If you hooked it up to me, most of the time it would read "Why in the fuck would you want to name your kid Gilligan?" This sounds like complete horseshit. We haven't cured the common cold, but the astro-scientists of Apple invented a mind-reading device? Fuck off.
The idea of something like this existing is one of my worst nightmares. As a general rule I dislike most new people, usually before I meet them. By the time it gets to the point where I am actually being introduced to someone I'm usually thinking of ways to exit the room. If I'm unlucky enough not to have figured out a way to escape then instead of paying attention during the usual meet and greet I'm probably thinking of the most inappropriate thing I could say or do right now even although I have no intention whatsoever of saying it. The worst part is I do this even when I don't immediately take a dislike to someone, and worse when I'm genuinely interested in whoever I'm talking to. I also do it during meetings at work, presentations I go to and lectures I give. It's like I'm daring myself to say something as offensive as possible at the most inappropriate moment possible- for fun.
The way it works isn't new or anything. There have been devices that do basically the same thing (although not as well) for ages. They use them for things like mice for disabled people so they can move a cursor around a screen by thinking really hard about left or right. Trying to actually read thoughts is astonishingly complex. It's basically a cryptography problem. Picking up the electrical impulses of our synapses firing even at amazingly high accuracy doesn't actually tell us much. It just gives us an encrypted data stream. Trying to turn that data stream into words or images requires building a huge library of data where test subjects tell us what they were thinking about and we look at the recorded data to try and find that particular thought in the noise, and isolate it to a pattern. But people don't think like a telegram. We shift context amazingly rapidly - and thinking about tits for a second in the middle of your concentrate on moving left while we record what that pattern looks like experiment adds huge noise and bias into the sample. Add to that the fact that we don't really share common thought patterns for complex concepts and it gets even harder. The way an american christian's brain processes left and right is probably pretty similar to the way a russian jew's brain processes left and right. But the way you conceptualize say Lasagne? Probably wildly different from the way your girlfriend processes lasagne. Colors, tastes, smells, sounds, all processed differently based on physical capabilities and changes. The more complex a concept is, the less likely your brain pattern for that concept even vaguely resembles another person's brain pattern for that concept. With a large enough library of high enough quality data, this kind of tech absolutely can read your mind. But actually getting there is probably years (maybe even decades) away, unless someone has an apple tree / gravity style epiphany that changes how we approach this subject dramatically. I'm a big enthusiast about this kind of tech and I desperately, desperately want it to be on the market and real. But sad reality is that it's a big data / processing analytics / cryptography problem that needs a shitload of data that's hard to collect and then a very expensive fuckload of infrastructure and time to crack. And the actual math that the infrastructure needs to execute to get the data quality usable and then figure out what the data means, isn't anything like simple. Also, applications of this tech aren't going to be like Professor X reading your brain. It'll be more like a way better keyboard and mouse. Instead of typing and moving a cursor around with a device or waving your hands/talking/whatever - you'll just think about shit, and it'll happen. But that alone is an amazing idea. What if being an artist didn't require hand/eye coordination or an understanding of drawing technique. What if it meant being able to imagine something really clearly, and the pictures you could imagine just showed up on screen? What if you could record your dreams and play them back later? The really awesome stuff happens from the follow on. If we figure out how to read a brain - the next step is how to write it. And then we get to learn shit like the matrix. You want to learn to speak french? Download the e-learning course to your fucking brain and bam, you speak french. Kung Fu? Piloting a helicopter? How to tie a cherry stem in a knot with your tongue? All available from an online store with direct download. Shit, what if you could plug a bunch of ram into your brain so that you now have a didactic memory? If we can read and write the brain - then the hardware to give everyone perfect and almost instant recall of EVERYTHING your brain is aware of after the tech goes online is trivial. This shit is really on the tech horizon as an actual thing that will happen. What will schools look like when the knowledge for any career path is just a download away? When native talent and capability is all that matters. Imagine how story telling will change when you can record, edit and share fucking feelings? Christ, imagine porn. How will human cloning change when you can back up every part of your brain and reload it into your own clone. Fuck a pension - I want a rejuvenation fund so that when I get old, I can just reboot and wake up 19 again. This technology is a foundation stone for human immortality. Not in a bullshit sci fi way - in a 'holy fuck, this is real' way. And the major benchmarks between here and there will change the human experience as profoundly as any other leap forward in human history. Like learning to communicate or use tools, this will change everything. There's a fair argument for the idea that we'll no longer be Homo Sapiens when this technology actually reaches its potential. We'll be something else entirely. The question really is just if it'll be the people reading this that live to see it happen, or if it'll be our children. Or if we nuke ourselves out of existence before we get there at all.
One of the biggest stumbling blocks to brain mapping right now is the sheer amount of data required to get the resolution down to the synapse level. Most of the current studies involve a (relatively) high level mapping of electrical current, which is fine for extremely generic tasks like cursor movement, but if we're going to get down to the level of actually reading minds or, better yet, doing the reverse - inserting data back into the brain - the imaging has to be significantly more complex. The instruments are actually available right now to do the mapping down to the synapse level, but it's incredibly time and data intensive. I was listening to an interview with a guy whose lab is doing this kind of work. He said their imaging tools currently generate about 1,000 terabytes - a million gigabytes - of data per cubic millimeter of brain. So the challenge is not just acquiring it, which is difficult enough, or storing it long term, which is even more problematic - it's actually being able to manipulate and compute that much data. Additionally, the process is slow right now. I can't remember exactly how long he said it takes them to scan a cubic millimeter, but I think he said they had it down to about a year or so. It's incredibly cool science, but we need to make some significant advancements in our ability to acquire and manipulate truly staggering amounts of data before we'll get to a real level of understanding.
I have no doubt that eventually it can be achieved if we don't all kill each other first, in the 1940's a computer the size of a living room had the power of a pocket calculator. But we are also no doubt a long way off. What Apple is doing here is just dumb to lay a claim like that. It's ean easy con. Look, I'll read your mind: I'll bet you feel sleepy when you wake up in the morning. SHAZAM!!!! Your mind is blown by this cosmic sorcerer.
The other problem with this kind of thing is medical imaging retention requirements. Most first world places have laws that you can't just take a medical image and ditch it. You have to keep it for X number of years and be able to get at it within Y hours at any given point in time for medical emergencies in the future that relate to that information or whatever. A super high resolution scanner for this kind of task might generate a PB per cubic millimeter from 5 minutes of observation or something. But the practical requirement for research is only for a tiny fraction of that data. My current day job involves big data. Not brain scanning, but very high resolution 3d mapping of large physical networks and then analytics over that. Our planning scale is for a petabyte in the next 12 months and 10 petabytes in the next 5 years so we've been talking to some other equivalent sort of projects that includes neural analytics stuff. The lab we've actually talked too about their storage and processing tech is working on neural interfaces for digital speech boxes. You think a word and the box talks like stephen hawking but without the wiggling a finger on board. It's not quite as in depth as converting your imagination into replayable pornography - but it's pretty intense. They get around the rules by stipulating that no medical advice will ever be given as a result of this work and advising that people have thorough brain health checks before participating in their research - because they can't tell someone that they found a brain tumor or something without then being in breach of the rules that let them dodge the retention requirements.