META’S New Insane AI SAM SHOCKS The Entire Industry! (FINALLY ANNOUNCED!)

So the AI race continues to heat up with Meta releasing their new software, which is Segment Anything Model. This is sam. It's a step towards the first foundation model for image segmentation. Essentially what this means is you can actually have an image and then essentially segment this image into different parts. Now, at first you might think, why is this groundbreaking information? Why is this segmenting no different than Photoshop? But. Well, I'm about to show you exactly why. This kind of segmenting is absolutely game changing for the entire AI industry. And this is about to honestly change a lot of softwares, especially those robots that are coming in. So let's take a quick look at Photoshop quickly. The reason I'm in Photoshop is because when I go back to SAM and I demo the tool, you're going to realize exactly why this is crazy. So this is Photoshop, okay? And what I'm hovering over is the object Selection tool. Now you might initially think that this is going to be quite similar to sam, but later on in the video I'll show you exactly why SAM completely blows out any image. I identification software out of the water with its image recognition software. So this is the website page that you see once you go to the announcement page and you can see right here it says Segment Anything Model. The acronym is sam, a new model from Meta AI that can cut out any object in any image with a single click. The only problem is I'm not sure why Meta decided that they were going to go with this as their headline, because this is not what the software actually does. It's far more superior than that. Only when you scroll down you realize the true capabilities. For example, take a look at this video here that's on the homepage. I've opened it in a new tab so you can easily understand exactly what I'm talking about. The real application of this software is the ability to use this in real time. You can see that right here. This is someone who's going and probably visiting a relative. But what's particularly interesting is they're able to ID everything in the their current environment. If you don't know, this is a form of augmented reality. And augmented reality is when you blur real physical reality with stuff like this, so you can see that the person clearly has some sort of glasses on. This is not VR, don't confuse it with VR, because VR is something that is completely different. And VR transports you to an entirely new world. And you essentially wear something on your head where you can't see the outside world, although it just puts you in A digital world that mimics 3D space. So as I was saying, the reason this is so cool is because this is real time. This is something that you can use to identify things in real time. And you can see right here it shows small IDs. You can see that it says right here, this is a stove. You can see that it's IDing it. And what's also interesting is that it is making accurate calculations as well. You can also see it says right here, 1.1 meters. This is truly game changing. I mean, this is honestly something I didn't think we would see for quite some time. Especially with the rise of large language models. These visual models are honestly truly insane. So you can see right here that it literally shows a cabinet which is, you know, 3.9 meters away. It then shows a television set over there. Then of course, it shows a book, it shows this, it shows that it is honestly breathtaking. Like how it can easily identify exactly what's in our environment. Now some of you might be wondering, so where's the real application for this? There's two main applications that I do think that this does have. Number one, of course this is going to be useful for ar. Maybe you're exploring an environment in which you're coming across foreign subjects and foreign objects that you don't know what they are. Potentially hazardous. Maybe you slap on these AR goggles and you're able to identify exactly what it is in your environment. Imagine an armchair was something that you'd never seen before and it was potentially dangerous, a potentially new weapon. It could be immediately programmed into your software, updated in real time, and you're able to identify exactly what that is. Now, of course there's the second one which is of course very interesting. You're able to put this in potentially an AI robot, which we all know they're going to be releasing sometime soon. And then once you're able to easily identify exactly what's in the environment, like this right here, we'll get to that in a moment. You're going to have highly sophisticated robots that are going to be able to easily understand exactly what's in our environment. So for example, you know, there's the multimodal model Palm, which is of course really good. And essentially what we're going to have, guys, is we're going to have robots that are going to be able to identify exactly what's going on in an environment and that's going to make multimodal robots, which of course, like I said before, are coming very, very soon. Much Sooner than you think. It's going to make their ability to distinguish what is what very, very accurate. And like I was talking about, Palm is obviously a multimodal model that is very good at this. This is definitely going to be interesting if Meta can somehow combine these technology with those robots. So another feature that this model does take into account is identifying multiple things. As you can see right here, the user does prompt the word cat and of course we are going to demo this tool to show you how it actually works. And we're going to upload some images to the software as well. So the user inputs the word cat and you can see immediately it identifies every single cat in the area. Now, I'm not going to lie to you guys, this is something that is pretty scary because this gives me flashbacks to many movies where, you know, it identifies certain things and it immediately highlights them as red or as bad and it then goes on to eliminate them. Now, of course I'm not trying to spread fear and of course talk about a dystopian future because of course this is good. Technological progress is definitely good for humanity. But there is that voice in the back your head that thinks if this is ever applied to a software that is potentially dangerous, you can just imagine something that wants to eliminate a certain object or person. It could get very dangerous very quick. Now, as I was talking about before, this is something that is very, very interesting and it's very, very game changing in the sense that it is real time. Now you can see right here that this software is detecting every image object in the environment. This is honestly, truly, truly crazy. It's literally detecting a dryer, it's going over here, it's detecting a dryer door, it's detecting a book, a pillow. It's really, really game changing to see just how quickly this AI software is going on and exactly what it's able to see. I mean, honestly, the future is truly, truly moving at a very rapid pace. It feels like only six months ago we just got ChatGPT and it seems like now we're getting these crazy tools in which we're able to literally identify everything in our environments. Now of course this is definitely going to be something that's more useful to robots than humans, but it just goes to show how crazy this is going to be. Now this is a short video that showcases some of the abilities you can see right here that you're able to identify different kinds of fruit. And like I said, this goes to show exactly what you can do with this because of course the maybe you're someone that doesn't know which kinds of fruit are inside the box. Maybe you don't know what exactly you want to get. But combining this with different pieces of technology is going to be very, very interesting. Imagine you're trying to sort something that's very complicated. Imagine you're trying to find a rare mineral in a cave. Imagine you're trying to find a specific poster or find specific clues. Maybe police officers. There's literally a million different things. Maybe you're a mechanic and you need to identify a certain fault. There are many different ways that this can actually be used. And I think that this is definitely something that is 100% going to change the way we work. I mean, AR was never something that was truly seen as disruptive until we start to get technologies like this, which can do real time stuff, which really do impact how we work and how we identify different things. Now, of course, like I was saying, AR VR is definitely very, very interesting. And of course, like I said, I was literally just reading the article. But you can see right here, this is someone, maybe they're forgetful. And you can see that right here it says Rex was fed 20 minutes ago. So combining this technology with AR goggles, maybe this is where Meta's big play is going to come in. Because they've already developed the Oculus headset, the Meta Quest, which is by far the most popular VR headset. When they do develop an AR headset, which I'm pretty sure that they're going to unveil sometime later this year or maybe next year, it's going to be pretty insane. Because if it does have these capabilities, you can where you're able to identify certain objects and tie them to your specific reality. Maybe you have a dog and you say, okay, this is my dog's bowl. And Sam is able to learn that you can easily, easily integrate stuff like this, where it says Rex was fed 20 minutes ago and you're able to successfully remember exactly what's going on. Maybe you forget to turn off the stove when you leave the house. Maybe you forget to turn off the lights. Maybe sometimes you forget where your keys are. And it's able to put a big red arrow in your environment to say your keys are right here. I mean, that would be something that would definitely be helpful. You can also see right here on the left hand side. You know, it also shows us exactly how this can be used to cook softer. Like I said, it's going to make workers more effective, it's going to make those in the workforce in particular industries, in particular jobs, much more effective. Because if you're following a tutorial, everybody knows that you have your phone there. Sometimes you get dirt on your phone, you have to keep pausing it. It's honestly just quite flexible, frustrating. But something like this, where you're able to watch the tutorial exactly with your glasses, is going to change how we use things. And honestly, you can also see right here, like I said, sorting items based on this is definitely going to be something that it does. Now there are computers that do do this already in a very sophisticated manner, but this is going to be something that is going to be in a very different asset. And like I said, it's able to, you know, you know, help biologists, maybe they want to identify rare animals or animals with specific injuries. This is truly, truly interesting. This is kind of like an AR AI thing. And I truly believe that this is going to be one of those game changing products which definitely gets applied to a multitude of applications if Facebook wants that. Now what's crazy about this is this was trained on 11 million images and had generated over a billion masks. So it was truly, truly a colossal task. But so far I do think that the final output is truly insane. Now we also do have a demo here which you can just simply select certain homages. As you guys can see right here, I can select this dog. I can also select this dog too. And then of course I can add the stick as well. So what's also cool is that in the demo you're actually able to upload certain images to try it out yourself. And you can see right here you have to accept the terms and conditions, something that I've already done. But they give you a variety of images right here. What you can also do is you can also upload an image. You can to potentially see exactly how this image works. So of course, right here, let's see how this image works. You can see right here that we can literally hover over completely everything. And so you can see right here in the demo that we're using, this is a certain image. Of course, Sam is running inside the browser. You can see that right here. I can simply select the background and of course it highlights every single corner. Now this is something that is really, really accurate. Of course we can undo this. What we can also do is we can also add or remove areas. And, and if we want to remove certain areas, we can simply add a mask to here. And for example, let's say we wanted to add this food right here and we want to add These beans. And then we wanted to add this right here and just add that. We can also. Let's say we wanted to remove the beans, we could. Let's say we wanted to move the grass, we could simply remove the grass by clicking Remove area and then clicking Remove the grass. I'm pretty sure that's not grass. I don't know what that is. Maybe I need Sam myself to identify exactly what those vegetables are. Yeah, you can see right here as well. This is where it identifies absolutely everything in the image. Right here on the left hand side we have a tool which simply says identify every single object in the image. And you can see it even, which is just truly insane. Identifies every single pea or whatever, you know, beans that is, whatever kind of beans that is. It literally identifies every single one. And I'm honestly, honestly shocked about this because this is truly, truly crazy. So of course here is where you can, you know, cut out different things. And I mean, if we go into Photoshop, you can already see that Photoshop is, does have a little bit of the same kind of characteristics, but it isn't on that level where Sam is. And I know you might be thinking that's the image that they gave it. Guys, if I find an image from Google and go head to head with Photoshop, I promise you it's not going to be as good. So this just goes to show exactly how crazy this is. And the only reason I do compare it to Photoshop is because I want to show you guys how far ahead Facebook is. So, yeah, Facebook has done a really good job. So here I've uploaded a certain image which is of the Boston Dynamics robot. But and the reason I've uploaded this is because I wanted to see if you give Sam an image that isn't a direct image from their library, is it going to still be able to do it? And so far, I mean, it seems to identify absolutely everything that's in the image. So it says, prompting Sam finding everything duplicates. And I guess that's how they identify everything in the image. And this is honestly really shocking because it just goes to show exactly what they identify as certain parts. And yeah, I'm pretty sure this has got it right. So I'm pretty sure if we hover and click, I want to see if we can get the entire robot. And. And there we go. You can see instantly we can get the entire robot. Maybe if I want to remove one of its legs, can I do that? I wonder if we can. And you can see right there, it only sticks on the leg. Oh. So I'm guessing that it just sticks on the leg right there. I'm just getting used to the software, but I'm not gonna lie to you, this is very, very impressive. Okay. Not because of this capability, but. But because of this capability. Because I know that once certain robots start to be deployed and want certain AR technologies are deployed, this kind of technology is truly going to change how we interact with the real world. Now, you can see right here that on this article from the Verge, it says that aside from the Quest lineup, which is Meta's VR headset, Meta also has thousands of employees building future AR glasses and wrist devices for controlling them. So the key difference from VR is that the company intends for augmented reality glasses to eventually be worn throughout the day as a replacement for smartphones. And Zuckerberg has called this the. The holy grail device that will define our relationship with technology by the end of this decade. Now, it's important to note that Meta does have some smart glasses at the moment, but these aren't on the level as to what we have with this new technology that is being released. It's essentially just some glasses that you can just record stuff. I mean, to be honest with you, this is definitely a little bit creepy if you do ask me, because you're never going to know if someone is recording. You're just going to think that they're recording only if they have their phone up. But with this, they essentially have a secret camera right here that literally could be recording anything you do. So that, for me, is a little bit creepy if you ask me. But what do you think about this software? Is this going to be the future? Is this going to be something that you use on a daily basis? I know that if this does come to the market, I will definitely be acquiring one of these because I feel like certain technologies, if you don't take advantage of them, for example, the iPhone, the Internet, or whatever it may be, you may eventually get left behind. But it will be interesting to see how this stuff changes over the future.