# HUGE AI NEWS: OpenAI Releases Sora, Fully AI Generated Games, New AI Agent, Recreating The Brain... ## Метаданные - **Канал:** TheAIGRID - **YouTube:** https://www.youtube.com/watch?v=GAeOXdfjIfg - **Дата:** 10.12.2024 - **Длительность:** 24:58 - **Просмотры:** 23,655 - **Источник:** https://ekstraktznaniy.ru/video/13603 ## Описание Prepare for AGI 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: https://x.com/tsarnick/status/1864147230187512308/video/1 https://x.com/Andrew_C_Payne/status/1863957226010144791 https://x.com/theworldlabs/status/1863617989549109328 https://x.com/GoogleDeepMind/status/1864341001214804370 https://x.com/GoogleDeepMind/status/1864367798132039836 https://x.com/OpenAI/status/1864735515121168695 https://x.com/adcock_brett/status/1865804506790838395 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 Music Used LEMMiNO - ## Транскрипт ### Segment 1 (00:00 - 05:00) [] so one of the first stories that actually released this week was of course the 01 preview transforming into the 01 full and of course this now actually has advanced image analysis so I think this is one of the things that is really underrated about 01 not only do you have prom mode but one of the main things that you do have is Advanced image analysis is something that is quite underrated because it's not used by everyone but if you have an image there is a billion different things that you can do and this is something that was released with the full version of A1 I know a lot of people aren't taking advantage of this but I can guarantee you that in a few weeks you're going to see some crazy prompts from individuals around the globe showing you all exactly what they've managed to do with 01 image analysis and of course with 01 now the 01 F version a lot of people at launch were stating that this version was kind of disappointing however the benchmarks actually kind of lied because the benchmarks that were actually there for 01 were a little underestimating exactly what 01 was so if you remember the benchmarks for 01 and 01 Pro mode a lot of people were talking about why would I pay you know $200 a month for this and essentially after a lot of people have actually began to use the 01 PR mode I'm starting to see a different vibe in terms of what people are not expecting out of open AI but the kinds of results that they are getting and the kind of I guess you could say Vibes that I'm hearing from people that are using this model for some of the most advanced use cases so for example I'm seeing people say wow the 01 Pro mode is really worth it and 01 is honestly so much better than I had thought so you can see right here as well what I've done is I've shown you guys the differences and one of one I wanted to highlight was you can see on competition code you can see that the 01 preview compared to 01 is actually you know like nearly 100% jump in terms of you know what you're seeing over there so we can see that 26 is 64 that's really crazy and then of course um I don't think everyone is going to buy this but I think some people might just because of how powerful it seemingly is but of course that is the fact that you do have the 01 Pro mode which is essentially the model that thinks for the longest and what's crazy about this is that this is just 01 and we can see that the benchmarks are just continually increasing with the amount of compute that we do add to the model and I do wonder what's going to happen with further models like O2 03 04 and one thing that I do want to quickly add here and this is something that people just continually forget anytime you're looking at any um you know AI Benchmark one thing that I want you guys to remember is the fact that when we look at competition math competition code PhD science of questions any absolute Benchmark remember that benchmarks almost all of them have some kind of error which basically means that these AI models are a lot smarter than we think because when we see the models getting 94 95% on the test usually that means the model has saturated The Benchmark I mean if all of those benchmarks that I just spoke about a lot of them have maybe 2 to 3% errors or maybe 5% errors so when you see a lot of these models hovering around the 995 to 96% area and you're thinking wow I wonder when these models are going to get 100% understand that with some benchmarks it just simply isn't possible so that's why I'm stating that in the future samman stated that look Benchmark saturation is going to be a thing these models going to be able to completely as in your benchmark so it's going to be really interesting to see what that means for the AI space and of course with this one of the things they focused on with 01 Pro mode which cost $200 a month is the reliability so these results were based on four out of four so the did four tests and if you got it right four times that was the only time they counted that response now something crazy that was really cool was Google's genie2 now Genie 2 was absolutely incredible because I think this was something that went really underrated SL under the radar simply because we don't have access to it yet but this was something that was launched by Google and I think it was one of the things that just is like mindblowing when you think about what it means for the future of AI now of course you can see on screen what you're able to do is simply create video games by Simple Text prompt so you can say create me a game that exists on Mars create me a firstperson shooter in this area and you can literally explore them now I do know that you know multiplayer games are pretty much a lot different compared to the single player games but I do still believe that the applications for this are going to be completely different like if we can manage to figure out how exactly we can you know solve these generative worlds and make them consistent in terms of you know the the consistency from looking to one way to looking another way and that's something that Genie has been able to solve it's going to be really interesting to see how the gaming industry is impacted by this I know some people would love an AI generated game the idea of playing games that are procedurally generated everything is completely new that has endless playability is going to be something that I think you know in the near future maybe let's say 10 to 15 years maybe games are going to be AI generated in a way that every person has their own unique tailored experience and I wonder what that will do to the replayability of a video game if every single time ### Segment 2 (05:00 - 10:00) [5:00] it's AI generated everyone has a new and crazy experience so this was something that I wouldn't underestimate because Genie 1 was 2D and now Genie 2 is actually 3D so this was a huge Improvement in terms of what they were able to do and this was actually even better than the Minecraft AI that was released not just I think maybe one or two weeks ago so this was something that I think went under the radar did make a video on it but I know that this is something that they did miss out on so it's going to be really interesting to see what happens once this stuff finally gets integrated into other games and of course when demos do become available so that's what I'm really excited for so that will be really interesting then we had Google's gen cast which is an advanced AI powered weather forecasting model developed by Google Deep Mind this basically leverages machine learning techniques to deliver faster and more accurate weather predictions compared to traditional models in this video you'll see an overview of exactly what it's capable of and how AI regardless of llms is actually changing the entire world for centuries humans have struggled to understand the weather from putting a wet finger to the wind listening to Cricut chirps and looking to the Stars For answers now we believe AI can help us advance and understand how we can predict it meet metet 3 and graph cast our deep learning models designed to generate more accurate weather forecasts traditional methods work by simulating entire systems with hundreds of variables interacting in practice this means abstracting those variables into equations and then manually coding them into algorithms our a ey models take a different approach they analyze past weather data to learn cause and effect relationships between weather conditions these learnings allow the models to find subtle Trends and patterns inside the data which they extrapolate to predict what could be coming metnet 3 and graph cast outperform the best systems in the world seeing further into the future more accurately efficiently faster and in better detail for every day and even extreme weather metnet 3 is already improving the accuracy of 24-hour weather forecasts on Google Search and we have open sourced the graph cast model which the ecmwf is using to generate forecasts and make them accessible on their website imagine a future where we can help people to prepare earlier for hurricanes enable Farmers to grow better crops or even help us decide if we should pack an umbrella AI won't just help us predict the weather it'll help people make better decisions about their lives I quickly do want to add that I know that the this video of gen cast was cool being able to predict the weather and stuff like that but I do think that like in the future a lot of AI tools are going to be able to really predict things that we may not have thought like there was even a study on how they Ed GPT 4 to I think with the degree accuracy of 80% accurately predict state by state what the you know the the results were which is just like mind-blowing like I couldn't even you know fathom how that was even possible with gbt 4 and this is why I say that these llms are a lot smarter than we think so I'm thinking in the future AI is going to be predicting a lot of different things and weather is just one of those many different things that's going to be able to do that now of course you know previously last week I spoke about how Amazon introduced a whole host you know whole Suite of models they introduced micro Nova light Nova Pro Nova Premiere but um I didn't know that Jeff Bezos was actually you know very engaged in terms of the AI going on at Amazon what is it that you're doing at Amazon ai yeah so I mean there are a few things but small but it's 95% AI and where do you it's just so because there we are literally working on a thousand applications internally so AI you have to remember AI modern AI is a horizontal enabling layer it it will it can be used to improve everything it will be in everything um this is most like electricity these kind of horizontal layers like electricity and compute and now artificial intelligence they go everywhere there isn't I guarantee you there is not a single application that you can think of that is not going to be made better by AI now whilst 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 AI 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 now one of the crazy things that actually happened this ### Segment 3 (10:00 - 15:00) [10:00] week was vahid kazimi who has a PHD machine learning and more importantly actually works as a member of technical staff at openi said something absolutely incredible he stated that in his opinion they already have achieved Ai and it's become even more clear with 01 that we have not achieved better than human at any task what we have achieved is better than most humans at most tasks now I got to be honest that first sentence that first very line that just like I'm even like speechless trying to say this right now because if it is true that open I have achieved AI that would be a Monumental milestone in terms of human achievement so I think this thing right here of course him stating that look you know I in my opinion we've achieved AGI and I think you know the fact that a lot of people would say okay open I didn't actually say that but this is a member of technical staff at opening ey and he's actually said in his opinion they've already achieved opening and AI I think it's really important to see where we are in terms of the AI space I remember before it was you know laughed about to you know talk about AGI coming very soon I remember you know certain YouTubers were saying AGI within 7even months or whatever and a lot of people were laughing at them but you know a lot of people are now stating that look we potentially have AGI in December which is definitely a lot quicker than most people did think and it's going to be really interesting because I think the biggest thing that you can take away from this is the fact that as we move from 01 to O2 to these thinking models it's going to get really blurry to see what AI is and the biggest thing that I would say is going to be really difficult for everyone is the fact that a lot of times when we do have new technology unless it impacts you directly on a day-to-day basis you aren't truly going to understand how impactful that technology was the reason chat gbt was so crazy is because you know all of these college students could use it anyone could use it to write an email you gra you know draft like a legal document sometimes it was something that on a day-to-day basis you could actually use and I think that even if we do get to AGI and remember I spoke about how the fact that you know with prom mode it's basically good at really Advanced tasks which the average person just doesn't have any use for I think it's going to be really difficult for the world to grasp the level of change that we're experiencing now someone actually talks about this quite a lot Brian chesky the founder of Airbnb actually talks about the fact that you know things are changing but in terms of other advancements we also got llama 3. 3 70 billion parameters so I got to be honest llama keep surprising me I said llama it's supposed to be meta but they keep surprising me in terms of how much they are doing and it seems like they really really ramped up their AI efforts with the kinds of things that they are producing they you know revealed a lot of different things that they working on in the air space and one thing I do like about meta is that they're working on things that are not just LMS which is just truly fascinating so llama 3. 3 70b is actually the latest iteration in the family of large language models from meta and this one of course 70b has 70 billion parameters now this one is designed for high performance natural language processing tasks while maintaining efficiency and accessibility so this one is you know fine-tuned for instruction following task making it highly effective for dialogue and task specific use cases and we can see that the improvements are pretty incredible because one thing that's really funny and it's actually crazy how often this happens is that if we actually just take a look at how llama 3. 3 70b performs and if we compare that to the recently released Amazon Nova Pro you can see that it manages to just Edge it out in various cases I mean the MML U is 86 and the Amazon Nova Pro is 85. 9 now I do think that different models have different qualitative differences which aren't visible by just numeric values which is of course something that a lot of people think of course once you use the model you're going to figure out quickly that some of them are good at certain tasks While others just completely fail but what I want to tell you guys is that like it's crazy that Amazon can release a model this you know week and then you know just next week or like 3 Days Later llama can release another model and you know it doesn't make the model obsolete but it just means that like ah you know you thought you had something thing and then someone else just releases their own model which is like it's crazy to realize how quickly this stuff is moving and of course the price difference is absolutely incredible which is what this model is meant to be it's meant to be a really effective model that you can use for a variety of different use cases so it's going to be really interesting to see a lot of people building stuff with this model I'm going to be someone that is of course starting to build these models as well like I said before I'm going to be doing stuff like this in my community it's going to be really interesting to see how well this model performs on certain benchmarks and it doesn't seem that far as well off GPT 40 so some people are basically saying that this is like a local GPT 40 so it will be interesting to see how that model performs compared to that now speaking of local models will you take a look at this so open source managed to do something crazy open source video has gone absolutely you know wild we've got open source video Hunan which is uh like I don't know like the craziest thing about this like genuinely guys I was seeing this video and I thought this stuff looked really good okay then I saw ### Segment 4 (15:00 - 20:00) [15:00] people actually run the model and they made videos and they like posted them over Twitter and it was genuinely one of the most difficult things to see whether or not content was AI generated because this open-source video model I don't know what they trained it with but stuff that was like fil filmed on like a raw camera iPhone so the footage that you see it's not polished but it's so awkward that you will believe that it is realistic which just makes you question things even more like there was a video of trump um I think he was kissing another woman that I genuinely sat at my desk for like 5 minutes trying to figure out whether or not it was AI generated eventually I realized it was but um it was something that I was like okay we've definitely reached yanan Valley because this video generator is just incredible and the crazi thing about all of this is that like this is open source like anyone can you know run this there's there's providers that you know you can run this with so this is something that I think is going to completely change the game because first it's their first iteration and secondly imagine what we have in 3 years when we have real open source video generation you know tools and software that is so effective so I honestly am stunned by this one genuinely like when I saw the videos of trump and Elon Musk doing certain things I was like what this is you know AI generated I was like okay okay new paradigm entered and the fact that like I said before it's locally running on people's devices is just absolutely insane so you know if you start to see videos of famous individuals doing certain things trust me guys you won't be able to tell and now I would you know put on my speculative hat even more than you may already have another thing that was released was E11 bios gamechanging leap in brain mapping cheaper faster and smarter so they basically announced a revolutionary method to map the brain at a 100 times lower cost than traditional methods and this advancement makes it feasible to study entire human and mouse brains unlocking the potential for breakthroughs and curing brain disorders and creating humanlike Ai and even simulating brains and of course you know brain mapping or connectomics helps us understand how brain works down to the connections between individual neurons and this actually could lead to building Advanced AI systems that think more like humans so they previously mapped the brain of a fruit flyer which was a pretty exciting Milestone and you know scaling to a human or even a mouse brain is just extraordinarily complex because 6,500 times larger than the fruit flies and a human brain is just like millions of times larger so you know when you're trying to map that size it it's like billions of dollars so basically one of the things that they've managed to do is reduce the cost and being able to do this is something that is really incredible because now we're on that Paradigm where we're going to be able to potentially do more mapping in terms of larger brain so this is going to be pretty incredible with uh you know the prism that they've developed and it dramatically lowers the cost by solving the AI error problem so neurons generate unique protein barcodes that basically act like IDs making them different making them easier to differentiate and proofread and this barcode system allows AI to become self- proofreading reducing the need for human involvement and it also integrates you know various techniques but when you just think about this overall this is just crazy for AI enthusiasts for those of you guys that are thinking you know we're going to get human AI soon this is going to be something that you know might lead to that so this is something that I thought was a niche story but it is AI nonetheless now of course for those of you who are wondering what else happened in AI we actually managed to get the crazy update from world laabs I did cover this in another video most people didn't watch this video but may my you're missing out because this company is doing something incredible so for those of you who are wondering what the future of VR is going to look like and how you explore certain worlds with a simple text prompt this is going to show you all that world so this thing is essentially a large World model and lwm which integrates physics semantics and spatial relationships to create realistic and immersive digital environment so this has 3D World Generation from images and it can transform a single image into a fully interactive 3D environment with smooth camera trols controls and you know you can interact with objects you know in real time so the generated World actually adhere to real physics you know spatial relationships and object interactions you know you can manipulate you know lighting you can manipulate the depth of field and it's just really crazy in terms of what you're able to do with a real that you literally just generated 5 seconds ago so this technology has various applications across game developers filmmakers I mean there's just a billion different things visual storytelling architecture and design Education and Training I mean this is the kind of thing that you know in it is infancy it's not seen like something crazy but I think once you start to realize how crazy this is going to be you're going to be like okay I understand that this moment was like the gpt2 moment for large language models we also had something crazy by Elon Musk released this image generation tool that was ### Segment 5 (20:00 - 24:00) [20:00] internally developed at x. and this was called Gro Aurora so this model you can see right here can develop images that are pretty lifelike in terms of celebrities and famous influencers and genuinely I've looked at many different AI models and most of them can't replicate people like some Alman or other individuals so it's going to be really interesting to see exactly how this stuff works because I've seen a few different things I've also seen this right here you guys can see grook Aurora celebrity likeness from this other user and I've also seen some other examples where we've got individuals like Elon Musk doing pretty hilarious things and I'm not going to lie I've seen them look super realistic so if you ever wanted to generate something that was super realistic this is of course the tool for you of course it's available on x. com you do have to you know pay that monthly membership or whatever but I do think that it's really interesting that they developed their own AI model they previously did have flux but for whatever reason they decided to develop their own image generation model and it actually has a few you know things that it was trained on things like the cybertruck the Tesla bot so if you wanted to generate those things and you didn't you weren't able to do that before this is going to be the model for you and of course we had 11 Labs introduce you know AI agents building AI agents that can speak is now easier than ever hi what would you like to talk about today introducing conversational AI with a Labs build test and deploy all in one platform for the most natural way to communicate with technology hey can I check the status of my order yes of course let me check for you now your order is due to arrive at around 20000 p. m. building conversational AI agents used to be immensely complex and demand huge resources but not anymore start building your agent by creating a voice or exploring our vast highquality library to discover one that's perfect for your application hello and welcome to dhk online how can I help you today with a wide range of the best llms to choose from you can then upload your knowledge base and Define the goal and personality of your agent or alternatively you can integrate your own server to take full control hi this is Eric from bound events how can I help hey I accidentally bought a ticket twice could I get a refund please sure please read out your reference number and I can begin to process that right now our system will analyze and evaluate your transcripts providing you with valuable insights conversation playback and the ability to test and define success criteria and it's all possible in 32 different languages welcome to DC medal how can I help hello and Best Yet effortlessly deploy your agent into websites with a single copy and paste or leverage our SDK for creating apps services or even video games so whatever you're creating get to production days not months and all with Enterprise grade security to keep your users data safe and secure it's time to explore the huge potential of conversational AI with 11 then of course one of the last things that we actually had was opening ey Sora this was something that I found to be really incredible in terms of you know just like how good this thing is like we look at the other Chinese models that have managed to surpass open ey in terms of deploying their models but when we actually take a look at what open ey is able to do in terms of the quality prompt adherence and in terms of the physics of these models somehow open ey just manages to Edge out the entire competition so this is something that you know if you're living in the UK like me unfortunately it isn't currently available unless of course you're pretty sneaky and you use avpn but this is something that I think is going to be pretty useful to a lot of creative individuals now one of the things that I actually think was probably the best thing to come from this and I say that because looking at how AI video is if you actually want to create AI films this is going to be what you use and this is going to be in the video I'm just going to scroll through here to show you guys exactly what I mean but basically I covered this in yesterday's tutorial there was something called the storyboard and the storyboard allows you to basically say I want a stalk to do this then I want the animal to do that and it's just going to be really useful in terms of you know creating fully-fledged films in an AI world so this is going to be something that I personally find to be really creative really interesting and this going to be a flood of AI videos so it will be interesting to see how people do that of course there's you know a feature like blend there's also some other features where you can you know recut there's a million different things that you can do with Sora so it's not just a video model that they've announced they've actually announced this platform and on that platform they've actually got a whole host of tools and you can really get to controlling exactly how the model works so this is going to be something that's really effective do apologize for the glitches there but if you did enjoy the video don't forget to see you guys in the next one