ElonMusk AI Is GAME Over! Googles New Health AI OpenAI Sora Exposed
25:03

ElonMusk AI Is GAME Over! Googles New Health AI OpenAI Sora Exposed

TheAIGRID 12.06.2024 31 840 просмотров 660 лайков

Machine-readable: Markdown · JSON API · Site index

Поделиться Telegram VK Бот
Транскрипт Скачать .md
Анализ с AI
Описание видео
Learn A.I With me - https://www.skool.com/postagiprepardness 🐤 Follow Me on Twitter https://twitter.com/TheAiGrid 🌐 Checkout My website - https://theaigrid.com/ Links From Todays Video: Welcome to my channel where i bring you the latest breakthroughs in AI. From deep learning to robotics, i cover it all. My videos offer valuable insights and perspectives that will expand your knowledge and understanding of this rapidly evolving field. Be sure to subscribe and stay updated on my latest videos. Was there anything i missed? (For Business Enquiries) contact@theaigrid.com #LLM #Largelanguagemodel #chatgpt #AI #ArtificialIntelligence #MachineLearning #DeepLearning #NeuralNetworks #Robotics #DataScience

Оглавление (6 сегментов)

Segment 1 (00:00 - 05:00)

so one of the first stories that I found to be truly fascinating was that Elon Musk actually dropped his suit against open Ai and Sam Alman now this was fascinating because if you weren't aware Elon Musk was in the process of suing Sam Alman and open AI because essentially what he was stating that they didn't really stick to the deal of open AI if you're unfamiliar with the backstory of how open AI came to be let me give you guys a little rundown basically in the early days of open AI they set out to develop AGI or artificial general intelligence for the benefit of humanity but of course over time that project had been transformed into a for-profit entity that's larly largely controlled by the principal shareholder Microsoft basically mus light look we started this company for one thing and the founding agreements that we were going to create AGI that was going to benefit the entirety of humanity but over time this nonprofit has turned into a for-profit that is now controlled by Microsoft what's going on here I'm going to sue you guys and essentially he did a 35 page complaint plus attached exhibit that he filed in March to remind the world of his position in the creation of a company that since has become one of the hottest startups on the planet and open a i ranked first on cnbc's disruptor 50 list largely thanks to the viral spread of chat gbt so essentially Elon Musk on Tuesday surprisingly withdrew his lawsuit against open AI in and two of the company's Founders Sam mman and Greg Brockman in California's state court so his decision to file to dismiss the suit came just one day after he publicly criticized open Ai and its new partnership with apple the case was dismissed without prejudice according to a court filing obtained by CNBC and of course the case was actually built on a questionable Legal Foundation because the contract at the heart of the suit was not a formal written agreement that was signed by all parties involved essentially this was more of like a verbal agreement rather than something they all wrote on a document and signed so this is something that had a very shaky legal ground so overall I think this is rather fascinating because now I'm trying to dive into the real reason with as to why Elon Musk decided to withdraw the suit did he somehow figure out that look maybe over at x. a were actually developing our own proprietary software SL AGI level systems that are doing quite well and consider considering the fact that you know Elon Musk himself has done some remarkable AI developments which I'm about to talk about in the next AI story I think maybe he's thinking you know what I can catch open AI somehow as long as I manag to do Tesla and x. a with a lot of the stuff that's going on in the background so him withdrawing from the suit I think maybe this is just you know a tactical move for him to just further his efforts and focus his time onto what really matters and if you're wondering what Elon musk might be working on well I'm going to show you all how crazy that is in the next story so Elon musk's Tesla recently did a kind of announcement that was rather fascinating they spoke about the rundown of what they've achieved since 2018 under Elon musk's leadership now this was truly incredible they detail every single thing that's been going on since Elon musk's leadership and the future things to come and of course they said incredibly right here and this is the main take way from this tweet which is why this is making it into this video is that apparently they've deployed two Optimus Bots performing tasks in the factory autonomously so if you're not aware with as to what I'm talking about here the Tesla team the Tesla company have made their robot Optimus which is apparently now effective enough to be autonomously performing tasks in the factory and like they said it's only June so this is very surprising because humanoid robots have been long thought of as a real problem that is not going to be solved anytime soon but currently what we are seeing here is that these robots are being deployed and doing things in the factory now it was previously thought that what we saw here was only just some beta testing or some Alpha Testing of how things were going to be but I think this should all show you the future of Robotics is really going to be here if this demo of the 100% onetime speed test level being autonomously completing tasks in this Factory doesn't show you how the entirety of society is going to change in the next couple of years when these robots get completely better and a lot more effective I think you're vastly underestimating how much the future is truly going to change we have a situation on our hands they're currently deploying these autonomously around the factory what do you think happens when these Tesla Bots get an upgrade they get

Segment 2 (05:00 - 10:00)

embodied AGI systems or like a baby AGI system combined with Tesla's years of Robotics and AI engineering to build something that is truly remarkable and effective I wouldn't be surprised if within the next 3 to 4 years we have autonomous fleets of Tesla Bots doing huge numbers of tasks autonomously which vastly increases their output efficiency and I think that Tesla has the infrastructure and has all of the necessary tools to be the leader in this space and with El mask behind it that is someone I'm not going to bet against so with embodied AGI on the horizon this announcement from Tesla stating that they have two working autonomously just shows us how close we are to that future of where we're going to be going now there was also an interesting piece of news regarding Elon Musk and Apple so Apple's State of the Union released a short video in which they detail exactly how their on device llm to work so essentially yesterday Elon Musk went on a very interesting tweet storm about how Apple intelligence works if you weren't familiar you didn't watch the video from yesterday a quick rundown Apple's intelligence is basically a kind of AI system that apple have natively built into iOS 18 and there were going to be AI systems running on device on your new iPhones and iOS systems now however there was a caveat to this because there's currently an issue that many people were kind of thinking about and it's this meme right here Elon Musk tweeted apple is you know handing off your data basically to open Ai and this is a remarkable meme because it's pretty funny however it's not true whatsoever so El musk actually saw this and basically said that right here that of course you know it's patently absurd that Apple isn't smart enough to make their own AI yet it's somehow capable of ensuring that opening eye will protect your security and privacy Apple has no clue what's actually going on and they'll hand your data over to open ey they're selling you down the river and of course he also stated that if Apple integrates openai at the operating system level then Apple devices will be banned by companies and this is an unacceptable security violation now whilst initially reading this tweet you might be thinking okay so Apple have done something that's pretty crazy because most people I'm not going to say most people didn't watch WWDC but most people weren't truly hyped about it because there wasn't any gamechanging announcement that was too crazy so essentially the reason that this is a story is because of course Elon Musk says that Apple isn't smart enough to make their own AI however the recent video which I just showed you guys shows that this is not really the truth so in Apple's State of the Union video they actually talk about the 3 billion parameter small language model that uses adapters trained for each specific feature the Fusion model also does the same thing adapters for each style and anything running locally or Apple's secures cloud is an apple model and is not opening eye now as someone like myself who's actually been paying attention to the fine-tunings of what's going on it can be quite easy for people who just watch the demo to easily Miss what's going on but Apple have been diligently actually very diligently working on their own AI systems in order to ensure the privacy of users that's why this video right here is of no surprise to me if you've ever been paying attention to the research papers which of course most people aren't you've seen that apple have developed a multitude of small language models that are on device capable and of course really comparable with the other top small language models like you know Google's Gemma and potentially even some in the Five series so basically this entire video summarizes the fact that what we have is Apple secure language models running but it's not actually handing off the data to open AI so this is something that I just did want people to understand because there is that misunderstanding there that could have easily happened but it's something that just isn't true so you don't need to worry about your data being handed off to open AI I mean many of people who are going to be using this anyways do use open AI but it's just something to just add to the information now Google actually did something rather fascinating too they released this paper which is called towards a personal health language model so a personal health llm and basically this is something that they've done with a app and it's basically something that is able to do really well in the personal health domain so you can see right here this is kind of your demo on how everything is working together so you can see it's able to offer coaching recommendations youri sleep medicine expert you are given the following sleep

Segment 3 (10:00 - 15:00)

data the user is a male 50 years of age list the most important insights you can see right here it's got the insights the domain knowledge the sensor grounding and of course the recommendation so this is going to be the structure of how they structure these answers you can see they having trouble falling asleep otherwise good sleep efficiency of 86% above the recommended range of 80 to 85% adequate deep sleep consistent sleep schedule no nap make sure your bedroom is cool and dark continue to avoid naps and keep a sleep schedule the other things right here you can see that what types of muscular contraction occurs in the pectoralis major during the slow control downward phase of a bench press then you can see it's The Eccentric and basically I'm not going to read every single interaction that we have here but the point is that we have a system of a personal health llm that is probably going to be the future of how these you know on device systems actually work with us and our health knowledge and I think that this is going to underscore the future of how we're going to be dealing with these models because Healthcare is an area where you know we really haven't seen that much Innovation just yet in terms of diagnosis from llms we've seen a lot of papers from Google that have shown us that they've been able to do quite a lot and I think that one thing that you know surprisingly hasn't been there yet is open Ai and I think open ey previously hinted that they're going to make a lot of grounds in areas that they previously haven't and I think that the healthcare app kind of area is going to be a huge area for open AI in the future that they're likely going to dominate remember I said this so if you see open AI release some kind of Health thing with GPT 5 do not be surprised because I can guarantee you if we actually look at exactly what is going on with Healthcare all you're really doing is just taking an analysis of certain symptoms and then kind of crosschecking that with a database of possible things and then of course you could use a vision system and you're basically trying to be as accurate as possible combined with the patient's information in order to diagnose what they have with remarkable accuracy that sounds like something that could be relatively easy for an llm or an AI system as long as it has multiple different inputs of course this doesn't negate the fact that doctors are going to need to be there for certain physical examinations but it's going to provide a remarkable level of affordable healthcare especially if these models are open- sourced or if someone creates an app that everyone can easily access and use and it's essentially just a base level of healthcare that you can really use if you want more information about whatever current condition you might be confused about so that's one of the areas of AI that I'm genuinely most excited for now something that I really wanted to talk about and something that is just not getting enough attention this is clling AI so clling AI has been the you know thing that has been absolutely crazy in the past few days for those of you that have been paying attention to the video in I guess you could say the text to video AI space essentially with king. a what we are seeing here is a revolution in text the video so this ladies and gentlemen is not actually the film created by open eyes recent you know Film Studio that they did a small collaboration with but this is actually the work of an individual talented traitor who decided to use China's newest textto video model in order to create something remarkably similar to what open AI has released so what you're currently looking at is a day with the balloon man now if you're wondering why on Earth are you showing me this video of a man with a balloon and talking about open AI well basically open AI did a recent thing where they showcased the capabilities of their text to video model by showcasing a man who had a head as a balloon and this was something that as you can see is being done by the newest text to video model now you can see here that this was what open AI Sora was able to do of course this does have remarkably a little bit better quality and it does look to have a lot more character consistency but I do want to state that open AI sora's version the version that you're seeing on screen right now isn't actually just raw output from the Sora model what you're actually seeing is essentially The Talented team the shy kids who created this masterpiece essentially you're seeing the derivatives of those generative outputs combined together to create this wonderful storytelling and this piece of cinematic history and I think what we have here is something that shows us that these tools whil yes because they are generative they can make a whole lot of mistakes they actually allow us to express our creativity in ways that we truly

Segment 4 (15:00 - 20:00)

couldn't before so what you're seeing here is you know there's a lot of Roto scoping changes going on in the background and this was something that they had to change a few times so when you're making the comparison between this what you're seeing on screen right now and the other one which is king. a it's important to note those caveats because I think if shy kids actually had uh access to clinging. a they'd be able to create something fundamentally on a very similar level now there was also this demo from cling AI which is really cool because it's so crazy that you can see that there's literally text to videos of like uh I don't even know if this is Call of Duty or what kind of a video game this is but I think what this shows us first of all is that in the future what actually might happen and what actually might occur somehow is essentially a situation on our hands where we truly have generative media that doesn't exist before I'm not entirely sure how this is going to work but I can see maybe four or five generations from now people and including myself being able to create you know full length videos where you can have entire novel scenes that is just incredibly fun to look at now there's also this thread where someone actually compiled all of these videos that you can see are truly fascinating you can literally see that there are these you know hamsters or I'm not even sure what they're called they're probably hamsters or rats just eating um popcorn they're watching a rabbit walk towards this area right here it's truly you know impressive that you know the rabbit was able to throw a few punches uh kind of the content that we would never ever see you know people like the rspa would definitely be uh mad about that and of course you don't rabbits fighting tigers that's just going to end horribly but um yeah what we can see here is a lot of content that you know we previously wouldn't be able to get so there's this thing called threads of humanity there's just you know there's a whole load of content that is now coming out because with text a video it's making the average person be able to you know create content that they otherwise wouldn't and when you combine that with the creativity of certain individuals you do get a lot of interesting things and right here this video we can see the water the ripples Rippling really nicely so I think this shows you that cling AI is kind of this next level um and I think what we're going to see is a lot more things happen in the future especially with the future models now one question I do have because this is something that is maybe changing the tides a bit is that this was released by China so I'm wondering since this has shown that you know open AI doesn't have the lead anymore in the sense that across all bounds they're not really you know like they don't have the frontier space and I think in the future now that everyone else is also in this space because open I basically had a year and a six months in order to kind of set the pace for what was the state-ofthe-art and the thing is with things like Sora uh you know a lot of people they got caught off God by it but the fact that you know China's managed to catch up with their own models and the fact that a lot of them are catching up to GPT 4 I think the next cycle is going to be even crazier in terms of what AI systems we can get because now we're truly going to see a level of innovation that we haven't before because there's not going to be that crazy standard unless open I does something crazy again and with clling AI this was one of my favorite examples this was animals being musicians I don't know I thought this was really cool because this is just something really weird that you know the millions and you know thousands of hours it takes to render this kind of stuff in a 3D world it's just something that people wouldn't really do so it's something that I think you know is kind of interesting now what's also kind of interesting was that I think Sora is probably not going to be a tool that you know many people including myself do have access to so it seems here that it says Brian Lord the CEO of creative artists said yesterday that open AI took the sky voice down when he asked them because they are mid deal with multiple Studios over Sora use and that those deals will be announced imminently he adds that open AI Executives were in the middle of multiple deals with Studios in Hollywood that are starting to be rolled out and they didn't want bad press so essentially it's kind of fascinating because maybe we could see a scenario where these AI companies kind of transition from B Toc which is business to Consumer to B2B business so I think that you know and is something that I've spoken about on my private patreon where I spoke about you know future AI systems are very likely to be B2B which means that open AI are going to be giving deals to these billion dooll companies because they're the ones that can afford it there's the ones that you know they can have regulations they don't really have to worry about like apple on the back end using it to create some nefarious technology and they also don't have to

Segment 5 (20:00 - 25:00)

worry about the public backlash when they release the technology one of the things that I think you know is going to lead to this is the fact that if opening are released Sora a lot of people are mad at them because they're like yo you're taking away people's careers you're taking away their livelihoods you're making generative art you're just taking away all the fun stuff but if they license this to Hollywood companies I guess on the back end it's something that benefits open a ey they don't get the bad press and it's something that kind of works for them I been playing around with Sora this latest thing that open AI launched that like generates video and I have a beta version of it and um it's pretty amazing like it's pretty good and you can create like good like 10 15 second videos that are like look very real it still makes mistakes it still doesn't quite understand physics it still like has some bobbling but if you look at the generation of this that existed one year ago go as compared to Sora it's Leaps and Bounds in fact there's footage in it that I would say you could easily use in a major motion picture or a television show like why would you go out and shoot an establishing shot of a house in a television show when you could just create the establishing shot for $100 like to go out and shoot it would cost you thousands of dollars action scenes of like me jumping off of this building you don't have to have a stunt person go do it you don't have to you can just go do it I shot I came up with a prompt of an Ultram marathoner running across the desert being chased by a sandstorm I didn't have to hire a CGI Department to do it I didn't have to hire I in five minutes rendered a video of an Ultram marathoner running across the desert in being chased by a sandstorm and it looks exactly like that so fast forward and let's go 30 times more performance you'll be able to render a whole movie you'll just come up with an idea for a movie then it will write the script then you'll input the script into the video generator and it will generate the movie so instead of watching some movie that somebody else came up with I just generate a movie and then watch my own movie and what's going to happen is there's going to be more content than there are eyeballs on the planet to consume it so any one piece of content is only going to be as valuable as you can get people to consume it um and so thus the sort of catalyzing water cooler uh version of something being good is going to have the bar to go way up because why are you going to watch my movie when you could just watch your own movie so those are some of the interesting dynamics that are going to be changing in the media industry and I think this shows us that I think you know authenticity and realness in content is something that has been a prevailing Trend among social media platforms you know when people are you know frequently real on camera authentic this is something that I've seen people you know truly enjoy like even sometimes in the videos that I tend to create sometimes I purposefully leave mistakes in so you know that it's literally just one guy and a microphone actually recording these videos not some giant Studio behind it because I think that is just a level of you know authenticity that you can't fake and I think people really resonate with that content more often you know like when someone has a you know video camera on their phone and they're recording the kind of footage that is going to be kind of the content that I think prevails in this industry because whilst yes you can you know generate an entire movie I think real person and real people like I've said before and humanness is going to be the prevailing thing in the future that in the age of AI abundance is going to be something that people really want like we could read any short story about anything but if I could read a story about a real human that went through a real experience that's going to be something that I want to read we could make a picture right now of anything the generative AI but if someone took a picture of something really cool happening down the street I'd actually want to see that because it happened in real life and it's not generative it's not fake it's grounded in reality and of course I could go watch a movie that is completely AI generated sure for entertainment but if there's a story like you know Mr Beast series or a human actually doing something that's completely real I think those things are going to be far more entertaining than you know just having this entirely AI generated thing now I think movies are a different genre because you're literally going there for something fake and of course that's the entire premise of movies but I think it's going to be an interesting Dynamic if we do get to that stage but there's a lot more that I can talk about on the back end on how you know uh this is going to require millions and millions of compute and I think just you know that the movie studios are probably going to be pumping out a lot of movies a lot quicker because they've been using AI but of course you could even talk about you know as some people even going to watch movies that have been you know using generative AI because you know no people were actually you know behind the scenes so you've got all of these kind of different things going on in this entire industry but I think overall what we're starting to see is potentially you know on the latter a shift to where we might

Segment 6 (25:00 - 25:00)

not even get some of these kind of products so let me know what you guys think about that

Другие видео автора — TheAIGRID

Ctrl+V

Экстракт Знаний в Telegram

Экстракты и дистилляты из лучших YouTube-каналов — сразу после публикации.

Подписаться

Дайджест Экстрактов

Лучшие методички за неделю — каждый понедельник