Microsoft MASSIVE Announcements: GPT-5, Copilot+ PC, Phi-3, Devin Partnership

Microsoft's build event is happening right now and they had their first day of Keynotes yesterday they announced a ton of new features product launches and almost all of it was related to Ai and day one is a 9h hour video so let's watch the whole thing together just kidding we'll watch the most interesting parts and I'll comment on it and I want to hear what you think in the comments below as well so let's watch all right so Satia Nadella obviously makes an appearance Sam Alman makes an appearance all right so first up sa and Nadella CEO of Microsoft makes his opening keynote good morning good morning it's fantastic to be back and by the way I believe that Saia Adela is top three CEOs potentially of all time but certainly of this generation of CEOs especially in his current form he is investing so heavily in Ai and not just opening AI he is investing across the board he's partnering with meta for open source he's investing in a ton of companies he's building out data centers for AI so he is putting Microsoft all in on AI and it will pay off in my opinion all right so he talks about scaling laws at this section and I find it pretty interesting to see the acceleration of compute that's happening right now so let me just play this little clip quickly of dnns are really along with the model architecture interesting way ways to use data generate data um are really driving this intelligence Revolution um you could say Moors law was probably you know more stable in the sense that it was scaling at maybe 15 months 18 months uh we now have these things that are scaling every 6 months or doubling every you know six months you know what we have though with the effect of these scaling laws is a new natural user interface that's multimodal that means supports text speech images video as input and output we have memory that retains important context recalls both our personal knowledge and data across our apps and devices we have new reasoning and planning capabilities that helps us understand very complex context and complete complex tasks right while reducing the cognitive load on us all right now he's going to talk about their three platforms that they've just built or they've been building one of which is Microsoft co-pilot which is essentially a built into every layer of the Windows operating system and Microsoft is going to be talking about co-pilot plus as well but let's see what he has to say about this is Microsoft co-pilot which is your everyday uh companion it puts knowledge and expertise at your fingertips helps you act on it and we built the co-pilot stack so that you can build your AI applications and solutions and experiences and just yesterday we introduced a new category of co-pilot plus PCS the fastest AI first PC all right so let's pause for a second co-pilot plus PC is really interesting these are fully featured laptops or computers in general that have special chips to accelerate local large language models but it's interesting and again I've said this from the beginning Microsoft is really investing in everything AI related so closed source with open AI open source with meta they're investing in Silicon in the hardware itself that accelerates on device large language models so again across the board just really putting bets on every single piece of the AI infrastructure all three of these things are exciting um platforms but I want to start with co-pilot plus PCS you know we are exposing AI as a first class name space for Windows this week we are introducing the windows co-pilot runtime to make Windows the best platform for you to be able to build your AI applications yeah W you know what win32 was to graphical user interface we believe the windows co-pilot runtime will free for AI that's kind of insane to think about like pause for a second he is saying the graphical user interface for the terminal what was really just a terminal before is the same level as we're seeing now that AI will be to our current Computing kind of the way we interact with computers so he is extremely bullish to say the least on AI and they are making it a first class citizen in Windows and I have some videos coming about this so stay tuned for that Windows co-pilot Library a collection of these ready to use local apis that help you integrate into your new experiences all of the AI capabilities that we and I think I'm going to be talking about this through the entire video but again they are investing heavily in local open source that is what fi is p hi that is a local open-source model and I love that they are investing heavily in that in parallel to close Source model so really bolstering both sides of the equation all right now he's going to talk about the different ways to access models and this is what I've been saying since the beginning of the video so let's take a look but of course if you want to access these models itself right you can directly call them through apis we have 40 plus models available out of the box including F silica our newest member of our small language family model which we can specific which we specifically designed to run locally on your npus on co-pilot plus PCS and if you heard that he called it an npu not a GPU and we'll talk about that a little bit later bringing that lightning fast local inference to the device you know the other thing is the copilot Library also makes it easy for you to incorporate rag inside of your applications on with on device data right it gives you the right tools to build a a vector store within your your app it enables you to do that semantic search that you saw with recall but now you can okay recall is something else I'm going to be showing you later in this video essentially what it is if you've seen rewind AI it basically memorizes or looks at everything you're doing on your PC which I don't know if I love to be honest but hopefully they have some good privacy settings and deletion settings but it can recall everything you do so it's super helpful if you're okay with that compromise of privacy okay now goes on to talk about some developer focused products that are launching it's it's quite technical I'm going to skip over it for the purpose of this video all right and Nvidia and Microsoft are partnering together as we kind of thought they would uh they have a strong partnership Nvidia is delivering a lot of the hardware obviously but also a lot of the software to fine-tune your models to train your models Cuda so it'll be really interesting and not investment advice but boy Microsoft is really firing in all cylinders right now all right now he's going to say a little bit about open Ai and how important they are to Microsoft so let's take a look and it all starts though with our most strategic and most important partnership with open AI honestly I think he's saying that because he's trying to appease Sam Alman if I were Sam Alman running open Ai and my partner who owns 49% of my company is out investing in direct competitors to me meta right meta with llama and the open source movement I would be a little upset now obious obviously I would not be in a position to make a closed Source company but assuming I were I would be upset okay now here's a really cool example of copilot on Windows and this is a demo somebody's playing Minecraft co-pilot can read the screen and give that person tips on how to play Minecraft now if this sounds familiar it is essentially what GPT 40 launched but it's coming sooner to co-pilot and it's branded as co-pilot so it's really interesting how Microsoft is playing this relationship with open AI this is a essentially the chat GPT desktop app again just branded co-pilot but it has all the same functionality it can read your screen it can communicate in a very natural voice let's watch this demo hey co-pilot how's it going hey I'm doing great and it looks like you're about to dive into some Minecraft ready to build explore and maybe dodge a few moms yeah I'm trying to figure out how to make a sword my son plays this game all the time and I have no idea what I'm doing no worries I'm here to help you impress your son I see you've got a crafting table in front of you which is a great start to craft a sword you'll need some materials can you open your inventory just press e on your keyboard yeah do I have what I need let's see you've got some sticks perfect for the swords handle but it looks like you're missing the material for the blade you can use wood Stone iron gold or Diamond let's start by gathering some wood or stone shall we head out and find some resources yeah that sounds great oh my God what's that oh no it's a zombie all right so I'm making a big assumption that this is all at 1X speed and unedited it does look that way and the latency is very very fast which is very impressive and overall it's just very impressive you can have this artificial intelligence assistant helping you with everything you're doing it doesn't have to be a game it could be your work it could be emails it could be everything this this is a really compelling offering from Microsoft for people to use Windows run you need to get away from it fast either Sprint away or quickly build a tower of blocks straight up to get out of its reach if you can find a shelter or dig into the side of a hill do it now just make sure it can't reach you all right I think I might be safe now that was a close one great job finding shelter did I just notice they added Echo to The Voice or is that just Reverb I'm a little confused that would be incredibly impressive if they actually added Echo to the voice because the player is in a cave I don't know let me know what you think am I overanalyzing this all right here is another demo that we'll watch I believe this one is going to be shopping which you know that's not my favorite use case to show off as a demo but let's watch it anyways all right let me ask the agent here hey Seth how's it going what can I help you with today I am fantastic I'm sorry I'm in such a hurry I'm going to go on an overnight camping hike where we're camping all this it's going to be dark cold should I should I use these shoes it's the mountains it's probably going to be cold he no worries let me take a look ah those are the trar hiking sandals while they're fantastic for summer hikes and keeping your feet cool they might not be the best choice for Co oh my good okay I'm sorry to interrupt you pick the best thing that I yeah so this is a lot of what we've seen already with GPT 40 he interrupts the model that is a new feature that I haven't had access to yet because we don't have access to the voice model but I can't wait to play with it in co-pilot and GPT 40 uh but I just don't have access yet but this is a lot of what we've seen I should use and can you just add it to my C just one product no worries let's go with the Tre ready Hawking boots they're durable provide great ankle support and most importantly they'll keep your feet warm in the cold mountains I'll add them to your cart now ready for check out I see that amazing okay so wow very impressive the AI can actually control your screen that is something that really only Microsoft could have enabled it is AI as a first class Citizen and it was actually able to control this user's screen this is really the future of how humans are going to interact with computers for the most part we're just going to talk to an agent and the agent is going to perform actions on our behalf okay so then without saying anything he switches to Spanish and so does the AI model again without saying a word about it not hey I want to speak in Spanish going forward he just started speaking and the model got it very cool I want to thank the sponsor of this video the open Institute of Technology opit is an online fully EU accredited higher education institution whose mission is to train the next generation of Technology leaders by offering highquality modern affordable degrees in this field for those who are only starting their path in Tech opit offers bachelor's degrees in digital business and modern computer science and for those who want to gain a deeper level of understanding of these topics and already have a bachelor's degree in any field opit offers master's degrees in applied data science and AI digital business Enterprise cyber security and responsible AI access courses anytime anywhere with live sessions and plenty of Community Support get career support internships and affordable pricing with eects recognition so check it out apply now at opt.com and now back to the video all right now I just want to show this screen for a second this really shows how they are absolutely investing in everybody and including everybody not just open AI so they have the desm model they have the mistal large here data bricks model which is fantastic I have a video that I tested that the cohere command R coher embed metal Lama three snowflake 53 mini more mistel more F so cool and I haven't seen the 53 Vision yet I saw 53 medium was just announced and I do have that downloaded and I'm going to be testing that soon but 53 Vision I can't wait to test that all right let's keep watching and yeah here it is Microsoft loves open source so here's an expanded partnership with hugging face and some of the features of the partnership additional models come in Azure AI deeper integration with Azure AI studio and TGI enabled for optimized run time so a little bit about that let's go ahead now and here is their 53 benchmarks so 53 medium the model quality is at about 78 higher than all the other models and then in terms of the size it sits right in the middle so here's Gemma 2 of course a little ding at Google here's llama 38b mixol which is just about the same size but worse in quality now what's interesting is the 53 small model is almost the same as the 53 medium in terms of quality but it is much smaller so I'm really impressed with the 53 model all right now he's going to talk about some additions to the 53 family let's watch that uh and today we are adding new models to the 53 family uh to add even more flexibility across that quality cost curve uh we're introducing 53 vision of 4.2 billion parameter multimodal model with language and vision capabilities it can be used to reason over real world images or generate insights and answer questions about images as you can see right here yeah and we're also making a 7 billion parameter 53 small and a 14 billion parameter 53 medium models available uh with Fi you can build apps that span the web your Android iOS windows and the edge uh they can take advantage of local hardware when available and fall back on the cloud we're not simplifying really all of what vs developers have to do to support multiple platforms using one AI model all right now he's going to talk a little bit about GitHub co-pilot and I want to preface with Devon the close Source Devon just announced a partnership with Microsoft that it is going to be powering a lot of development productivity increases through that partnership and integration and again Microsoft is kind of betting everywhere GitHub co-pilot has this new workspaces product which is directly competitive with Devon but now they support both all right let's watch was the first I would say hit product of this generative AI age uh and it's the most widely adopted AI developer tools 1.8 million Subs across 50,000 organizations are using it yeah GitHub co-pilot really changed the way people code including myself completely completely changed it then when chat gbt came out it really just made it so to be honest I'm not writing a lot of code from scratch anymore and I don't regret that at all all right so let's talk about co-pilot a little bit more he's going to give a few other demos about how it's going to be helpful and I'm really excited about it I may actually end up using my window machine more often now it can be in Loop it can be in planner and many any many other places someone think about it right it can be your meeting facilitator uh when you're in teams creating agendas tracking time taking notes for you or a collaborator writing chats surfacing the most important information tracking action items addressing unresolved issues um and it can even be your project manager ensuring that every project that you're working on as a team is running smoothly project managers across the world just shuddered all right the next segment I want to talk about is Microsoft CTO which is who you're seeing here bringing out Sam mman and a talk with him and there's some interesting points in here so let's watch it together do the next round of amazing things with him and so with that I'd like to bring Sam Alman to the stage hey good to see you you too so uh you are one of the busiest people on the planet wild week it's yeah it's a wild week it's a wild year man um but so I I really appreciate you taking uh time out to chat with us today um so I I guess what I really wanted to start our conversation uh about and like I asked you this question last week is you know there there's just been an extraordinary amount of change over so I find this conversation to actually be really interesting so I'm going to do kind of a super cut of just this segment of the conversation between Sam Alman and Microsoft's CTO so I'm going to play that now yeah there's just been an extraordinary amount of change over the past year and a half year uh like what has been the thing that has surprised you most uh particularly relevant to an audience of developers I mean I'm delighted to be here uh and obviously great to see you but developers have been such a core part of what's been happening this last year and a half um there's millions of people building on the platform what people are doing is totally amazing and the speed of adoption and talent and figuring out what to build with all of this over what has really not been very long like when we put gpt3 out in the API uh some people thought it was cool but it was narrow where the happened and seeing what people have done with gp4 and seeing now what's happening with gpg 40 even though it's new and hasn't been out that long uh is quite remarkable I've never seen a technology get adopted so quickly in such a meaningful way uh the what people are building how people are finding out how to do things that we never even thought of possible which is why it's always great to have an API uh that's been very cool to see there's a version of AI that could have existed that is uh you know like a bunch of smart people like building uh you know things at extraordinary scale and then just building it into a bunch of products where everybody gets to passively use them like the the really brilliant thing that you all have done is like taken the exact same set of things and like decided to make it available to like any developer who's able to sign up for an API key yeah we we try to be really thoughtful about what makes a good API for this there's going to be all kinds ways people can use this but the more this can just be a layer that gets built into every product every Service uh the better and we've tried to make it such that if you want to add intelligence to whatever you doing uh any product any service we make that very easy what are the category of things that people should be expecting over the next you know K months so the the most important thing and this sounds like the most boring obvious Tri thing I can say but I think it's actually much deeper than than it sounds the most important thing is that the models are just going to get smarter generally across the board there will be a lot of other things to which we can talk about but if you think about what happened from GPT 3 to 3.5 to 4 it just got smarter and you could use it for all these things it got a little more robust it got much safer uh both because the model got smarter and we put much more work into building the safety tools around it um it got more useful but the underlying cap capability this amazing emergent property of like we actually are seeming to increase the general capability of the model across the board that's going to keep happening and the the jump that we have seen in the utility that a model can deliver with each of those halfstep jumps and smartness it's quite significant each time so as we think about the next model and the next one and the incredible things that developers are going to build with that I think that's the most important thing to keep in mind uh also speed and cost really matter to us so with GPT 40 we were able to bring the price down by half and double the the speed um new modalities really matter uh voice mode has been actually a genuine surprise for me in how much I like the new voice mode and I when people start integrating that I think that'll matter but but it's the overall intelligence that'll be coming that I think matters the most so one last thing before we let you go so you know you and I and like members of your team and members of the Microsoft team have been doing really an extraordinary volume of work over the past uh year and a half two years thinking about safe deployment of an awful lot of AI capability like everything from you know apis and developer tools to end products uh and you know I think we you know have accumulated a really interesting volume of experience like experience is sort of hard to get if you're not doing deployments at this scale um so I I you know and I think you just mentioned something that's like really really really interesting like part of uh you know part of the interesting and surprising progression of capabilities of these models means that they're more useful in like helping to like make AI systems safer so I I don't know whether you had some thoughts you wanted to share there as well you know when we first developed this technology we spent a lot of time talking about all right we've made this thing it's cool are we ever going to be able to get it to an acceptable level of robustness and safety and now we kind of take that for granted with gp4 um you know if you use it it's far from perfect we have more work to do but it is generally considered robust enough and safe enough for a wide variety of uses and that took an enormous amount of work across both teams and fundamental research like when we started this we like we've got this thing we've got this language model it looks like kind of impressive and kind of not and even then how are we going to like get it aligned and um what it what do it need you know what is it going to take to be able to deploy it the number of different teams we've had to build up uh to go from research and creation of the model to Safety Systems to figuring out policy to how we do the monitoring um that's a huge amount of work but it's it's necessary uh to be able to deploy these and use them like you know when you take a medicine you want to know it's going to be safe when you use an AI model you want to know it's going to be robust and behave the way you want and have been super proud of the work that teams have done together and I think it's amazing how fast this much work has happened and that we can all now use this and say oh yeah basically it basically works as the models get more powerful there will be many new things we have to figure out as we move towards AGI um the level of complexity and I think the new research that it'll take will increase I'm sure we'll do that together but we view this as a gate on being able to put these things out into the world which we really want to do yeah it's definitely table sakes so thank you so much for uh for being with us here today like I really appreciate your time uh it's awesome to hear from you awesome okay so those were the highlights of day one of Microsoft build event I think that the Windows operating system is becoming more and more compelling at a rapid pace and honestly Apple where are you I want to see some AI functionality being built into apple and I think it's going to happen this year I think they're going to make some big announcements and really show us the next evolution of their operating systems not only Mac OS but IOS as well so you know I'm going to be following closely I hope you enjoyed this video if you liked it please consider giving a like And subscribe and I'll see you in the next one

Channel: Matthew Berman

Views: 100,522

Keywords: ai, microsoft, satya nadella, openai, sam altman, microsoft build, msft, llm, phi, recall, copilot

Id: 6H8NPVGC6Ak

Length: 25min 0sec (1500 seconds)

Published: Thu May 23 2024