so with a crazy week in artificial intelligence let's actually take a look at the newest news stories that have happened this week so one of the craziest things here is that Elon Musk has actually spoken about grock 3 now the craziest thing about this is that grock 3 as you guys know Elon Musk has you know numerous times referred to this as the most powerful AI model and I do wonder if that is going to be the case because since they've been making grock 3 there has of course been the 01 Paradigm and he literally says that you know pre-training is now complete with 10 times more compute than grock 2 so of course grock 2 is actually pretty decent but grock 3 is going to be 10 times more compute so it's a 10 times increasing compute is something that's rather fascinating because usually when you add more compute to the model when you add more data and high quality data the model gets a lot smarter now the thing is that most people haven't realized is that there is actually a very interesting area where Gro could actually dominate most people haven't realized that tra gbt has been dethroned quite a long time ago if you currently look around at the AI space the best model to use is either Claude or actually Gemini in terms of a raw chatboard experience I would say that chat GPT is best because it has multiple features like being able to upload files and such but when we actually take a look at the fact that you know we look at what is going on with these models we can see that grock 3 could actually come into the mix and disrupt things even further this is because they have I think probably the most computer at the moment maybe not more than Google but they certainly have a ridiculous amount of compute for a startup that has only existed for I think around 1 year and 18 months they are definitely behind the ball but they've been moving a lot faster than these other organizations so it's going to be really interesting to see what grock 3 has you know coming out with I think that's going to come out q1 of 2025 and this one is highly anticipated but as many people have said before don't doubt Elon Musk and I remember when I was looking at you know a few tweets between the teams they actually spoke about the fact that like all they do right now is work and that's why they've been able to literally move as fast as they have like they've literally just been working 18 hours a day like six days a week so that's why they were able to catch up with the amount of time that they did now if you're wondering you know how crazy this is going to be well there was an example of someone actually using Gro to diagnose medical injury so you can see right here it says you know Elon Musk says Gro can diagnose medical injuries and then this is talking about a true story of how grock managed to diagnose someone's daughter with a broken wrist last week now if you're wondering how this happens this has been the silent revolution in ai ai models like cat GPT Claude and Gemini and of course now Gro they're pretty good at analyzing images and large pieces of data and essentially what makes these models really good is they're able to summarize data and because they've seen a lot of images in their training data it's easy for them to identify when they know a similar image and that's why these models are excelling at Medical cases so this person talking about you know you know her daughter you know she didn't sleep by 6:00 a. m. and she was convinced that her arm was broken and she says that was in a huge amount of pain and then they went to Urgent Care and they examined her and took some X-rays okay but the doctor who saw her and the radiologist declared her free of breaks and and after answering some of the questions they sent her home with an ace WP and an ibuprofen so they basically said you know what it's not that bad but then on the way out and this is really smart by this person they actually asked him to please print out the x-ray images for us so I could take them to her PCP and she was like I hope they were right but she wasn't solid so her mother's intuition basically were like you know what let me actually just do some more digging so her hand kept going cold and tingly and she couldn't move her thumb and it didn't look right so she had more aches and pains and then she was like you know what I'm going to start doing my own research and she was deciding to you know looking over X-rays of normal wrists and broken wrists and she remembered a post from a few weeks ago by Elon Musk that grock 2 could read medical images and then she decided to you know upload this x-ray to Gro and actually should have had the screenshot here but the long story short was that grock was basically stating that look this is a pretty obvious break and I'm not sure how the how the you know people who looked at your arm didn't see this and you can see that they were like like okay let's go to the primary care physician and you know get a referral to the author and then the next day the risk specialist actually took new x-rays with multiple views and they examined them and they confirmed that grock's diagnosis of a distal radial head fracture with dorsal displacement and they actually didn't tell them about the foray into Gro because they don't want to think you know she's insane but they set the cast and that's it so you know if they didn't do this and if they you know didn't decide to do extra research and had it gone untreated she would have lik needed surgery that she can probably avoid now so this is pretty crazy and I do apologize for my horrible scribbles here but this would have gone untreated at least for a while because the doctor said it was fine and I'm just a mom right so who the hell am I to question their diagnosis so this is something that's pretty crazy because when we take a look at this most people are like you know I'm just a mom like I'm not a person I'm just a mom I'm just you know a person I can't really like know what's going on but this is something that's pretty crazy because doctors and Radiologists we know that these people are overworked you know the healthare system you they could have been you know as she said they were probably tired they could have you know been coming off a long night maybe they were looking for only an obvious clean break or maybe they in need of some CME but both of them missed what grock agreed was an obvious break so this is pretty crazy because I think once again like I said before every time I see a research paper or something along the lines of AI in the medical field I'm not saying it's a silver bullet what I am saying is that so far I've seen some clear clear indications that this is something that is going going to be an area in the future where we definitely get a lot of more exploration and I think the future of healthcare does look like one that is relatively cheap relatively fast and something that is a lot better than we do have now especially at the lower levels when these systems are going to be there now this is AI news but this is like kind of like crazy air news cuz this one went super viral on Twitter so I don't even know how to even you know talk about this but there was a Twitter post that went completely viral on Twitter and I got tagged in this a few times you can see that this tweet had over 1.
3 million views 20,000 likes 7,000 retweets but this was because billionaire Larry Ellison says that AI fuel surveillance systems can ensure citizens will be on their best behavior so I might make an entire video about this but it's pretty crazy because he's basically talking about you know we were going to have supervision he said every police officer is going to be supervised at all times I mean most of them pretty much are right now but if there's a problem AI is going to report that problem and report it to the appropriate person citizens will be on their best behavior okay because we are constantly recording and Reporting everything that's going on and this doesn't sound like the very best future that we could have it's not really what like this isn't really what people think of when we decide to think about you know the future world we don't decide to you know think about the fact that like everyone's going to have to be on their best behavior I think this is you know some references to you know certain novels about certain dyo dystopian Futures but um yeah he like this is pretty crazy okay like I think this is something that is probably the average person's worst nightmare not that because the average person is a criminal but I think people are very you know concerned about their privacy we've known about you know different things that the government has done that have you know constantly invaded individuals privacy so this is something that I think the average person isn't going to want so it's going to be super interesting to see exactly how this you know reaction you can see that this reaction of course you know saying that citizens are going to be on their best behavior we're going to be constantly recording everything um isn't the kind of PR that you know you kind of want for yourself but maybe it was a slip up I don't know what's going on here but this isn't a very good scenario and for a billionaire to be saying this of course billionaires are people with immense power this is not really a good sign because it's just pretty dystopian and he also expects AI drones to replace police cars and high spe chases he says you just have a drone follow the car it's very simple in the age of autonomous drones he did not say if those drones would broadcast chases on you know networking you so I don't really want to cover this story too much because this is like really Doom and Gloom but at the end of the day there is uh you know a lot of information about there like about how you know the society will change due to Ai and I think sometimes we do tend to forget the negatives of AI because AI is something that is 24/7 it doesn't sleep and of course it's going to enable surveillance now I don't think surveillance of things like security and safety are bad I just think that when you do have certain people in power I just think that people you know humans are a bit like valuable like in the sense of like humans are just you know they're not perfect meaning that you know they can go on power trips they can you know exercise their power in the wrong ways it's just pretty scary stuff now what's crazy about all of this as well is that we've already seen how this plays out you know with China's dystopian AI you know they've got like some social credit system I don't want to get into that too much but um you know hopefully this dystopian AI doesn't turn out to be re reality I mean we never know how future is going to happen it's going to move pretty quickly but it's something that you know we have to consider now ai video has gotten pretty crazy this is something that you know has gotten even crazier as time has gone on basically this is Will Smith in V2 we did see examples with this Benchmark being crushed by Sora but um recently we got to see an update with V2 because V2 is largely the better model and I would say that every example I've seen from V2 not even the Cherry piix versions have just constantly continued to break the internet this was one that was doing the rounds uh and I think this just goes to show the progress that we've made in just two years so I mean where thing's going to be 2 years from now that is going to be something that's going to be super interesting to see considering might be small areas for improvement but it's pretty crazy whil everybody is focusing on AI I also am focusing on AI and I'm using this 60-second window to shamelessly plug my AI grid Academy this is something that I've recently revamped where over 200 people are using AI agents and prompt templates to make money and every week we share a bunch of different resources on how people can take advantage of this AI Revolution as we know that many companies are focusing on this as their future one of the things I have is a workshop Wednesday which is going to start in 29 hours which is I break down winning strategies into simple actionable steps I have private videos where I'm showing you guys exactly how to use AI agents to make money then on Fridays I actually showcase the best a tools to automate a lot of this stuff and of course on Sunday I actually break down various case studies from in the community and of course around the internet on people that are actually making serious money with AI if that is something that interests you don't forget to check it out it's going to be one of the first links in the description and now let's get back to the video also had a economic estimate on AGI so Stuart Russell actually says that building AI would be the biggest event in human history because it would mean supporting everybody on Earth with a splendid style of life since the minimum cash value of AGI would be $15 quadrillion dollar now the reason I think this is important is because a lot of times people say that yes we're in an AI bubble because startups are valued at ridiculous amounts but the startups in the AI space I personally think that the majority of them aren't overvalued because if there is a like 10 to 15% chance that they can manage to succeed then the upside is so great that it's worthwhile taking that chance like if openi can achieve AGI and they can you know embed that system into various parts of the economy they're probably going to be the most valuable company on Earth next to Nvidia and I think the economic value that they generate from that would just be absolutely incredible so um it's something to account for when you know you're looking at the AIS and these valuations so I'm going to leave aside the question of whether we will succeed in 2027 or 2037 or 2047 and just talk a little bit about why uh success would be the biggest event in human history um one obvious reason if we think about the upside is that if you have real general purpose AI then you can do everything that humans have been able to do namely produce a civilization that supports at least hundreds of millions of people in an objectively Splendid style of Life we'd be able to do that at much greater scale at much lower cost meaning that we could provide that objectively Splendid style of life not to hundreds of millions but to everybody on Earth and if you just take uh average Western middle class lifestyle and you say well now everyone gets to enjoy that lifestyle that would be about a tenfold increase in the GDP of the world and the net present value of that is about 15 quadrillion dollars so that's the cash value minimum cash value of AGI as a technology so then you can see why we're investing relatively large amounts of money because in comparison those amounts of money are minuscule now a bad news story this week was the fact that meta is planning to flood platforms with AI powered users and I made a video on this and I spoke about why this is probably not the best thing to do and why it's like it's a nuanced issue because there's just a variety of things that you can talk about but I'm going to read this and then I'm going to you know get into that stuff but they basic talk about how you know meta is aiming to have Facebook filled with AI generated characters to drive up engagement on on its platform as a part of its broader roll out of AI products and they're basically talking about you know these AI characters will be created through you know you know meta's AI studio with the idea being that you can interact with them almost like you would a real human on the website now for me personally I think this is 50/50 because when I made a video about this that I'll link in the description I basically spoke about how whilst yes fitting filling a website with AI powered users is awful because that defeats the purpose of social media the point of social media is to interact with other humans so if you do that it defeats the entire purpose of the platform and that makes no sense but I also said that there are so many people on social media that have little to no interaction with other humans that social media may be better for those specific users if they have someone to talk to so once again this goes back to that loneliness issue where will this situation actually make social media more usable by humans if those said humans that are you know very humanlike they DM you they message you they interact with you content and you struggle to tell whether or not it is a human or real person do those social media platforms actually get more engagement that's going to be something that's going to be interesting to test and I do wonder how advertisers are going to you know get into that because of course if you're putting your content out there you want it to be seen by real humans you want real purchases but of course if you now have a platform filled with AI powered users that is going to be really strange for meta because their bread and butter is having eyeballs and allowing people to advertise on their platform so it's going to be really interesting to see I've note that they've you know cancelled some of the profiles that they've had but I think this is something that they really gotting on but I wouldn't be too worried about this because meta have been wrong several times when it comes to AI not in terms of the technology side but in terms of the implementation side it wasn't until recently where you know they were investing billions and billions of dollars into the metaverse so this might be another fied plan but if it does work it will Mark a very interesting change where the AI manages to merge with the internet now from more you know in meta we actually had Yan Lun talk about human level AI being inevitable for meta as it develops the vision for super intelligent AI assistants that connect people to other people so how long is it going to take so I I think to have a possibly a system that at least to most people feels like it has sever intelligence as humans if all of the plans that all of the things that we are imagining will work okay so those japer architectures and you know some other ideas that we're we're playing with succeed I don't see this happening in less than five or six years okay but now is it going to happen in five or six years and I think the there's a a distribution with a tail that's very long and the the history of AI is that people just keep underestimating how hard it is I'm probably making the same mistakes right the same mistake right now you know when I say five six years this is is if you know we don't run into a major obstacle that we didn't foresee if all the things that we're planning to try out actually work if things can scale if computers you know accelerate and all that stuff like you know there's a lot of things a lot of planets that need to line up for this to happen so that's the best case right it's not going to happen next year like you might have heard from from from you know some other folks Sam yeah right Yeahs you know you know you know various various people uh or or you know Dario yeah it's going to happen within the next two years or something uh no I know people you know find his ideas honestly fascinating because like I said you know if everyone is thinking the same thing then nobody is thinking so if everyone thinks that we're going to get ASI in you know the next two years then no one is thinking and I think it's important to have people that you know share a different perspective on how the world is because it's important to have you know criticisms of any view that's so you can have an open discourse and that's how you can progressively move forward allowing yourself to actually think okay maybe he's got a point maybe that architecture might be more effective than some of the other ones and even if not maybe it has other applications and use case so the next thing that we got here was something from clone Robotics and this was super super interesting so they were showing off their humanoid which is you know the Torso and I think clone robotics is probably one of the standout companies for me because their you know technique their entry into humanoid robotics is an angle that no other company is taking which means that this could turn out to be one of the most realistic humanoid robotics companies ever I mean this is definitely something out of Westworld the water hydr system that they're using seems you know Innovative and I mean right now it definitely looks a little bit crazy but I think with companies like this as long as they can get significant funding they could definitely be pioneering the humanoid robot race now I personally think that humanoid robots is extremely hard and honestly I'm not even sure how you would you know have this kind of system in maybe nvidia's Isaac Sim so that you could you know train it with reinforcement learning I'm honestly not even sure how this kind of Hardware platform is going to be trained I mean I'm sure these Geniuses will figure that out but these torsos are using pneumatic actuation with valves that produce you know noise from the air exhaust and it's pretty crazy it's pretty crazy because you know they've got 9910 muscle fibers animating its 164 degrees of freedom it's got like 182 senses for feedback control but I mean imagine what this looks like in 20 years when they figured out all the Kinks like imagine this is something that is smooth it's relaxed it's say to walk around it's able to talk to you it's strong I mean definitely seems like we're living in the future hopefully honestly this isn't just a fancy demo but they also did showcase some pictures of what it could look like in the future I don't know about you guys but this to me looks like some kind of alien I wouldn't like to come home to this kind of robot cooking in my kitchen I'm sorry maybe I'd want one that just has like a smiley face on it but now if we're going to talk humanoid robotics it might be worthwhile mentioning nvidia's new Cosmos which is something that's pretty crazy in terms of you know the development of humanoid robots and overall robotics so I would watch this because it's a development that you don't want to miss the AI industry robotics industry is continually accelerating at a rate that we can't even fathom the next Frontier of AI is physical AI model performance is directly related to data availability but physical world data is costly to capture curate and label Nvidia Cosmos is a world Foundation model development platform to Advance Physical AI it includes Auto regressive World Foundation models diffusion-based World Foundation models Advanced tokenizers and an Nvidia Cuda an AI accelerated data pipeline Cosmos models in just text image or video prompts and generate virtual world States as videos Cosmos Generations prioritize the unique requirements of Av and Robotics use cases like real world environments lighting and object permanence developers use Nvidia Omniverse to build physics-based geospatially accurate scenarios then output Omniverse renders into Cosmos which generates photoreal physically based synthetic data whether diverse objects or environments conditions like weather or time of day or Edge case scenarios developers use Cosmos to generate worlds for reinforcement learning AI feedback to improve policy models or to test and validate model performance even across multisensor views Cosmos can generate tokens in real time bringing the power of forite and Multiverse simulation to AI models generating every possible future to help the model select the right path working with the world's developer ecosystem Nvidia is helping Advance the next wave of physical Ai and also might want to pay attention to nvidia's road simulation which is where they're talking about autonomous cars because that is going to sneak up on a lot of people we now have wayo we now have zukes and there are other companies that are doing autonomous driving so it's worthwhile to see how that's going cuz I do think that it's going to come to a city near you a lot quicker than you do think and most people are going to be surprised and be like wait I knew AI was developing but I didn't know autonomous cars were this far along but it's something that is expanding rapidly and you know today instead of showing you a whole bunch of uh uh videos of of cars driving on the road I'll show you some of that too um but I want to show you how we use the car to reconstruct digital twins automatically using Ai and use that capability to train future am models okay let's play it the autonomous vehicle Revolution is here building autonomous vehicles like all robots requires three computers Nvidia dgx to train AI models Omniverse to test drive and generate synthetic data and drive agx a supercomputer in the car building safe autonomous vehicles means addressing Edge scenarios but real world data is limited so synthetic data is essential for training the autonomous vehicle data Factory powered by Nvidia Omniverse AI models and Cosmos generates synthetic driving scenarios that enhance training data by orders of magnetude first omnimap fuses map and geospatial data to construct drivable 3D environments driving scenario variations can be generated from replay Drve logs or AI traffic generators next a neural Rec construction engine uses autonomous vehicle sensor logs to create High Fidelity 4D simulation environments it replays previous drives in 3D and generates scenario variations to amplify training data finally edify 3DS automatically searches through existing asset libraries or generates new assets to create Sim ready scenes the Omniverse scenarios are used to condition Cosmos to generate massive amounts of photorealistic data reducing the Sim to real Gap and with text prompts generate near infinite variations of the driving scenario with Cosmos neotron video search the massively scaled synthetic data set combined with recorded drives can be curated to train models nvidia's AI data Factory scales hundreds of drives into billions of effective miles setting the standard for safe and advanced autonomous driving now the most interesting thing I personally think that happened this week was that samman actually said we are near the singularity he says I always wanted to write a six-word story here it is near the singularity unclear which side so this is something that is you know like there's just so much that goes into this cuz I think I made a 25 or 28 minute video speaking about this and essentially what he's talking about here is that like the singularity is the point at which we move forward and we don't know which side we're on and in this he was talking about you know the fact that this could have been the you know simulation Theory or the fact that we have reached a point where AI is just developing so fast that technology is in that Loop where you know the advanced technology makes us more advanced technology and things just start to move quicker and quicker and quicker and of course I think this is something that sounds crazy to people that are not in the AI space but I've read two research papers in the past week and I actually spent a lot of time reading and digesting those research papers because I wanted to not only create a video on that so I could educate the masses but I also wanted to truly understand how the system works and you know how the algorithms work and when I truly understood the kind of technology that they were developing you know from the ground up here it was something that I could start to understand why these individuals at these companies truly believe that we are so close to artificial super intelligence like if you watch those videos and you understand you know the math you understand the results that they've had you understand that when they run certain algorithms how they just keep getting smarter the benchmarks keep increasing and they're not getting Dumber in the sense that like there doesn't seem to be a real slowdown like there doesn't seem to be any bottleneck just yet I mean it's clear to us that we are very close to artificial super intelligence and you have to understand that right now the biggest thing that is slowing down this is not the fact that we don't know how to get there it's just the fact that we don't have enough compute and we don't have enough power to actually get to you know super intelligence so that is something that I think is a key differentiator and I think that's why a lot of people in the industry are starting to say that look super intelligence is around the corner AGI around the corner so it's going to be super interesting to see what happens with that regard now for all of you guys who want to know the actual sizes of the models and I know this might not seem like news but to me this is pretty big news because we actually get to see how many parameters these models have and it's so important because we really want to know how smart the models are compared to their relative sizes and these architectures are highly highly secretive things so this paper you know it's a Microsoft paper and previously when Microsoft did say that GPT 4 was 1. 8 trillion parameters and a mixture of experts that was something that later was revealed to be true and now they basically spoke about you know these model is so they said that claw 3. 5 Sonic is a 175 billion parameter model I'm not sure if I believe that at least the latest iteration of the model considering number one just how smart that model is and of course just how restricted the model is meaning that you know if you have like 10 to 15 responses from Claude it just I don't want to say it stops responding but they start to restrict your account of course GPT 4 was 1.
