This A.I. can literally read your mind using brain scan!

hey guys the project here so despite all the craziness that's going on in the AI world today when I found out yesterday about this project this obscure little project oh my God you see I'm actually very much into AI I've been building computers since I was 10 years old now I am interested in AI I think everybody is but I'm especially interested in AI because the human brain more specifically can wear imagination and prediction happens which is our frontal cortex is great at predicting linear change it's evolved to the point where it's able to predict intuitively anything that's linear change but technology has so far outpaced us that we are now dealing with exponential change which our brains are completely incapable of dealing with at an intuitive level I'll give you an example if you have a unlimited large sheet of printing paper if you fold that that's two sheets and before that again that's height of four sheets if I fold that 100 times how thick would that layer be just give me an answer intuitively don't calculate okay so some people may say okay it might reach as high as the ceiling of its house some may say no no it's going to reach to the top of Empire State Building some some may say well you might even reach to the moon and uh here is roughly the real answer if you look at our solar system is mostly empty and if our solar system is to scale the Earth build a tiny grain of sand right here the sun will be like a tennis ball all the way over here so our solar system is actually quite large and uh our galaxy even more so there's vast spacing between stars and then between galaxies even more so by the way light takes one seconds to go around the entire Earth six times but for the light Traverse our entire galaxy gonna take a hundred thousand years if you talk about the observable universe there are 200 billion galaxies that we can see that paper folded a hundred times actually about 102 times will reach the entire length of the observable universe so we're really bad at intuitively understanding exponential change but and and here's the problem AI is all about exponential change change and that's where the progress of AI is there right now we are at the beginning where the exponential change is about to go up every day you wake up and there's something new about it the truth is our economic system and our legal system is not equipped to handle any of this our legal system can barely handle anything that is that's at a change at a slow pace linear change you can barely handle that no imagine exponential change pretty soon videos and audios are just as credible as a crappy image in the era of photoshops our economic system is not be able to handle that because if you think about it logically if a company or a manufacturing facility and or you know Supermarket replace most of its human workers with machines and they can do everything more accurately honestly and more efficiently but the amount of value and amount of revenue and profit those companies generate will be the same even if humans are replaced by machines which means that humans should therefore in theory be having more for doing less we would more time and we'll be able to receive just as many if not more and better goods and services That's The Logical conclusion except humans are very greedy and we are very power hungry our spiritual maturity is nowhere near our intellectually wise speaking so we're not wise but we're smart that's not a good combination at all so that's already we're already gonna face some crazy change doesn't come out in the next few years but yesterday I discovered some new development that even that I'm aware of all this is going to happen that the next few years is probably gonna be really crazy knowing that all that and knowing the next few years can be crazy and our economic and legal system is gonna they will not know how to handle all of this what is this thing came out that scared the crap out of me well I'm just going to show you on the screen okay so you guys must have some understanding of AI generative art okay this is based on a model called the stable diffusion on the screen you can see a experiment in research by a Japanese group of AI scientists and what they did is that they used a stable diffusion technology which in case let me just give you like a really really simple version of how that works basically imagine a colored balls right in a box and when you have a very limited number of balls and you check the box they will spread out kind of evenly but not very clearly that's a fuzzy image that's very noised image and what AI is able to do because it's so good at pattern recognition and the process of diffusion which means as you shake the Box the number of pixels or balls in any area will be roughly the same right so when you do that and AI can predict that's going to happen and I can insert its own basically pixels and or color the balls into this box and the more you insert in there the clearer the images gets but AI does something else on top of that it it can you can modify the variable variable of how each particles move so that it is closer to the patterns it's already learned through the billions and billions of images you learn online so when the entire canvas so to speak is filled 100 with these pixels now or the colored balls you end up getting pictures that have a pattern to them now the scary part is this group of Japanese scientists is able to use this on a human brain's MRI scan and they basically look at how the human brain lights up how the how the the neural activities on specific part of the brains move and work and they basically is able to extrapolate extract a fatty image a fuzzy a verified image it's like if you imagine a thing right it's not picture clear it's it's kind of weird and fuzzy in your head they were able to extract that out use this technology to apply the diffusion technology to it and then it creates an image that's almost identical to what the viewer is actually looking at look at this part here so you have the test subjects looking at the picture of a teddy bear and then the bottom is what the freaking AI looking at the MRI scan of the brain and is able to generate an image and you can see if teddy bear looks like a teddy bear the house it looks there's no house there but you can tell there's something bright in the the road the trees in there and this is only at the beginning which means this is not just an attempt at reading your mind they are reading your mind because what you see what you imagine what you remember this process can be applied to it's just right now first iteration the accuracy is not 100 yet soon they'll be able to read not just pictures they'll be able to read words giving audio because they are all just different patterns of information and when this model is perfected and it gets into the wrong hand or being used in the wrong way that means people can literally read your mind and your memories guys stop that scene Science Fiction it's pretty much here remember everything is improving and expect your rate which means in an hour a self you know learning AI can run iterations a billion times to improve at doing certain tasks so something that takes a human to improve for like a thousand years and AI can do it in like three seconds so imagine what happens in a month in three months in a year imagine what happens if all these are combined to be used in a bad way this is so far above everything else in terms of its potential impact all right just want to talk to you about that alright see ya