Everyday AI Podcast – An AI and ChatGPT Podcast - Ep 272: How OpenAI’s Shot at Google (and GPT-4o model) Will Change How We All Work

Episode Date: May 14, 2024

GPT-4o looks cool. No doubt. But, here's what you probably aren't thinking about.  ↳ It was definitely a shot at Google ↳ It will likely change the way the most efficient organizations work forever ↳ It is actually just a super powerful Trojan Horse  What's all that mean? We break down GPT-4o and what it means for us all. Newsletter: Sign up for our free daily newsletterMore on this Episode: Episode PageJoin the discussion: Ask Jordan questions on GPT-4oRelated Episodes: OpenAI Releases GPT-4o: 12 things you need to knowUpcoming Episodes: Check out the upcoming Everyday AI Livestream lineupWebsite: YourEverydayAI.comEmail The Show: [e-mail] with Jordan on LinkedInTopics Covered in This Episode:1. Future impacts of GPT-4o2. OpenAI, Microsoft and Apple Collaboration3. Live demonstrations of GPT-4o4. OpenAI's strategic releases5. New Features and Plans for GPT-4oTimestamps:01:45 Daily AI news06:12 Generative AI update widens efficiency gap.10:04 GPT-4o combines transcription, intelligence, and text to speech.10:49 New GPT-4o assistant sees, hears, reduced cost.15:18 Eminem-inspired reverse reveal and criticism by OpenAI.17:41 Exclusive GPT-4o features will differentiate tiers.21:45 Google's Gemini rollout led to widespread mistrust.25:02 Large language models improve with usage statistics.27:32 Prediction of future technology transforming work processes.33:00 Teenager learning on iPad with real-time assistance.34:52 Discussing potential AGI, questioning if it's present.38:39 Latency reduced with new unified GPT-4 model.Keywords:Jordan Wilson, artificial general intelligence, AGI, OpenAI, Microsoft, Apple, GPT-4o Omni model, generative AI tools, live giveaway, YouTube comments, future of work, impact of technology, desktop app, AI assistant, coding, iPad app, access to education, GPT-4 model, US-China Geneva meeting, Klarna, Google IO developer conference, rock paper scissors demo, deceptive marketing strategy, training data, technology development, Microsoft's Copilot, GPT-4o, API cost, voice communication, Google's Gemini modelSend 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. The future of work was debuted yesterday.

I'm not just talking about the new GPT40 model from OpenAI. I'm talking about how this model can be used. And the reported kind of partnership or coming together of three of the biggest names in tech. and how this kind of unlikely but connected partnership of these three big companies is likely going to change the way that we interact with technology in the future. All right. So we're going to be talking about that today and more on Everyday AI. What's going on, y'all?

My name's Jordan Wilson. I'm the host, and Everyday AI is for you. This is a daily live stream podcast and free daily newsletter that serves as your guide to how you can leverage generative AI to grow your company. and to grow your careers. I hope that's you. So thanks for showing up. If you're listening on the podcast,

make sure to check out your show notes, as always, and go to your everyday AI.com, not just for a recap of today's show later in the newsletter, but you can visit an entire library of great and free hundreds of hours of free generative AI content from experts in the field on our website. So make sure you go check that out. All right.

So before we get into how I really think this, new announcement from OpenAI yesterday and their GPT40 model is going to change the way we work. Let's first start as we do every day by looking at the AI news. All right. So the U.S. and China will be meeting on AI in Geneva. So high level envoys from the U.S. and China will meet in Geneva to discuss the risk and shared standards of artificial intelligence and the first meeting under an intergovernmental dialogue on AI. Both countries view AI as crucial for national security and economic growth.

So the U.S. plans to focus on developing safe, secure, and trustworthy AI through voluntary commitments with leading companies and safety tests of AI products. China and the U.S. will reportedly today take up issues, including technological risks and global governance of AI during the meeting. The talks, while they're looking, both sides, are looking to build trust and understanding between the two countries on AI issues. but immediate binding agreements are not expected. All right, our next piece of AI news for the day.

Klarna is all in on AI. So the Swedish fintech company has reported that almost 90% of its employees are now using generative AI tools in their daily work with a high adoption rate among even non-technical groups. So the use of AI has been touted as a major boon to the company's bottom line with its first quarterly profit in four years attributed to investment. in AI. So a couple kind of key takeaways here, Klarna reported the high adoption of generative AI tools among employees with over 87% using tools such as OpenAI's chat GPT and its own

internal AI assistant. AI has been especially beneficial for the company's communication and legal teams with tasks such as evaluating press articles and drafting contracts made much more efficient. Klarna has seen success in implementing AI in its business with its first quarterly profit in four years and $40 million in reported savings from its AI chatbot. Yeah, so we talk about that. People are always like, oh, how would I use this? Well, there you go. Karna reported that their usage has saved them $40 million in employees' time.

Last but not least, you know, hot off the heels of Open AI's announcement. Today, we are expecting Google's announcements at their Google IO developer conference as it kicks off in hours and the focus will obviously be on AI. So Google's annual developer conference, Google I.O, will take place today and tomorrow in Mountain View, California. So the event is expected to focus on artificial intelligence with potential announcements regarding AI, apps and services, and possibly even debuting some new hardware. Google is expected to make major announcements regarding AI, potentially updating its Gemini model and some other AI updates. Also, Google already teased a live multimole. Votal AI video assistant, similar to what Open AI announced yesterday.

We'll be talking about that more today. All right. So let's get into it, y'all. Thank you for joining us live. And we're doing something special today. Well, something different anyways for our live stream audience. So just go ahead. We're going to do a giveaway at the end of the show.

I'll tell you what it is then. So just go ahead and use the hashtag Hot Take Tuesday in your comments. And then we're going to be drawing someone live that comment. So we're trying trying something new today. So go ahead and give that a try. And, you know, hey, if you're on the podcast, make sure to come join the live stream every once in a while. It's a lot of fun.

I also want to know from our audience, what do you think of the new Open AIs GPT40 model? So A, do you not know anything about it? B, are you not super impressed? C, are you pretty impressed? Or D, are you blown away? All right. So let's just start.

I'm going to give you our hot takes. It is hot take Tuesday after all. So here are my thoughts. And then we're going to be diving into this in-depth here a little bit. So I'll say this is our first mainstream taste of AI agents. Yeah, this is it. I'm going to describe what that means to you here in a second. But, you know, we've been talking about AI agents here on the show for probably a year plus, right?

Going back to the very beginning, I think this is our first taste of actually working with an AI agent and not just typing something to a large language model and hoping that you get a good response. I think this is also a big step toward AGI, artificial general intelligence. I don't think people want to admit this, but when you kind of see at least what was previewed yesterday, not all of this is out yet, I think you'll understand what I mean on that. Also, I think this is actually going to create huge gaps between whether it's companies, departments who are using AI versus not using AI. You know, as an example, we talked about Klarna, you know, getting this 90% of their employees using generative AI.

I actually think this newest update is going to drive a divide in the gap between people who are using it and those who are not, in companies who are using it and companies who are not. So both in efficiency and productivity, you know, that's the downside that people don't talk about generative AI and all of its great potential and upside in helping us be more efficient, more productive, et cetera, is those not using it, right? There's still going to be companies and people and individuals and departments who are not using generative AI.

And I think that this latest update is going to make that efficiency gap, that skill gap, much wider, which is actually extremely problematic for people. And I don't think enough people are realizing this or talking about it. It is a literacy issue, is an educational issue. But as we take a look at this GPT40 announcement from OpenAI, I think that will make sense. And then last but not least, Google is in big trouble. They are in huge trouble. So we're going to talk a little bit.

So, you know, obviously it was reported over the last day or two that Open AI and Apple are going to be working together and that Apple will reportedly be using OpenAI's models in their next iOS 18 that will bring, for the first time, AI to all of their billions of devices around the globe, which is huge. Obviously, we already know about the Microsoft and Open AI partnership. So you kind of have this unofficial now, Big Three, this unofficial super team. So combine that with the new model that we saw from OpenAI, the accessibility, them making major aspects of it free.

Google's in trouble, y'all. Like, I know they have some announcements here in a couple of hours, and they're essentially going to be debuting a lot of things that Open AI did yesterday, but I think they're in trouble. That's me. All right, so 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.adobie.com. Let's first do a very quick recap. And hey, thanks. Thanks for everyone joining.

in, you know, trying to get into this hot take Tuesday. Hey, make sure you get it all in one, all in one kind of sentence here, Rolando. I want to make sure everyone who's putting it in gets entered. It's just a software. So if you put a space, it's not going to grab it. All right. So let's go ahead. Let's go over a quick recap of what actually happened yesterday. We did an entire episode on this. We did actually two episodes yesterday. So I'm not going to take a super long time. I just want to do a very high level and quick recap for those of you that maybe missed it. All right. So the new version of GPT4 is called GPT40, which stands for Omni Model. All right. GPT40 is already available right now to paid users and it will be going out to free users as well.

I checked last night. It wasn't available to free users yet. Could be as of this morning. So make sure if you do have access to it on the free account, let me know. the most powerful model as of now will be available to free users and paid users as well. Right now, paid users will have five times the capacity limit as free users. So it will kind of be throttled and capped for free users. And paid users won't have that same limitation. Number four, even free users will soon be able to access the GVT store. That is huge.

All right. And I'll probably have more on that in another episode. But yes, even free users to chat GBT will be able to use GVT. So these custom GPTs that anyone will build, free users cannot build them, but they can use them, which I think really changes how chat GBT will ultimately be used in a team environment. Another one, GPD 4-0 combines transcription, intelligence, and text-to-speech all in one mode. Whereas before, you kind of behind the scenes had the different modes working with each other, which created some latency, which, hey, if you want to talk about a helpful agent, an AI agent, you can't really have a lot of latency on multiple ends. So this new GPT4O model changes that. So we are going to see a new, so kind of

0.6 here. We're going to see a new desktop assistant that can hear and see what you're working on. And I'm going to show you guys a demo of that here in a second. Seven, GPD4O is rolling out to the API at a reduced cost. It is 50% the cost of the previous version of GPT4 Turbo's API. So it is twice as fast and half the cost. All right. So that's going to change a lot for the tens of thousands of, you know, other products and services that use OpenAIs API. Number eight, OpenAI demoed a live view mode,

presumably being able to use vision in real time, which is huge. And a lot of people are confused on like, hey, versus like the free versus the paid. If this GPD40 is going out to all free users, why would I remain a paid user? well because these features that haven't been rolled out yet, specifically kind of this live view mode will not, at least right now, will not be available to free users.

So there are features that have not been rolled out and they will be rolling out in the coming weeks that are only available to paid users. So a lot of people kind of miss that and they're like, oh, what's the point? Well, that's the point. All right. Number nine things you need to know here is there's a reduced latency with a real time feel in voice-to-voice communication. Number 10, it is a much more human feel.

I mean, the emotion and kind of this range in the voice of the AI agent is pretty outstanding. Is it scary? Yes, it is actually scary how emotive this new GPT40 voice assistant is and how responsive it is as well. Number 11, like I said, a lot of these newer features, aside from the actual base model, are going to be rolling out to paid users in the coming weeks, right? So we're going to show some demos of this here in a second from Open AI. And last but not least, hey, I already talked about this, but Google, I think, is in big trouble, is in big trouble.

All right. So let's talk about that first. All right. Let's talk about why Google is in trouble. And, hey, Woozy with a comment here is picking up on the point I'm about to make. Yeah, they were using. They were using Macs and Apple everywhere, which is interesting, right? Because reportedly, right, well, not reportedly, but Microsoft has a huge equity stake in OpenAI, right?

Reportedly, they've invested between $10 billion and $13 billion and have a 49% ownership or equity stake in Open AI. So you would think, oh, okay, so hey, any demo, you know, presumably Open AI is going to be using Windows machines, right? right? They're going to be using Microsoft products. No, they were using all Mac. Again, I think this is because we had our kind of first official-ish reports yesterday that Apple will be moving forward and announcing at some point between now and its worldwide developer conference, WWDC in June in less than a month, that they will be using open AI's GPT model in their devices, in their iOS, bringing kind of edge AI to Apple, which is a huge announcement. right? So yeah, an interesting takeaway there. But that's another huge shot at Google. But let's,

let's actually kind of look at the shots, right? Because this was, I'd say, very intentional, very intentional. And I don't think everyone picked up on it. So the thing that people did pick up on, obviously, was the timing. Right. So Google announced their I.O conference months ago, right? I think at least three or four months ago, this date has been set in stone. Open AI about four days ago said, oh, we're going to have an announcement. And it's going to be Monday right before, literally 24 hours before on the dot to Google's I.O. announcement. So presumably, right, and we talk about this here on the show all the time. These companies are poaching each other's top talent. So each company knows what the other is doing.

They know what they're working on. Presumably they know what they're going to announce. So I think a huge power move here by Open AI, essentially at the last minute coming in and saying, like, I was thinking this, I don't know, does anyone in here, maybe I'm a dork, you know, I like 90s rap. I like early 2000s rap. But does anyone remember 8 Mile, the movie, you know, with Eminem? I kind of thought Open AI did like a reverse inverse of the, you know, the Eminem, you know, 8 Mile, where essentially, you know, he's in this rap. And at the very, you know, at the very end, he's, you know, battling someone. And he, you know, he essentially says, hey, here's, here's all the, all the facts. Here's everything that you're going to say about me. So good luck coming back at that, right.

I think actually, Open AI did the inverse reverse of that. And they said, we already know your quote unquote showstopper that you're going to announce. So we're going to announce it first. And then we're going to take some very subtle, not so subtle shots at you because you guys really screwed up with your Gemini model in your original Gemini announcement. So that's the way that I took it, but I want to show you a little bit what that means. So number one, it was the timing. My gosh, you couldn't. I mean, that was a direct, straight, you know, Will Smith slap in the face from Open AI straight across the face of Google. Number two, the 1X speed callout. Yeah, I'm coming in

hot today, Chrissy. I must have slept enough and had the right amount of caffeine. So let's talk about this one X callout. So on the Open AI blog post, and again, make sure to go subscribe to our newsletter at Your EverydayAI.com. We're going to be breaking this down in a lot more detail. But you got to love to see this because it said, it said, all videos on this page are at one X real time. which if you don't really know anything, or maybe if you don't follow this space closely, you're like, okay, that's good to know.

That just tells us that any video that they put out there are real videos and they're not sped up. They're not slowed down. You know, they're not edited and pieced together. Seems like a pretty common thing to say, right? Well, not when you talk about what Google did. All right. So, and more on that right now.

So let's talk about the demos chosen And even how that correlates to that, hey, all these videos are in 1X speed, right? So, and it's definitely worth checking out on OpenAIs YouTube channel. And we're going to be linking some of our favorites in today's newsletter. But one of the videos, and I didn't see anyone else talk about this or anything. So one of the things they did is a rock paper scissors demo with GPT40. Right. So presumably this is the feature that no one has access to yet or not even paid users have access to. And this is one of those features that will kind of separate the paid tier from the free tier, right?

Because everyone gets access to the GPT40 model, but not everyone is going to get access to essentially this live assistant or what I am going to call as a live agent. I really do think it's a live AI agent. But here it is. And I love how Open AI did this. They showed a video of the program running on an iPhone, right? So it's a front-facing camera that is showing two people playing, and I know I say this wrong, paper, rock scissors. I know it's rock paper, scissors. I always say paper, rock, scissors, right?

But you have this live demo of them going, you know, paper, rock scissors in real time. And then chat GPT, the voice assistant says, oh, it looks like this person. person one, right? And you might say, all right, well, what's the big deal? Well, because that is literally the exact demo that Google originally did with its Gemini initial release in the first week of December, 23. So about six months ago. So you might be like, all right, so why, why does this matter? Well, here's the thing. Google straight up lied, right? They straight up. lied, which is why I think a lot of people didn't really investigate Google's model at the level that they maybe could have.

You know, Google's Gemini is a very powerful model, but I think the distrust was at an all-time high because as an example, you know, people saw this Google marketing video, which I will say it was not a demo video, right? What we got from Open AI, presumably is live 1X, 1.1. real demo videos. What we initially got from Google Gemini in December was a marketing video, right? So they had this same sequence of one person kind of doing this paper rock scissors, right? And then the model saying, oh, it looks like you're playing a game. It looks like you're playing paper rock scissors. So here's the difference. It wasn't live, right? What Google

did is they put out this marketing video that made it seem that Google Gemini you can interact with in real time. It can see a video live. It can process that video live and respond in voice live because that is what the marketing video showed. Here's the thing. It was all a farce. It was all a farce.

It was deceptive to say the least because they actually shared in their research paper, ah, here's actually how we did it. We actually took a bunch of screenshots, very strategic screenshots. And then we did a bunch of back and forth prompting, multiple shot prompting in order to get Google Gemini to say this. And then we just kind of put it in, you know, text to speech and made it look like it did this all on its own with its own reasoning. Nope. So, you know, open AI, geez, I mean, my gosh, just straight flame thrower with this subtle or not so subtle just shots fired at Google saying, yeah, these are. One X speed.

And hey, Google, ahead of your big announcement, here's the demo you were actually trying to do. And we got it figured out before you did. Right. And we're showing this live and in real time. All right. And hey, full disclosure, full disclosure. I think a lot of people aren't paying attention or as much attention to Google because

of this. Because Google, I think, created a huge, just this huge level of mistrust and distrust. with their original Gemini rollout. You know, every single big publication essentially covered it verbatim, fell for it, got duped, and then had to run stories two to three days later. I'm saying the biggest news publications in the world that either retracted their original reporting or they had to run a story that said, oh, well, hey, Google essentially deceived us all.

And this wasn't actually real. It was actually very manufactured, right? So I think, you know, Open AI saw that opportunity to really, you know, talk about like finishing blow, you know, Mortal Kombat, you know, woozy opponent finish, you know, finish him. That was an actual wordplay there with, with Woozy Rogers, but we'll say it was. But they knocked them out.

I'd say when I saw this, I was like, geez, they just went for Google's throat there. All right. So, hey, as a reminder, if you join late, make sure comment, hot take Tuesday. We're going to be giving that away here at the end. All right. So let's talk about what this actually means, the big picture, right? Because we already talked about this partnership, right, between Microsoft and Open AI.

Microsoft reportedly has invested 10 to 13 billion for a 49% equity stake. And now you have this now reported marriage between Apple and Open AI. Again, that's not official. But it does seem like the most recent round of reporting from Bloomberg makes it seem like, yep, this is a done deal. Apple is going to be using OpenAI for their next iPhone, which is a huge deal. So you now have this, you know, if you're an MBA fan, you know, this term like a super team here. Now you have this. So technically they're not working together. And what's very interesting here is you have people that are technically enemies, now becoming frenemies, right, in Microsoft and Apple, right?

So even when we talked about the demo, Open AI was using in everything. They were using MacBooks. They were using iPhones. They were using iPads. Even though the very company that invested more than $10 billion, their main line of business is PCs. Their main line of business is Windows operating system. Right. So this just goes to show you that Open AI is in such a powerful position because

they are so far ahead of everyone, right? And we'll be sharing the benchmarks. I didn't want to make this a benchmark episode and talking about MMLU and human eval and all of these benchmarking tests of the large language models. But obviously, I believe in every single benchmark except one, this new GPT4O model is out benchmarking

every single model. You know, we'll obviously see, I'm sure Gemini, Google Gemini will release a new one today. But regardless, now you have this, almost unfair fight of these three companies that were kind of frenemies, but now they're all essentially on the same page. And here's why that matters, right? And something that people aren't, they aren't taking this into consideration. So a big part of large language models in usage statistics, right, in feedback, reinforcement learning, this partnership, if this is true,

If OpenAI is going to be used and their GPT model is going to be used with iPhone, they're going to get so much training data. And guess what? That makes the GPT4, if that is the model that will be used on, you know, devices, whether that's 3.5, whether it's the new 4.40, we're not sure. But all of that training data between those billions of devices that Apple has out in the wild that are presumably going to be using GBT technology, guess what? All that training data makes the GPT model and the future of that technology exponentially

better. And guess who benefits from that Microsoft? Yeah, yeah, the Apple OpenAI work. That partnership, Microsoft benefits because the base of their co-pilot is obviously GPT4. So as that model improves from either hundreds of millions or billions of devices and all of this usage and the reinforcement learning that will come from this, Microsoft benefits. The three of them now have this almost unfair advantage, right? Which I think now is why you have a lot of, you know, government scrutiny into these high level partnerships. Because it creates an almost unfair advantage.

Google has an uphill battle, right? I'm going to be curious to see how Alphabet, you know, the parent company of Google, how their stock reacts both today and in the coming weeks, especially after the Apple WWDC announcement, when if this does become official, right? Because if you're a smart analyst or if you're just someone that understands how technology in the world works, you're already seeing. If this is true, Google is not in good. condition. You know, we're not even talking about the future of search necessarily, but huge uphill

battle. All right. Let's talk about now how I think that this is going to change the way that we work. All right. So something that, again, not a lot of people are talking about, but I think is actually the biggest thing is you have the desktop app. All right. This is literally bringing what I would say are agent capabilities. What I would say is a hint of AI. I know people are going to disagree with me. But, I mean, as you see some demos, I think you'll start to understand what I'm saying when this like, like, yeah, this brings in a hint of AGI.

So I think the combination of the desktop app, the iPhone and the iPad app, and kind of this agent assistant is going to create, is going to completely change the future of how we work. I do think up until now, right? And again, we don't have all of these capabilities yet. They will be, Open AI said that they will be rolling out in the coming weeks. All we have access to now is the base model, but we don't have all of these kind of agent or

assistance capabilities yet. But I'm going to go ahead. Hopefully this will work for our lives, for our podcast audience as well. So I'm going to go ahead and play a video here from Open AI. So let's go ahead and take a listen. So this is a quick demo of the desktop app and someone presumably doing this live. So let's just go ahead, take a watch, take a listen. And hey, for our commenters, I never know.

Can you let me know if you hear the audio on this? All right, here we go. I think so. Okay, next let's try something coding related. So let me take out a computer for this. Okay. So I'm going to open the chat GPT desktop app like Mira was talking about before. Okay, and to give a bit of background of what's going on, so here we have a computer and on the screen we have some code and then the chat Chaptee voice app is on the right.

So chat Cheptie will be able to hear me, but it can't see anything on the screen. So I'm going to highlight the code, command see it, and then that will send it to chat ChpT and then I'm going to talk about the code to chat ChpT. Hey, chat, GPT. Hey there, how's it going? Yeah, it's going really well. I was wondering if you could help me with a coding problem today. Of course. I'd love to help you. out. What's the coding problem you're dealing with? Okay. I'm going to share with you some code. One second.

Sure thing. Take your time. Okay. So I just shared some code with you. Could you give me a really brief one-sentence description of what's going on in the code? This code fetches daily weather data for a specific location and time period, smooths the temperature data using a rolling average, annotates a significant weather event on the resulting plot, and then displays the plot with the average, minimum, and maximum temperatures over the year. Awesome. Thank you. All right. So hopefully even for our podcast audience, that made sense. But essentially what we had there was this demo where they launched the desktop

app and it has an overlay, right? So we're going to find out a little bit more exactly how this desktop app works because it does look like there's an icon that you can click that will share your current screen. But in this case, they just copied the code. They didn't have to paste it in anywhere. The desktop app knew that they copied the code and then they talked in real time to go through and change and kind of debug and decipher and tell the person what this code actually meant and they were able to get their answers in real time. Okay. So let's also now talk a little bit about this. So this is, I think has the potential to change education. So we see here and right before this demo video, we'll link to it in the newsletter as well, is you have a,

what they said, a dad and his son working on an iPad. So on the iPad, essentially on the left hand side, it's like kind of a split screen. So it looks like there's two different things running at once. So there is a essentially a math problem on the right hand side, you know, with a triangle and angles, et cetera. And then on the left hand side, so split screen on this iPad, yeah, iPad, not a surface or, you know, again, interesting. And on the left side, you have the app for chatGBT with this feature that is not yet available yet, this, you know, this AI assistant or AI agent that can see here react and talk in real time.

So this one's a little shorter. So let's go ahead and watch and listen to this one here. Let's look at the problem together. Can you first identify which sides of the triangle are the five? Opposite, adjacent, and hypotenuse relative to angle alpha? All right. So I'm pretty sure this is the angle alpha right here, right? Ficked.

Correct. Now, looking at the triangle, which side do you think is the hypotenuse? Remember the hypotenuse? I'm not totally sure. I think it might be this one, but I really am not sure. This side, AC. You're close. Actually, side AC is called the adjacent side to the angle alpha. The hypotenuse is the longest side of a right triangle and is directly opposite the right angle.

Can you find which one is the hypotenuse? Oh, okay, I see. So I think the hypotenuse is this really long side from A to B. Would that be correct? Exactly. Well done. Now that you've identified. this is wild. This is wild. So, you know, again, hopefully even our podcast audience could understand that here. But you have this student, you know, appears to be a teenager, you know, maybe, maybe in high school. But going through, you know, what could be a homework assignment on the iPad in learning, literally in real time, got something incorrect on the first take. And then you heard this assistant in a very human tone could actually see what would.

was going on on the iPad. That's wild. That is literally the equivalent of having a live human being standing over your shoulder, watching you, dedicated, right? When you talk about access, I mean, part of this just has me so excited. You know, I think people always talk about the downsides of AI and will it take our jobs. Like, yeah, it's going to take our jobs, obviously. But, you know, people don't understand, I think, or don't look at or talk enough about the overwhelmingly positive aspects of this insanely powerful technology, right?

How about access to education, right? You know, I was lucky enough. I went to a good school growing up, but there are, you know, millions or hundreds of millions of children and kids around the world that either don't have access to high quality education or they, don't have access to education at all. Right. So yes, you know, these devices are expensive. iPads aren't cheap. You know, and $20 a month for certain people is very, maybe insurmountable.

So there's challenges there. Yes. However, what this does, not just for the future of work, but the future of education, it is hard to wrap your brain around, right? And this is where I start talking about, are we witnessing kind of the first glimpses of AGI, of artificial general intelligence. And I say kind of, if I'm being honest, right? And again, there's arguments on this on both sides.

This isn't a, is this AGI or not episode? But I mean, when you have an agent that you can talk to, right, in real time, who presumably and with some training, this is the very first iteration of this, it's not live yet, but essentially, if you know how to prompt it correctly, if you know how to use it. It is smarter than the average human in just about any general task, right? Is that, is that kind of math tutor right there going to be smarter than the smartest, you know, math professor? Absolutely not. Is it instantly smarter than the average human being? Absolutely, right? I have a master's degree. I don't remember any of that. It's far smarter than me.

So when you think about the applications that this can be used in in both how we work, how we learn, how we connect with each other, it's truly mind boggling, right? As someone that covers generative AI on a daily basis, so like, yes, a lot of my time is covered, you know, or is spent trying to understand the technology. But so much of my time is also thinking, how does this impact the future of how we work? Right. And I'll tell you this. I'll tell you this as we wrap up today's show. This changes everything, right? And I'm not one of those that, that, you know, speaks in hyperbole. Yes, this is hot take Tuesday. So I come in with hot takes, but I don't say this often. But right now, I think the combination of this, you know, super trio, this super team, you know, kind of behind the scenes, unofficial pairing of OpenAI, Microsoft, and Apple, and what that partnership and the data

sharing and the improvements of the model for those three means over time. And this is just a glimpse. This is just a glimpse. This isn't even, you know, GPT5 or, you know, whatever the next model may be called. This is just a glimpse of what is possible, right? This is just the first iterations. Yes, these are demos. We don't know.

Did they have to do 50 takes? Yeah, they told us this is 1X kind of unedited. Did they have to do 50 takes to get this? I don't know. But if this technology functions like it is functioning in these live demos, the future of work is uncertain to me, right? I don't see, hey, it's hot take Tuesday as we wrap this up. Here's what I'll say.

if you aren't already using generative AI in your day to day, your company, if you haven't already implemented generative AI, I've been saying this for a long time, you are in for a tough rest of 2024. There's good and there's bad. One of the things that up until this release, there wasn't quite that line in the sand, right? Essentially, if you were using generative AI, you were just much further ahead. than your peers.

But now, when we almost have this agent workflow, we have this assistant who in real time, in real language, can see. So this is one model, right? So before technically chat GPT could do all of these things, but they had a different model, right? So behind the scenes, there is this latency. There is this delay because if you were talking to chat GPT, it had to use their whisper technology to first change your voice into text. And then it had to reason with one part of the model.

And then it had to use kind of a separate model for text to speech and speak back to you. Now it's all one model with this GPT40, which is for Omni, right, or everywhere. Now it's one model. And the latency is next to nothing, right? I think there's some tricks that they did to do that. I noticed that usually the first, you know, two words to eight syllables were all just general responses. So I think that's one of the reasons why they were able to get that latency down to like faster than a human. Anyways, this is literally, it does appear that you have an expert in whatever you need,

readily available at all times that can see, that can hear, that can reason, that can speak with almost no latency. And if you know how to use that correctly, if you know prompt engineering 101, if you know the limitations of models, if you know their capabilities, and if you use, you use this correctly, y'all, this changes everything about how we work. All right, y'all, let's go ahead. Let's wrap this up. We're going to try this. So we're going to try this for everyone that did this, did this hot take Tuesday. So let's go ahead. Let's make sure we can do this. Hopefully this doesn't blow up in my face. I've never done this before. So we're going to start collecting comments here.

All right, let's see if this work. All right. So I'm going to start doing this. Let me know if this is fun, but essentially I'm going to be giving away an hour consult to the winner. So if you are the winner, I'm going to try to reach out to you. I don't know how this works. We're going to find out live, right? That's one of the things about a live stream. Who knows what works. But, you know, this is something normally we charge a decent amount of money for.

So, you know, we're probably going to do this once a week if you all like it. So are you ready? Should we draw for this? All right. Let's go ahead and see if this works here. All right. So actually here. Before, sorry.

I know we started there. I saw a quick problem. I saw a quick problem. It looks like it wasn't bringing in all the comments. I tried to stop it. I try to stop it right before it finished. So let's do this one more time. It looked like it was only grabbing three comments.

So let's try this one more time. All right, here we go. Let's try it one more time. I think I got to it right before it finished. So all right, let's try it one more time. Here we go. Hopefully this works better. There we go.

Ah, well, it seemed like there's still only like four people. All right. So whoever gets this is still going to win. All right, Kristen, we got you. Kristen, I can't reach out to you on YouTube. Make sure you reach out to us. You can just reply. And hey, I don't know if that got everyone.

So we'll have to do this again next week and I'll have to test it out. It looks like it only pulled in a handful of them. So sorry, y'all. All right. Yeah, I think it did just grab YouTube, Kevin. You're right. All right, I'll have to make sure next time that we get all of our LinkedIn friends as well. Yeah, yeah, total hot take Tuesday bias.

I agree. All right. So I hope this was helpful, y'all. Make sure to go to your everyday AI.com. If this was helpful, share about it. You know, leave us a review. Also, if you're listening on Spotify or Apple, appreciate y'all. Make sure to join us 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 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.adop.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 EverydayAI.com 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.