Everyday AI Podcast – An AI and ChatGPT Podcast - EP 278: Microsoft Build AI Recap - 5 things you need to know

Episode Date: May 22, 2024

To end a week-ish full of AI happenings, Microsoft has thrown all kinds of monkey wrenches into the GenAI race. What did they announce at their Microsoft Build conference? And how might it impact you? Our last takeaway may surprise you. Newsletter: Sign up for our free daily newsletterMore on this Episode: Episode PageJoin the discussion: Ask Jordan questions on Microsoft AIUpcoming Episodes: Check out the upcoming Everyday AI Livestream lineupWebsite: YourEverydayAI.comEmail The Show: [e-mail] with Jordan on LinkedInTopics Covered in This Episode:1. Microsoft Build Conference Key AI Features 2. Microsoft Copilot Updates3. On-device AI and its futureTimestamps:01:50 Startup Humane seeks sale amid product criticism.09:00 Using Copilot increases latency and potential errors.11:15 Copilot changing work with edge AI technology.13:51 Cloud may be more secure than personal devices.19:26 Recall technology may change required worker skills.20:24 Semantic search understands context, improving productivity.28:41 Impressive integration of GPT-4 in Copilot demo.31:41 New Copilot technology changes how we work.36:13 Customize and deploy AI agent to automate tasks.38:08 Uncertainty ahead for enterprise companies, especially Apple.46:09 Recap of 5 key announcements from build conference.Keywords:Microsoft CEO, Satya Nadella, Copilot stack, personal Copilot, team's Copilot, Copilot agents, Copilot Studio, Apple ecosystem, enterprise companies, Microsoft Teams, OpenAI, Jordan Wilson, Microsoft Build Conference, edge AI, Copilot Plus PC, recall feature, gpt4o capabilities, iPhone users, AI technology, data privacy and security, GPT 4 o desktop app, AI systems, recall, mainstream AI agents, Humane AI, Scarlett Johansson, ChatGPT, Anthropic Claude, COPilot Studio Agent, Microsoft product.Send Everyday AI and Jordan a text message. (We can't reply back unless you leave contact info) Start Here ▶️Not sure where to start when it comes to AI? Start with our Start Here Series. You can listen to the first drop -- Episode 691 -- or get free access to our Inner Cricle community and all episodes: StartHereSeries.com Also, here's a link to the entire series on a Spotify playlist. 

This is the Everyday AI Show, the everyday podcast where we simplify AI and bring its power to your fingertips. Listen daily for practical advice to boost your career, business, and everyday life. Meet Firefly AI Assistant, now live and Adobe Firefly, the all-in-one creative AI studio. Just describe what you want to create and the assistant handles the rest, orchestrating multi-step workflows across Photoshop, Premiere Express, and more in one conversational interface. You direct the outcome. The assistant accelerates execution. It's been a wild week in AI.

In just the course of nine days, we've seen pretty groundbreaking advancements and announcements from not just OpenAI in Google, but now Microsoft. So today, we're going to be talking about how I think the five ways that Microsoft is going to change the way that we worked and going over essentially the five things that you need to keep your eye on that were announced at the Microsoft Build Conference. All right, I am excited for today's show. What's going on, y'all?

My name's Jordan Wilson. I'm the host of Everyday AI. And this is for you. We are a daily live stream podcast and free daily newsletter that serves as your guide to learn generative AI. So if that sounds like you, thank you for joining us. Appreciate it as always. Make sure if you haven't already, go to your EverydayAI.com.

Sign up for that free daily newsletter. we will be recapping a lot more from the ongoing Microsoft Build conference and everything else that's going on in the world of AI. So make sure you go check that out if you haven't already. No, hey, Matthew, no, we are coming live. We are coming live today. Yeah, we had a pre-recorded show yesterday. I was actually on the road. So, hey, I love that people, you know, are following along that closely. But yeah, we're joining live here, you know, what is it, Wednesday morning. So before we talk about those five things that you need to know from the Microsoft Build Conference. Let's start as we do almost every day by going over what's happening in the world of AI News. All right. So first,

the company behind the Humane AI pin may be looking for a buyer. So Humane, a startup behind the poorly reviewed AI pin wearable computer is seeking a potential sale, according to reports for its business after facing criticism for an underdeveloped product. So Humane is a startup that's led by former Apple employees and it's seeking a possible sale for its business with a reported price tag of $750 million to a billion. So the startup has raised $230 million in funding with some pretty notable investors, including OpenAI CEO Sam Altman. So the AI PIN device was pitched as a wearable AI assistant and it's received some

pretty negative reviews for slow responses and consistent software and hardware issues, among other things. Essentially, any seasoned tech reviewer kind of said, this is a half-baked product. So it should be interesting to see if QMaine is even able to find a buyer at that valuation. All right. Next, in AI News. So we talked about this a little bit on the show on Monday, actually, but some updates. So actress Scarlett Johansson and O.E.

Open AI's Sam Altman are kind of fighting. So Scarlett Johansson released a statement and claimed that OpenAI created a voice that sounds just like her without her consent for their product, ChatGPT. She is now seeking some transparency and legal action. So to quickly recap, OpenAI did even release a statement that said that they were pulling down this Sky voice option. So, you know, chat GPT, as do a lot of large. language models have options where it can read back responses and you can choose from different voices. So there was one that was called Sky, which has since been pulled from chat GPT, I believe as of late Sunday night, early Monday morning of this week, because there were complaints of it sounding too much like Scarlett Johansson. So in this statement, you know, Scarlett Johansson said that essentially Open AI reached out to her asking if she wanted to be the voice of chat chbt chat GPT. She reportedly declined. And then Open AI kind of found, they said that this was a separate voice actor that sounded kind of pretty much exactly like her. So some ongoing battle there.

You know, who know that we would have celebrity drama in the middle of large language models? All right. And then last but not least, in AI news, a new anthropic paper breaks down how large language models work. And it's pretty noteworthy kind of discoveries, actually. So Anthropic, a leading AI company recently published a research paper aiming to explain the decision-making process of its AI chatbot clawed. So AI systems operate similar to the human brain using layered neural networks to process information and make predictions based on training data. So this whole field of AI interpretation is, you know, it's kind of a cloudy one because, you know, people are always focusing how all of these algorithms work and, you know, trying to explain, you know, AI and explain this black box.

So a pretty interesting paper here from Anthropic researchers. And they kind of utilize what's called dictionary learning to decode Claude's neural network, revealing millions of features that provide insights into the model's reasoning. So I think this paper is really going to help the rest of us understand how these models work. So this was just released a couple hours ago. I haven't had the chance to dive into it yet. It'll probably take a couple hours to read. So make sure to check that out.

So that's what we've got going on in AI News. So let's just get straight into it. And let's talk about the five things that you need to know from Microsoft's Build conference. So they are in day two of the three-day event there in Seattle. But presumably most of everything has already been announced. So I'm not going to make you wait to the end. I'm just going to give you the five things. And then we're going to go in one by one.

Ready? So five, I mean, Microsoft is bringing. edge AI or local AI to all of us with co-pilots plus PC. Number four is recall. I think that's going to be a game-changing co-pilot feature that's also kind of creepy in a privacy nightmare. Number three, I think we see a much more powerful co-pilot now. I think this is we are going to be seeing this next version of co-pilot is what they

originally announced, you know, more than a year ago. So I think we're going to finally be seeing that. Number two, I think AI agents are now going to be coming mainstream with co-pilot studio agent. And then last but not least, I think Microsoft is fighting the wrong battle when it comes to AI. All right. So let's get into it and go over. But I'm curious from our live audience joining us like Cecilia, joining us from the windy city, the very windy city. My gosh, right?

We had some tornado watch here's. All right. So, all right, Woozy said he's trying to figure out the volume. So hey, everyone else, let me know. I'm in a little bit of a different setup today. So hopefully y'all can hear me. But for our live stream audience, I would love to hear what are your thoughts on this co-pilot announcement so far from the Microsoft Build Conference?

A, maybe you didn't pay much attention. B, are you thinking it's pretty impressive? C, is it very impressive? Or D, are you just not impressed at all? I'd love to hear and get a gauge from our live stream audience. You know, we always try and, you know, bring news that matters to you and updates that matter. So, you know, I'm really curious what people think of co-pilot so far. All right.

So let's now, now that I told you all five things, let's jump into each one a little bit more. So number five, Microsoft is. is Microsoft is bringing edge AI to the masses now with copilot plus PC. So essentially this is a new kind of PC or a new classification of computer. And the biggest thing here, it is running generative AI locally. All right. And that's actually a huge deal. So they didn't actually come out and say what features will be running locally versus which ones won't.

They didn't say if it is going to be the GPT4.0. model that's going to be running locally, but it does look like that is the case. So let me just quickly explain what local AI or, you know, it has all these different names, you know, edge AI, local AI, on device AI. But, you know, it essentially means this. How we've been interacting with generative AI and large language models for the past couple of years has all been in the cloud, right? And there's obviously some advantages to that, but there's a lot of disadvantages. And there's, you know, huge pros to having this edge AI. You know, so think of it like this.

You know, right now if I go into co-pilot, right, and I type something in, I type a prompt. So it sends that to Microsoft's servers. So it sends it to the cloud. The cloud computes it, and then it sends it back to me. So there's obviously latency there, right? It increases my wait time. But it also increases the likelihood of something's going wrong, right? Like the more connections that you have, the higher the likelihood that something could go wrong, right?

Like we see this as an example, you know, if chat GPT or Google, you know, if they release new features, a lot of times it's pretty slow, right? Because their servers are getting crushed. So local AI is huge. And we haven't seen this mainly because of the size of the models and the technology just really wasn't there to be able to do this level of compute, right? So large language models are huge, both in size and their robust capabilities. And no one, at least until now, has found out how to bring a large language model to local devices, right?

At least not in mass, right? So, yes, you can download some models on, you know, hugging face as an example. Smaller models, you know, especially are much easier. So, you know, when you talk about, you know, Mata's, Lama, some of their smaller models, Microsoft Phi, mistral. So smaller models, yes, you can download and run locally. they're not integrated into your workflow. They're not integrated into your computer. So that's another big thing here, right?

It's not like you're going to just be launching one program and running, you know, a GPT model locally. It is going to be ingrained in your workspace, right? So across Microsoft's suite of products. So this cannot be understated how big this copilot plus PC could be, right? Again, this is just an announcement right now. it's marketing, you know, we haven't seen it yet. We've seen shiny demos, but we've seen that go wrong before with Google. So it should be pretty interesting to see how this ultimately plays out. And if this does, in fact, change the way that we work, right? It is very, that could be the case.

So, you know, I have said for a long time that I do think co-pilot in general is changing the way that we work. And I don't think I was wrong, right? Like I said that, you know, about eight months ago right before co-pilot came out. And right now, if you work at a larger company and you are a Windows company, which is most enterprises, there's a good chance you have co-pilot available to you, even though I do think the rollout has been a little slow. But Edge AI and Copilot plus PC could change that, right? So it obviously is going to require a higher, you know, more

capable device, right? So a lot of these devices, I believe the date that they threw out there was June 18th is when a lot of these PCs are going to be available. So yeah, it's not like you can just run co-pilot plus, you know, on an older PC. It is only going to be some of these newer ones. And it's kind of made possible with this NPU. Okay, so this neuro processing unit. So essentially, how this is going to work without getting it, without getting too far into the technical beads is it's Microsoft copilot plus PCs are going to use a combination of a standard CPU processing unit, right, graphics processing unit, GPU. And then this new technology that they essentially have created in the NPU.

And here's another good thing or another good thing to keep in mind, right? Because you might be thinking, okay, this is going to bring things locally. right? So a lot of times, one of the biggest reasons that companies aren't using copilot or aren't using generative AI as a whole is because of data and private, like privacy, right? Because right now, let's just say as an example, if you want to upload a document or into co-pilot, it's technically you are sharing it with Microsoft servers, right? You are sending it to their cloud. So the big advantage here is not having to do that, is, you know, still having a level of, you know, of data security and data privacy and being able to work with these files,

kind of quote unquote offline, right? And being able to run this very powerful, you know, generative AI technology locally on your on your hard drive with this new combination of working with the CPU, GPU, and NPU. All right. So on the flip side, right? Yeah. This is also, I think, okay, here's what I think, y'all. I think a lot of people aren't necessarily, going to understand how this works, right? And they're still going to, you know, bring in documents that they maybe shouldn't be using generative AI for, even locally.

And here's what I mean, which, which it sounds crazy to say this, but I think that sometimes cloud is more secure than your personal device, right? So let's just say, right, that that we all start putting, you know, the most important files of, like for our company. on our hard drive. It's not usually how it works, right? Those are, you know, those files normally live in a secure drive for a reason. So I do think that there's going to be a lot of decision that, that business owners or, you know, CTOs or CIOs or CMOs have to make, you know, saying like, hey, are we going to allow, you know, our employees to use this new co-pilot plus PC and, you know, this new more powerful co-pilot technology and bring some of our most, you know, our most

sensitive, our most confidential documents to their external hard drive, right? Yes, it allows them to work locally. Yes, you know, Microsoft does say that it's not sharing your data or sending your data when you're working with it on your local device, but that also does create a little bit of a security risk, right? So now all of a sudden, you know, whereas a lot of times maybe employees wouldn't have very, you know, confidential or sensitive data living on their hard drive, you know, it's on a shared company drive. Maybe they will now. So it's going to be interesting to see how that kind of impacts how we work. All right. Our next thing, number four, recall, I think is a game-changing copilot feature. People are going to be talking more

about copilot plus PC, I believe. They're going to be talking more about GPT4. Oh, I think this recall feature is going to be the one that gets people to use co-pilot and gets people, kind of the gateway drug, so to speak. I think that's going to be the one that does it, is this recall feature. So here's what recall is. It remembers everything. So think how your browser has internet history, right? Now recall, think of it as computer history, right? So it's going to change. Sorry, y'all, I got some sun blaring in the eyes there. There we go. So it is going to, I think, change how we use our computers, how we use the PC. So essentially with recall, it is going to remember every single thing that you are using within Microsoft's products. So they haven't said if you are using, you know, know, as an example, you know, Mike or sorry, Google Chrome, if it's still going to remember everything

or if it is just limited to their browser, Microsoft Edge and all of their products. But still, I mean, it's kind of crazy, right? So let's say, all right, your typical day, let's just say, right, you open up Microsoft Edge, you're browsing, browsing the web, then you open up Outlook, check your email. someone's asking for, you know, an updated PowerPoint. So then you have to go into Excel, grab the latest figures, go put it into PowerPoint. Then you upload it onto your shared team drive. Then you go into, you know, Microsoft Word and you type up a quick memo.

You should go back into Outlook, shoot that over, right? It might sound like your date. But then you're like, oh, man, what was it that they asked me for? You know, last week, right? With this recall, it literally, you can scrub, through your entire history. I also have no clue how they're going to be able to pull this off. You know, I do feel that this is going to fill up your hard drive rather quickly. But you can essentially then scroll through and just type in.

You know, let's say someone was that that imaginary presentation there was about a new, let's say you work at a coffee company and you're coming up with a new, I don't know, a new cold brew machine, right? and someone's asking about the drip rate, right? Like how fast is the drip? And you know that was in a conversation on a team's call last week, but you're like, I have no clue where it is, right? So you could just, in theory, go into recall and type in drip rate,

and it's going to bring up everything, right? It's going to bring up that conversation in the team's call. It's going to bring up the email in your outlook. It's going to bring up the facts and the presentation, right? So this is crazy, right? And I think it's going to maybe make us, well, I think it's going to make certain people much more productive. But I also think it's going to make some people a little more lazy, right? So think of even just as an example, right, since we've had AI assistance on our meetings,

you could in theory be a lot lazier. You could check out, right? Because the meeting's being recorded. There's a transcript. You can go talk to the meeting afterward, right? maybe only 10% of the meeting pertain to you, so you tune out, right? I think that, yes, this is a very powerful feature that is going to help people who want to be very good at their jobs.

It's going to help them be better. But I also think on the flip side, this is going to cause a lot of people to tune out, right, because they can just kind of go through the motions of their day-to-day work and then use recall to get all the information, which maybe isn't a bad thing, right? maybe yes, then it does just kind of free us up to focus on more creative and strategic work. But this does, I think, change what even a required or enviable skill set for a worker is, right? A lot of times, you know, the people that you might promote or even you yourself, you know, maybe you got to the point in your career that you could because you're very organized.

You know, you're always going back and reading notes. And so then when you're in a meeting, you can, you know, accurately. communicate information about a project and, you know, people see that as, oh, you're a great employee. So this might kind of change that, right? When you can now, in theory, at the snap of your finger, go in and be able to scroll through and scrub through everything that's ever happened on your computer, right? Your teams, your browsing history, your email, all with one simple search and semantic search as well, right? So it's not like, you know, if you're talking about that drip you know, that drip coffee maker, if you don't get it exactly right, right, it's going to use

semantic search and it's going to understand, oh, this is what you meant. It wasn't called that in the team's call, but this is what you meant based on context. So pretty crazy. And, you know, I do think that we are people who are going to be using this are going to work a little differently. I do think that people are going to start bookmarking or breadcrumbing, kind of key moments, right? And kind of just essentially, you know, jotting down. keywords that then they can refer to later versus, you know, either taking a bunch of notes or really being tuned into a meeting, to a presentation, or to even something you create yourself, right? Before I send something to something, right? If we have a client, you know, at everyday AI,

let's see where, let's say we're helping them on consulting. I'm typing up these documents, even though, yes, I use AI all the time. I'm still typing them up. I'm rereading them manually all the time because I like to, you know, especially as someone that talks about AI live, I like to be able to recall things or if I'm, you know, doing consults for people live, I like to be able to recall things instantly, right, no matter where they are. But, you know, maybe recall changes that. You know, when you're on a call on a pitch, on a presentation, you know, there's another feature coming that from Microsoft that they just announced that I think is going to help on this. But maybe instead, you're just instantly typing things into recall async in real

time and all of your answers are up there for you. So I do think that that one is, pretty, it's going to be a pretty game-changing feature. The flip side, this is a potential privacy nightmare. It's like, let's just be honest. Let's just be honest. Having every single thing recorded, there's a lot of people doing things probably on their computers that they don't want other people to see, right? And that could be, yes, it could be bad things, but it also could be, oh, all of a sudden,

right, let's just say you're on a work computer and you're spending, I don't know, 80% of your day and your company enables, you know, this recall feature and you're spending 80% of your day, I don't know, looking at fantasy football or you're looking at, you know, unrelated things that, you know, have nothing to do with your job, right? And who knows now? You know, I do think maybe companies are going to use this feature in a bad way, right? Like we think of the great things that something like this could do. Who knows? Maybe you might see in the future, you know, companies collecting, you know, especially with remote work, hybrid work, with AI everywhere, you might see companies once a month, who knows, having people check in and checking their devices

in and literally running through your entire recall history and being like, let's see if this employee is actually a good employee or not, right? So that's something I thought about right away. Yes, people are thinking about, you know, data privacy, data, you know, security. Absolutely. I mean, just the thought of ransomware with this thing, right? Like if this is, you know, between co-pilot plus PC, bringing, you know, I think allowing people to store or to use, you know, more, to have more powerful features to work with confidential private docs on their, on their local device. And then this recall feature, I mean, ransomware is going to be wild. But, I mean, I think employers could use this maybe in a bad way in doing routine checks. and, you know, essentially being, you know, now being overly micromanagers and being able to

literally scrub through a timeline, right? You can literally do that. Scrub through a timeline of everything that's been happening on your computer. So some weird, I think, implications that could happen down the road. And I would not be surprised to see something like that. But on the plus side, I can't wait, right? Like, let me just be honest. I'm everything Mac. I'm everything Apple, you know, with co-pilot pro.

right, you can get that on your Mac, which I think if that wasn't available, if all these copilot features weren't available on the Mac, maybe I would have already jumped over to PC. But now with this copilot plus PC with this recall feature, this might be the thing that really pushes me over to the PC side. So yeah, super excited to see what happens with recall. All right. Yeah, so there we see. Yeah, should have this up here on my screen the whole time for.

our live stream audience. And you know, I'm curious, you know, for a live stream audience, between co-pilot plus PC and recall, is this something? Do you think this is going to change how you work? Or do you think this is just going to be marketing features? So let's let's now go on to number three here. I think that we are going to see a more powerful co-pilot that just gets you. So yes, we were asking, Juan was asking here, if this, sounds like a rival to the OpenAI desktop app for Mac. Yes, so ChatGPT is rolling out with their new GPT4O for paid users, a desktop app. And yes, I did talk about this at the time that, oh, it's curious, you know, given that Microsoft has a reported 49% equity stake in OpenAI, you know, reportedly investing between $10 billion and $13 billion into the company. Yeah, it was kind of curious at the time that when Open AI had their event one week ago from Monday, so about nine days ago, that they just said that their desktop app was coming to Mac and everything that they showcased was Mac. It was all Apple products, right? I do think that is because they will be announcing, you know, Apple will be announcing Open AI as their kind of large language model or their generative AI partner for future iOS and, you know, iOS features and devices.

coming up here in June at WWDC. But the other reason, I think, Juan, to get to your question, you know, because a lot of people were like, huh, why wasn't this announcement coming to Windows? Well, I think that this number three here might just be a version of that, right? I do think that this new version of co-pilot is essentially what we are seeing in GPT4. So I'm going to go ahead. I'm going to play a quick little video here. I hope this will kind of work well for both our live stream audience and our podcast audience.

So essentially here, let me just kind of preview what's going on in this video. So someone is browsing a website in Edge and they are using kind of the co-pilot companion, right? So I think this new version of co-pilot, there's a refreshed app. It is kind of like snappable. It is very easy to use wherever you are, one click. But what we are going to see here is someone talking to the new version of co-pilot while they browse the web. All right. This is a pretty quick, pretty quick video here.

So let's go ahead and take a look, watch, and listen. All right. And actually, let me know live stream audience that you can hear this. Hopefully, hopefully you can. 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 use these shoes? It's the mountains.

It's probably going to be cold. Hey, no worries. Let me take a look. Ah, those are the Trek Star hiking sandals. While they're fantastic for summer hikes and keeping your feet cool, they might not be the best choice for cold mountains. Okay, I'm sorry to interrupt you. Pick the best thing that I should use and can you just add it to my car?

Just one product. No worries. Let's go with the Trek-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. Muchisimas gracias. I hope that me vaeer in the campamento. They nodded, Seth. Esprit of your campamento and maintain te caliente with those potas.

Bueno suerte and diviartate in those mountains. Adobe just introduced an entirely new way to create, bringing the power and precision of its creative suite into one conversational. experience. Meet Firefly AI Assistant, now live in the Adobe Firefly app, the all-in-one creative AI studio. Powered by Adobe's creative agent, Firefly AI assistant lets you start with your vision, just describe what you want, and shape the outcome as it takes form with the assistant. The assistant orchestrates multi-step workflows, drawing on 60-plus pro-grade tools across Adobe Creative Cloud apps, including Photoshop, Illustrator, Premiere, Lightroom Express, and more to help bring your ideas to life.

You can also get started with creative skills, a growing library of pre-built workflows for common creative tasks, like batch editing photos, creating mood boards, portrait retouching, and creating social variations. Every step the assistant takes is visible so you can refine, redirect, or take over at any time. You stay in the driver's seat as the creative director. Adobe Firefly AI assistant now in public beta. See it today at Firefly.

adobe.com. So pretty, pretty impressive, pretty impressive little demo there with co-pilot. So yes, there's a lot new in co-pilot, but I'd say one of the most noteworthy things is this integration of GPT4.0, right? And we saw this with the partnership in the past where Microsoft actually got or was able to use a lot of these open AI features before they, they even came out in Open AI's own products. So it will be curious if that's the case here, because even though the GPT40 announcement was about now nine days ago, they've only just now started rolling out some of these features, right?

So the biggest thing that you just saw there in this demo, I mean, if you weren't impressed, I mean, come on. Talk about a future of work. And there was another example that Microsoft did at their build conference where it was someone playing Roblox, right? It was a dad playing Roblox, and he's like, hey, you know, I don't really know anything about Roblox. You know, this is, you know, my son plays this.

What's going on here, right? So essentially when you launch co-pilot, you can just click on a window to give it access to essentially see, right? While. So we'll link that demo in our newsletter as well. So make sure to sign up at your everyday AI.com. But same thing there, right? So this assistant, just like we saw in this demo video,

was able to look at a website, read, understand, and talk with the user and talk them through it, add an item to their cart, kind of wild. Similarly, in the demo of Minecraft, right, it's able to see. And it says, oh, these guys are after you. You know, you can either run away this way or you can quickly turn around and, you know, build blocks up to get out of their path, right? To be able to see, react, and offer advice in real time, this is what I think is the future of how we work, right? And I've been saying this, this isn't new, right? I'm not just parroting back

some marketing message from Microsoft or Open AI. You know, I've been saying this now for seven, eight, nine months, right, that we are all going to have a personal AI assistant that is going to help us in every aspect of our work. And this is it right here, right? I cannot understate how big this piece is, how this new co-pilot and with the GPT40 integration being able to essentially see anything that you give it access to and it can look, it can react, it can process, it can give you advice in real time. I mean, that's huge, y'all. Like, think of all the times that you wish, right, oh man, I wish I had, you know, a mentor to help me along with this project or, hey, I wish I could ask my director, you know, these 10 questions, but I'd feel silly. I'd feel dumb.

I'd feel like, you know, that they would judge me that, you know, I wasn't prepared enough, right? I think there are so many times that we wish we had someone smarter kind of working over our shoulder that we all do, right? So yes, this is, you know, GPT40 technology. Yes, we also saw an announcement of this from Google as well with their project Astra. But here we are in the new version of co-pilot out of the box, the ability to, you know, essentially launch the co-pilot app, click on anything you want to give it access to, and then it can see, it can understand, it can reason, and you can talk to it in real time, low latency. I mean, y'all, this is, I think, the original premise of when Microsoft announced co-pilots,

you know, a year ago. You know, I don't think it was there yet, but I mean, this does change how we not just interact with AI systems, but this changes how we work, right? Yeah, it's, I do. this is the future where you are just going to, you know, if you walk into a co-working space or a library or something, you know, normally you might see someone kind of zoned in with headphones, you know, working silently. I think in the future it's just going to be people talking, right? People are going to be talking into their computers, you know, giving them access. Here's this code I'm writing. Here's this

spreadsheet. I can't figure out this, this programming piece. You know, here's a sales deck. You know, I'm going on a call. You know, how should I focus on this type of clock? what should I be telling them, right? Like we are all going to have the power of the world literally one click, one conversation away to help us on anything instantly. Pretty impressive future. All right. Let's keep this going.

Our last two here. So number two, Microsoft may bring us our first mainstream AI agents. So AI agents aren't necessarily anything new, right? So we've heard, you know, recent announcements from Google, from meta, and there's from a lot of other, you know, companies, right? Like there's Devon that's kind of more for coding AI agent or assistant, you know, Langchain has been around and been popular in the developer community for many months. But this might be the, I think, the most main.

and the first big mainstream play of AI agents. So yesterday at the Build Conference, Microsoft CEO, Sadie and Adela, kind of walked everyone through kind of this, what he called the new co-pilot stack. So essentially you have your personal co-pilot. You have your team's co-pilot team. So I'm going to talk about that here in a second as well as agents. Okay. So let's first talk a little bit about co-pilot team.

Okay. So not only do you have this new agent capability, but the team capability as well. So it's a essentially an employee, quote unquote, that you can bring anywhere. So we saw this same, this same concept from Google last week where you are literally assigning like an AI agent as a actual coworker, right? Like you're giving them a seat. You're giving them access to, you know, certain files just like you would like a remote employee. And we're seeing the same thing here with co-pilot teams. So, you know, in the example that they showed at the conference, you know, literally it's admitting co-pilot into your team's meeting, right, giving it access to certain files. You know, when you're chatting in teams, you can chat with them, right? And they have access to all of that

knowledge and all of the documents that you give them access to. So pretty big on the team side, but also on the agents, right? So there's not a lot. lot of, there's not a lot of information out about it yet. But I do believe that this is going to be one of the more popular features along with recall. But essentially from what little we know, the ability now within co-pilot studio to essentially go customize a co-pilot, you know, where you can essentially, I won't say fine-tune it, but you can kind of tweak its capabilities, give it certain knowledge, and then call on it at any point. So think of, you know, a lot of times when you go into a chat GPT or a Gemini or Claude, right,

you're just kind of working with this base model. So the way that I would say this agent is think of it more of like a custom GPT that you can build, right? And you can obviously do that within co-pilot. But then one that can actually help perform actions on your behalf, right? So let's just say as an example, you can kind of customize a co-pilot agent that is going to always attend your team's meetings, and then it's going to type up a recap, and it's going to create a quick slide presentation, and then type up an email in Outlook, right? So you can essentially

customize those agent functionalities, so kind of create it, specialize for what you want it to do, and then deploy it at scale, right? So have it act as a team member, but you can customize its capabilities and what it should be doing. So extremely powerful. powerful, right? Especially when you think, right, I always tell people, hey, how do you start using generative AI at your company? You know, I say, what are the most manual kind of not like low knowledge, but, you know, what are those most manual, low hanging fruits tasks that you do on a repetitive basis, right? And then you should be building simple workflows that help you, you know, let's say that process that I said. Let's say you're a project manager and you're attending a team's meeting

and then you're making a quick, you know, 10 slide recap. Let's say it's a huge, you know, all hands meeting where hundreds or thousands of employees attend. So then you kind of, you know, you take notes. You create a little document, a little slide, and then you email it out to everyone. Let's just say as an example, along with a, you know, a link to the recap where people can watch. Maybe that process takes you five, six, seven, eight hours. Maybe that's the majority of your job, right? Now you can build essentially a customized co-pilot.

agent that, you know, it might not do it perfectly every time, but it might get you at 90% of the way there within a couple of minutes, right? So that's why I think this new version of co-pilot is extremely enticing, right? Especially, we don't even know what's going to be coming to Apple and to Mac here in June, but I do think that this Microsoft's co-pilot studio and kind of this new co-pilot stack between personal team and agents, man, I mean, this is, if you're not already looking at Windows, looking at Microsoft, you know, let's just say your organization is all in on Mac, all in on Apple. This makes it tough. I mean, yeah, we'll see what Apple announces at their WWDC events here, I believe, you know, the second week of June. But it's going to be tough.

It's going to be tough, I think, right, for, for, for, for, for, for, for, for, for, for, for, for, for, for, for, for, for, for, companies for enterprise companies, if they are Mac, it's going to be hard to justify sticking with that eco system much longer if Apple doesn't completely blow something out of the water here with their next announcement. Which brings me to our last one. Number five, here we go. Speaking of that, kind of a hot take here, right? I had a recorded show yesterday. So I didn't get the hot take Tuesday, but I'm going to bring it right here. I think Microsoft is fighting the wrong battle, and they have been for years. But I only think now it is evident once we have seen just how far ahead they are of everyone else when it comes to their generative AI and large

language model capabilities within their operating system, within their suite of products. You know, so what we saw a lot in the developer conference, so the Microsoft Build conference, we saw a lot of this. We saw a lot of, you know, a lot of time and effort and energy go into the marketing of, you know, hey, this new co-pilot plus PC. So looking at the new Surface laptop and, you know, this Copilot Plus that runs models locally versus the newest MacBook Air. And you know what?

It got me thinking. I think Microsoft has been doing it wrong for a while. I'm being honest, right? They're trying to say, you know, our Surface laptop is better, is more capable, is more powerful than a MacBook Air. Yeah. Yeah, duh. You know, I don't understand this stance. from Microsoft, right?

If I was Microsoft, I would have been trying so hard over the last couple of years. Is it too late now? Maybe. I would have been trying so hard over the last couple of years to convince companies, individuals, everyone to untether from the Apple ecosystem. And I would have been, if I was Microsoft, I would have been paying so much more attention, marketing dollars, everything else about, hey, how you can stay. how you can still pair an Apple, you know, an iPhone and how you can still integrate it into the

Windows ecosystem, into Microsoft, right? I'm not a Microsoft user. I used to actually build PCs back when I was a teenager. I was actually a Microsoft certified professional like, I don't know, 22 years ago. But I don't think that they put in enough effort. I think that they've, Microsoft has really just been able to build products and software in a silo, completely neglecting, you know, any capability, any functionality for people who are using iPhones, right? That's the thing right now. There's, there's no comparison for the iPhone. And I think of even myself in how I work, right? And I'm probably on the extreme side of this. So I have an Apple iPhone. I have a Mac. I have a watch, right? I have all the Apple stuff.

And I use it because it all works so seamlessly together. And I know that if you are in kind of the Windows ecosystem, you get that functionality as well. But here's the reality. iPhone and Apple are not losing market share anytime soon. So I think Microsoft is spending all of this, all of their marketing dollars trying to compare, you know, hey, our surface laptop is better than the MacBook Air. Yeah, duh. It's not even close right now. Right? Because right now, Microsoft has had a year plus head start on everything generative AI versus Apple, right? Apple, their announcement in June, presumably they're going to be working with Open AI. Guess what? Microsoft is light years ahead. So I think even the absolute best case scenario for whatever Apple announces here in June, the absolute best case scenario, right? Like we've seen reports of like, oh, it's going to help.

you know, summarize your text messages, which I would love, you know, summarize web pages in Safari, great, right? But absolute best case for Apple and for Macs, you know, looking at the end of 2024, is there a year behind? They're a year behind where PCs are, especially when you look at the new kind of co-pilot plus PCs. Y'all, if I was Microsoft, I wouldn't be fighting that battle. I would be spending more and more time trying to develop new solutions that make it easier for people with iPhones to essentially work on a PC, period. That's the wrong battle. Is it even possible? I don't know. I know Apple and Mac, you know, all of their, you know, software programs are very proprietary. It's hard to, you know, you talk about like blue tax versus

green tax, right? But I think Microsoft is fighting the wrong battle here. You saw this over and over and over again, these comparisons to, you know, the co-pilot PC Plus versus, you know, MacBook Air, it's like, okay, that does nothing. That does nothing. I think what Microsoft has to do or what they should have been doing for the last couple of years, maybe it's too little too late, is convince iPhone users to untether, untether from the Apple ecosystem. Hey, if you have an iPhone, here's how you can still work great on our devices, right? Like an example is, you know, Google Photos, right? That makes it easier for Apple users to get into this Google ecosystem of software, of services, right? Hey, unlimited, unlimited cloud storage of your Google Photos.

What a great entryway into Google's suite in products. I don't think you see that from Microsoft, right? I don't think you see it, right? Like, why is there no, you know, all-encompassing way for iPhone users or for people who are, you know, kind of quote unquote, I won't even say it stuck. But now I kind of feel stuck if I'm being honest. I kind of feel stuck in the Apple ecosystem, in the Mac. Like, right? And it's like, I want all this. But everything I have, right, I love being able to, as an example, text message from my computer, you know, with people who just have I message, which you can't do right now, you know, on the window side.

So I do think that in the build conference, Microsoft spent. so much time, you know, essentially fighting a battle that I think anyone who can read, anyone who can decipher information, they already understand that they're winning, right? The hardware's better. The software with generative AI, with co-pilot has the potential, if it works well, has the potential to be exponentially better. So I don't know why they are fighting a battle that I think they already won. Yeah, like Rolando is saying, I love this, said, sounds like Microsoft needs to watch

Simon Sinex TED Talk start with why. Yeah, why? Why are you comparing yourself to MacBook Air? It makes no sense. All right. So to recap, y'all, let's go ahead and recap the five things that I think you need to take away, the five things that matter from this new announcements from the build conference. Again, we're in day two of day three.

it's wrapping up. But first and foremost, you know, kind of point number five is we are getting edge AI with copilot plus PC, kind of the first big mainstream of bringing generative AI capabilities with large language models to the desktop, bringing it online off the cloud to the device. Number four, recall, which I think is a game changing feature for how we work, essentially like internet history, but for everything that is happening on your computer. Number three is just a much more powerful co-pilot with GPT4O capabilities and the ability to kind of act with an agent in real time on no matter what you're working on with one click. AI agents and co-pilot studio is our second to last one there.

Just being able to build your own co-pilots and then to bring in this agent capabilities, which we've seen announcements, but Microsoft, a lot of these features are reportedly going to be rolling out June 18th. They could be the first major player in the AI agent space. And then last but not least, number one, I think Microsoft is just fighting the wrong battle. I think they are so far ahead, yet they are still obsessed with just comparing, you know, apples to apples.

I think they're fighting the wrong battle. They should be fighting to get iPhone users into the Microsoft Windows and co-pilot PC ecosystem of products and services. I think they're fighting the wrong battle. They're so far ahead. All right. That is it for today. I know a longer show.

There was a lot here. If this was helpful, I'd appreciate it. If you're listening in on the podcast, please leave us a review and rating, we come to you guys live, almost every single day. This is unedited,

unscripted, right? I think that's important, right? When we're learning about AI, there's so much fake, so much manufactured, so much marketing out here. So that's what we try to do at everyday AI. We try to break it all down for everyday people.

So if you don't have a PhD in machine learning, I hope that every day, Day AI podcast, live stream, and newsletter is helpful for you. So if you could, leave us a review. If you could share this, we'd appreciate it. And also go to your everyday AI.com, sign up for that free daily newsletter. Thank you for tuning in. I hope to see you back tomorrow and every day for more.

Everyday AI. Thanks y'all. Meet Firefly AI assistant. Now live in Adobe Firefly, the Allman One Creative AI Studio. Just describe what you want to create in your own words and the assistant handles the rest, orchestrating multi-step workflows across Adobe Creative Creative cloud apps, including Photoshop, Premiere Express, and more in one conversational interface. You direct the outcome while the assistant accelerates execution.

Stand control with the ability to step in and refine at any time. See it today at firefly.adobie.com. And that's a wrap for today's edition of Everyday AI. Thanks for joining us. If you enjoyed this episode, please subscribe and leave us a rating. It helps keep us going. For a little more AI magic, visit Your Everyday A. and sign up to our daily newsletter so you don't get left behind.

Go break some barriers and we'll see you next time.