Steve Hoffman discusses brain-computer interfaces and neural sensing technology, from EEG headsets used in Korean schools for attention training to experimental systems that map brain activity to words and images. He covers applications for ADD and PTSD, and explores speculative approaches to reading dreams and thoughts directly from neural patterns.
Chapters
Okay, we'll dive into the real thing. Okay, mind tech. I'm going to talk now about the future of humanity and technology specifically focused on the brain. So a lot is changing. Our brains have been virtually unchanged for over 100,000 years. So prehistoric cavemen had exactly the same brains that we do today. And if you think about that, that is pretty amazing. Because living in a cave and living in San Francisco or Beijing or London, totally different, right? We, with the exact same hardware, are able now to use cell phones, create rockets to the moon, all of this stuff. What computer— like, we're used to our computers. Our computers are upgrading their hardware like every month, right? It's getting better. The hardware is— there's something new coming out for our computers. Our brains haven't been upgraded in over 100,000 years. But that is actually about to change. We're actually entering a phase now where, you know, we get a new phone every year, or even more often for some people, but we're getting new hardware all the time. In the future, we will be upgrading our bodies and our minds much the way we do machines. And I know this sounds like science fiction, and what I'm going to talk about today— I'm not going to talk— I'm going to talk about, at first, what we're really doing today, and then in the latter part of my talk, I'm going to speculate on where we might be going in the future with this hardware and software. So, this is one of our startups. They're out of Korea. They're called Soso, and they're one of many startups working on what's called BCI. BCI is brain-computer interface. So, what it does is there's a whole range of devices now that you can put on your head, and they will actually measure the activity in your brain. And the standard technology for this is EEG. EEG. Now, EEG is still pretty primitive, but what's coming next will literally blow your mind. There are new technologies, and a lot of them are being developed by startups and the military, that are going to allow us to much more accurately measure and actually communicate with your brain. Now, the nice thing about this early technology is it's noninvasive, meaning they don't have to crack open your skull and put a chip inside. You just put on this device, and these devices, you can— with these devices, you can do simple things. So you can do things like measuring concentration. So this startup and many others They have worked on different applications for this, which I'll talk about. So the startup we know in Korea has gotten it out to hundreds of schools in Korea, and they allow children to play interactive games on their iPad, and as they're playing, they can measure the concentration levels, and they can actually control the games by concentrating on them. And these— this software-hardware combination has proved really valuable for kids with ADD, attention deficit disorder, where they can't focus. So it's training them how to focus, and they can end up learning a lot of things that they wouldn't have learned— that they would have had to take drugs to actually focus at that level. Now, there are other companies using this technology for post-traumatic stress disorder. So people who have suffered, like, in a war or some experience, and they want to have the biofeedback to calm themselves down. They're using this technology experimentally now, as you can see in this picture, for how could you control yourself just by thinking in a game, and what would that mean? So there are going to be many, many more applications. And as I said, the current hardware is still pretty primitive, but the hardware that's coming out right now, literally as we speak, is going to blow open the doors for this. We're going to get to a whole new level. So scientists have also done several experiments on what they call dream reading. And again, it seems like science fiction, So in a primitive way now, they can— what they do is they show you a series of pictures while you're awake before you go to sleep, and they measure the patterns in your brain as you look at these pictures of a dog and a house and all these things. Then when you sleep, they are able to identify if those same patterns come up in your brain And know that you're thinking of that particular image. So you can imagine, if you get a big enough database of these images, you can start to interpret somebody's dreams through the technology. Now, MIT has a technology, and there are a couple startups that are also working on this, that are building mind-reading software. Along the same lines as I told you there. So the mind-reading software is really interesting because it's much simpler. When I— all of us speak to ourselves in our head. That's how we— it's strange that we have to communicate to ourselves by speaking, right? But we will just sit there talking to ourselves. So as we are talking, there's a certain pattern for every word we say. What they're doing is looking at the pattern in the brain, and then they're matching that pattern, they're putting that into a database. And when you're wearing one of these devices, if you think of those words, and the beauty of this is there's, you know, most of the words in the dictionary we don't say. But a subset of these words, actually a small subset, we use all the time. So if you can match the patterns in somebody's brain to the patterns in the database, you can literally start to read their mind. Every time a word comes up, you know it. Oh, that word came up for that individual. Now, will the patterns be identical in every person? Probably not. The information will be stored in slightly different places. But it wouldn't take long for you to train a machine to read your mind. Now, that's just the beginning. They may be able to have algorithms sophisticated enough, and they get enough samples, enough data, that the AI could actually learn over time what the general rules are for different words, and then you could apply it to anybody. So we're at the very beginning of this kind of mind-reading technology.
