 min set up. And you can say most words. Alterego reads volumetric fine-grained neuromuscular signals in multiple languages and translates them into text. It allows you to understand what the user's saying with no noise while they barely move their mouth. Hardware is still hard. And there's a long way to go before this is end-user usable. But it takes some serious courage and exceptional engineering skills to dedicate over a decade of your life to building hardware that resembles the future. We need more engineers like Arnav and his team.

We’re excited to finally share this with you. Follow along, we’ll be revealing more soon. Welcome to the future of communication. Sign up for early access alterego.io

This is just the beginning. Imagine a world where AI and computing are woven as an extension of the mind and disappear into the background. That’s the future Alterego is building. 🌍✨

Introducing Alterego: the world’s first near-telepathic wearable that enables silent communication at the speed of thought. Alterego makes AI an extension of the human mind. We’ve made several breakthroughs since our work started at MIT. We’re announcing those today.