Good Morning, AI - ChatGPT 4o, HeyGen Updates and HoverAir UNBOXING

Hey everybody, it's Noelle Russell Russell. Good to see you. It is, I think it's Tuesday. I'm so excited for today. There are a lot of things I want to talk to you about today. So I decided to just hop on unannounced, no preamble, and just give it to you straight. And I got something super fun in the mail. So I wanted to share with you. And I can't wait until Friday, which is normally when I do these. So I'm going to do it right now. But first and foremost, for those of you who don't know, the spring updates have come out from OpenAI. The video is out. It's also available in our community, the iHeartAI community. You can see the QR code right over here if you want to join us. If you're not already a member of our community, come on, join us. It's fun. I share with you things I learn the second I learn it, and then Of course, we have tons of other enthusiasts around who give us lots of information from all over the world. So would love to have you. For those of you who don't know, I'm Noelle Russell. I am the host of Good Morning AI, a podcast where I've traveled the world talking on stages all about how to build responsible AI at scale. It's extremely exciting time because I've been talking about this for about 10 years, but now people are actually listening. So it's a really exciting time to be around the technology that I've had a chance to be around. So let's get started first with the spring updates. What? So I'm super excited to announce that part of OpenAI, they basically announced a new model. They call it GPT-4o, and not to be confused like GPT-4.0, but 4o. And this O is an Omni model, and the O stands for the ability for it to do some of the things that we, those of us who have been using maybe features that have not been released to the public yet, we've been working with for a while, but the ability for it to very efficiently combine text and vision and speech all together. It's quite a beautiful experience. I've been using the voice model consistently for the last few months., and I have loved it. I just open it up and just start talking to it like a friend, uh, and it's really great. It's helped me build a much better symbiotic relationship with the, the tool, right? So that I can anytime I want just open it up and start having a conversation. It also allows me to kind of play through what I'm asking for, so I don't have to know exactly the perfect prompt to type in to get the answer. I kind of can dialogue with it and get to the answers that I need, and it's been really incredible. So the exciting part about the announcement from OpenAI recently is that it now is available to you for free as part of the baseline. I think OpenAI is going back to its original roots, like how do we make this technology accessible to everyone? So I'm super excited to see that, like, right? Cannot be more happy to see how many more of these features they're releasing to the public so you don't have to be a premier user. You don't have to buy a $20 a month subscription. However, I couldn't buy it fast enough. I'd buy more. I have a team that uses it. I'm a big fan. I do like having 5 times more capacity than those who don't choose to pay for it. But that's kind of the only thing that's different now, where there was quite a few features that were different before. Some of the features I wanted to bring up I already mentioned the omni-model with voice and text and vision. It also includes video. There are some really cool live demos. Not all the features have released to everyone yet. Even though you might run to your web app or your mobile app to launch ChatGPT, you may not see all the things that I'm talking about yet, but it will be releasing over time. They typically stagger releases, and so you'll see it soon. I just always just keep checking the app, but you will already see the model available. So when you go into either your web or mobile app, you can drop down the models that are available. You'll still see good old-fashioned 3.5, but you'll now see 4 and 4.0 as options. Another new feature that you'll see is the ability to do— I can't remember what they're called— private chats, transparent chats. I have to go look now, um, but it there's a really cool reason why I like them. It's called Temporary Chat. Of course, Temporary Chat. And in Temporary Chat, it allows you to just have— and I do this quite a bit— I want to have a conversation, and in the past I'd have to go have that conversation and then delete it so it wouldn't kind of clog up my memory. It wouldn't clog up the list of things that I actually want to keep track of. So Temporary Chat's kind of nice when you're, again, just trying to iterate through some prompts. Without worrying too much about it. So temporary chat is cool. It doesn't actually create new memories. But that being said, I have to lead us into some of the new features of the GPT-4. Oh, I am a huge fan already, but I've been using some of these features for a while. So one of them, of course, is vision. But vision, not just computer vision with images, but also with video. Being able to actually talk to a robotic process while you're recording something. Like, I have cameras all over my office, lots and lots of cool stuff, uh, that I'm recording. Being able to now say, hey, based on what you're seeing, guide me through what you think I should do. I mean, imagine a world where we can ask it to help us configure. I mean, I have a— I have what are called the OBS, uh, for those of you who are familiar, OBS bots, right? This the tail. I have TailAir, I think it's called. I have 3 of them. Configuring them and synchronizing them is a bit of a beast. So I would love to have a device, a ChatGPT, just be able to be like, oh yeah, yeah, you have to do this and this and this, and just tell me what, what to do with what it sees. That would be awesome. So you'll see that in my future. Maybe I will end up using YouTube to inform me of it in the short term, but I definitely think there's such a huge opportunity for us to integrate voice with video display and the extreme latency improvements that we're seeing in this model too. So another thing I mentioned earlier about memories. So it now, it actually, I think at the last version of GPT-4 Turbo, I actually started to see a lot of improvement on cross-conversational memory. However, now I can see those memories preserved and documented in the app. So I can see what memories it's capturing and using to inform. So sometimes I'll have a conversation and go down a rabbit trail, or rabbit hole, I guess. Anyway, go down a rabbit hole and I'll be like, huh, I don't actually want that memory to now taint the rest of my conversation. So having control over what memories stay and which ones go, that's pretty awesome. So that's a big part of it. Another one is advanced data analytics. And the reason again, that these are important is for those of you, some of you are probably already GPT Plus users, in which case you might have been using this already. However, if you are not and you've just played around with ChatGPT, you now have the opportunity to use what we used to pay for for free. And I'm a big believer in that. I don't mind being early and paying for the early access, but I am absolutely a huge fan of giving this available to anyone who uses it. There are nuances to leveraging ChatGPT in the wild, meaning as a tool that you use every day. But my goodness, there's so much more reason to go and actually start spending time building a relationship and memories with this device, with this technology, so that you have the ability to, to get better answers. Some of the really cool things— oh yes, so I'm a builder, of course. I build a lot of applications. I leverage the ChatGPT APIs I actually don't do it through OpenAI. I leverage the ones that come available through Azure. And the reason I do that, and some of you may also use GPT APIs, but use them through a cloud platform. And that's just because the apps that I'm building are sitting inside Azure, inside AWS. They're already sitting in a secured, you know, well-architected cloud environment. This service is so much safer when I run it inside that secured environment. What I always tell people is like Microsoft tends to be about n-1. As soon as a model releases, it's a little bit of a lag before it becomes available on Azure OpenAI. However, super fun, I saw a post from Eric Boyd. I don't know if it was yesterday or early this morning, but I saw this post and it said, We've created an early access playground. So now, oh, I wonder if I could share my screen. I, I don't know if I will. I'm gonna try to share my screen at the end so that I can give you all the meat up front and those who wanna stay and watch me play around with my computer and try and get sharing to work, you guys can watch some of the demos. But inside the Azure AI Studio, you now can play around with GPT-4.0, which is awesome, but it's inside your Azure tenant. And I like to be inside my Azure tenant. I find comfort there. It helps me sleep better at night, as it should. So rather than using, you know, a public free version of ChatGPT, I can run it inside my cloud tenant and it's much more secure. So that being said, that's available. So awesome, go do that. Hi everybody. Oh my gosh, I can actually see some of the chats. Oh yes, what's up my friends? If you, uh, Alexandra, what's up? Hi, uh, it's so great to see you. If you are here, don't forget to say hello. I can see your chat messages. I am live streaming on like 8 platforms or something, so I might not see every platform at the same time, but I can see some of you. So great to see you, thanks for joining me. Um, okay, and if you would like, now's a great time to share this post out with your friends and let them know, hey, Noelle's live. I like just surprising you all with a little live action. So GPT-4.0, amazing stuff. Some of the other advancements, I mentioned memory. Another capability is the browse capability. I've been talking about generative AI for, I don't know, almost 18 months. And one of the biggest complaints initially was, hey, you know, I'm not really sure how recent the data is that it's trained on. I don't know how current the information is. So now having the ability to actually browse datasets, browse locations and sources, get sourced information— that has, uh, immensely improved. But they've been improving that again for a while. We're just now seeing that show up in the free application. Um, if you're paid, you get 5 times the capacity, so that's awesome. But a lot of the features we thought were special to us are now available to everyone. I think that's pretty great. Um, the API— so as I was mentioning about the API, Some of the things I wrote down that I like that I didn't really notice, but I get that they've done. One, it's 2 times faster, it's 50% cheaper. I was running on some credits that I was sponsored, so I didn't really feel that pain as much, but being 50% cheaper is much better because especially when you add voice, you're likely going to use more tokens. Little like tiny little thing to be aware of is that as you start talking, you'll end up saying more words than if you were a little bit more intentional in the prompts that you build, which means you'll be paying more for it. However, because of that, um, there is, uh, you know, the efficiency is much higher. And that being said, all of those numbers, the 2 times faster, 50% cheaper, that's of GPT-4 Turbo. And last year when GPT-4 Turbo came out, I moved all of my generative AI projects to GPT-4 Turbo in order to get those efficiencies. So now to have out of the box a GPT for O that I can use and I automatically get those benefits and it's cheaper, even more cheaper. Awesome. I'll take it. Some of the things that I always think about, like, okay, so we're now opening it up to more people. We are now adding the idea of real-time video. We're adding the idea of real-time speech. So what about responsible use? So some very interesting developments. We're going to see more of this in the coming weeks. The team at OpenAI talked about it briefly in the updates, but there are very specific mitigation strategies that have been implemented. And just know the OpenAI team and Microsoft, they've been working directly with some of the biggest generative users in some of the most highly restrictive environments. So I recently had an opportunity. I was over at Scaled AI or Scaled Agile. I don't know if you know the Scaled Agile Team. They're some of my favorite people. If you are an agilist, you know, I heart you. I think you're amazing. But people, usually these are the people that are the glue that holds the engineering work together. At least that's what I have found. I covet these people on my teams. And I had an opportunity to go to the Scaled Agile Government Day last week, and I had a chance to interview Adam Cox, who is the director of Science and Technology for the Department of Homeland Security. Now, earlier this year, I was over at DHS— I was with DHS's Undersecretary talking at CES. So I have a heart, of course, for the work that DHS does. They tend to— well, they work in so many areas. There are so many subdivisions inside that agency. But one that's very close to me is this drug problem we have, certainly in the US, around fentanyl and tracking fentanyl and how much better they've been able to do their job because of this technology. And also child trafficking and human trafficking. They tackle these problems and the improvement in the number of cases they've been able to solve because of this work inspires me every single day. And so one of the things that I know is that they are constantly looking at those types of scenarios, right? They're also thinking about who are those individuals, the humans that have to look at those images to do the identification and how do we create content filters that allow us to identify that something harmful has happened, but not actually force a human to watch every single harmful thing. And I think that that's— I mean, this is knowledge work, right? People who are in the field of knowledge work, like, it's especially as the world becomes, you know, in certain areas is very, very dark. Being the person who's doing that content moderation is really tough. Yes, not yesterday, a couple days ago, I was in Washington, DC, actually the same— the day after the Scaled Agile event. I went to the National Association of Broadcasters to be on a, a panel and we were talking about kind of misinformation, malinformation, and very similarly, right? Like we're worried about how people will use this technology in a negative way, but some of the safeguards and mitigations that have been put in place, I'm gonna talk about them at the end. Microsoft has also productized those and made them features you can add to any Azure OpenAI service implementation and it's, Pretty incredible stuff. I actually think people don't know that it's there. So I wanna show it to you so that you know it's there and then you can start telling people that it's also there. All right, so that is the first piece. Huge opportunity. So excited about the advancements. Go and play around with it today. You will, I think it defaults to GPT-4, so you'll have to go up and choose the model from the dropdown to go and use GPT-4.0. But really exciting stuff. I'm very happy. I've been using voice for a long time with GPT-4, but I've been a bit of an anomaly. No one else was using it. I think this will help us kick it into gear a little bit on that side. So, so that's exciting. The second thing I wanted to talk to you about, I hope I made notes to myself to remind myself what I was going to say. But the second thing, I might skip to the— oh no, oh my gosh, you guys are gonna love this. So I was— I don't know what I was doing, but I got a note from the team at HeyGen. For those of you who haven't seen me do a keynote, I have done a keynote. I do keynotes, and at the beginning of that keynote, I have a agen deepfake that I built like at the end of last year. So it's kind of old now, but it's still pretty good. And people look at it and they're like, what am I watching here? But that deepfake I built with agen, I built it, it was an instant avatar. So if you don't know what that means, that's one of those that you just spend 2 minutes in front of the camera and it processes and uses it. It's not fancy. It's not like a super expensive, you know, feature. It was free actually when I did it. So I love it. I love the idea of it. People are always entertained when they see it. They can't tell that it's not me. So it gives the right effect. Well, last week, or maybe earlier this week, I got an email or a note from the team and they said that they had just launched a new feature and this feature blew my mind and I've used it a ton. So one of the things that I've used is I'm actually going to try and share my screen. There we go. What? Oh my gosh, look at Noelle. So I— and hey, Jen. Now what you're seeing right now is actually one of my new avatars. This is not an instant avatar, but what they call a studio avatar, I think. I can't remember all the product names that these guys, these people use, but, but you can kind of see it. I'll play it. I don't think you'll be able to hear the audio. But you can kind of see, like, it's in 4K. Like, it's pretty crazy. Hello, I'm just gonna let you see a little bit about me. I promise you'll be able to tell the difference between me, like, real, just my vision, but then redefined by the future of artificial intelligence. This is the story of a Russian lioness. Anyway, this is not what I wanted to tell you, but I did fine-tune this model. I also fine-tuned the voice model, my, the sound of my voice, which is pretty fascinating. But what am I really excited about? They have this new cool feature called Instant Highlights. I don't know about you, for you content creators out there, you have definitely done this. I've spent way too much of my time going through existing content. So what I'll do is I will, let me, I will, I'll actually show you. So this here, this video that you see here, this is a YouTube video. It was actually down here at the bottom. It was— that's Eric Boyd, CVP of Azure AI. He introduced me to do a demo on computer vision. I think that was the only thing I demoed. But anyway, on stage at Build. Fun fact, I'll be at Build next week. Woo! But anyway, he introduced me to go on stage to do this demo and it's like 5 or 6 minutes. So what I did is I went out to YouTube and it did not ask— I don't need to own the content. I could take any content I want. So I go out to YouTube, I grab my video because I want my stuff, and then I inject it. So you can see here, drop your URL, auto-detect the highlights. Uh, this is the language that you can choose, so it'll even translate the captions in all of these languages. And then I'll go in and it created for me a couple of these. This one's 23 seconds. How do you tell the difference between Chip and Dale? I don't know if you can hear. Right? The chipmunks. Look at those captions! Anyone know? They created— man, it's snapshotting. You can't tell how long this is. 35 seconds? Something? I decided to build— Oh my gosh, it went through my whole video. Now this was a 5-minute video, so it created 1, 2, 3, 4, 5, 6 different videos, different lengths, TikTok style, Instagram Reel style. YouTube Short style. But all of my content, I mean, some of these are so— I'm just like, you can create Q&A. Oh my gosh. In minutes for departments that need it the most. As you can see here, right? I have content that I have built. And this is what I tell people who are marketers, like, as you start building this, you know, if you're a marketer or if you are a professional and you're building your own brand, just start talking about the stuff that you do. Like what I'm doing right now. Like build a little, you know, do a live stream. Talk about the things you care about. And now guess what I'm going to do to this very video? I'm gonna go, I'm gonna take the YouTube stream, I'm gonna put it into Hey Jen Instant Highlights, and poof, it's gonna give me 20 or more, uh, videos. I'm gonna go into a longer piece of content. This is me at the Women in Analytics conference that I spoke at quite a bit ago. Um, some of these you'll see because the content is longer, and when I do a keynote, my bites are longer because I'm not doing, you know, I wasn't super short, 5 minutes on the stage. This, I had 45 minutes on the page. So it grabbed, and again, I gave no edits, no guidance as to what to grab. Um, but you can see it takes— now granted, the video I gave it was portrait— no, was landscape style. So you can see what it did here. It also, um, grabs content. So if it is this way and I happen to be all the way on the right or left side of the frame, it might grab me and I'm not even in the picture. But I'm I'm okay with that because this is just little bits of content that I can use to drive people, um, to the bigger piece of content. So, so amazing. I've already provided feedback to the team about what I want to see more of, uh, but wow, really cool stuff. I don't know, I hope you like it. Yes, amazing. Priscilla, awesome. Good to see you here. Good to see you. I'm gonna be in your city like this weekend, so I'll text you and let you know. Um, I'm headed to Orlando. Uh, so anyway, hey Jen, instant highlights, like bam, awesome. I also have an AI voice. This is, uh, my own AI voice. There's also a bunch of other ones. And then my video avatar. And like I said, this is the one that I built for free, uh, which was my instant avatar. And this is my fine-tuned Noelle model. And I'm just super excited about it. How will I use it? I'm still figuring that part out. Love to hear any ideas that you all have. But, but yes, I'm thinking about maybe just like I do with my keynotes, have it be the beginning or the end, and I can kind of either catch the content from my AI avatar or pitch it to my AI avatar to close things out. We'll see what happens. Really exciting times. Okay. Finally, last thing I want to share with you today. We've been talking here for 23 minutes already, so I don't want to keep you too long. But I do have an unboxing to do. I don't know if you've heard about this. I hope I don't have to do too much. I am a big fan of a concept known as frustration-free packaging. I don't know if you know what that is, but frustration-free packaging was a project originated in the mind of someone who worked at Amazon, and they were like, for the love of heaven, can we please just make packaging easier to get into? I'm a huge fan of that project. I support it wholeheartedly. This looks like an example of frustration-free packaging. That was pretty easy. This is the HoverAir. What? Oh, I should not share my screen anymore. I'm gonna stop doing that so you can see what I'm doing. So this is the HoverAir. Have you heard of it? Do you know? So I am, um, last week I kicked off a new project, a new generative model that I'm building for a company. I kicked it off with a couple of my friends, amazing at Unstuck Labs, phenomenal team, by the way. They're in DC. They help startup founders do their magic and get into the world. Absolutely love them. And so we're kicking off a new project. You'll hear more about it in hopefully a couple of weeks. But he had this device, this, you know, drone. However, what he did was he like had it in his hand and it just floated up into the air, followed him around, took audio and video. And I was like, okay, I'm, I'm gonna have to get me some of that. So I ordered it. Now I will say Wyle is his name. He has— every time I see him, he's got a new tool, a new AI thing, and I'm like, ooh, I want one. So I buy it. And so hopefully I can be that person for you. However, I'm in a situation where, like, it's not terrible that I, you know, get an AI gadget every so often. If you are already on that list with your spouse or partner that you're not allowed to buy more gadgets, do not buy more gadgets. I am not here to sponsor that. That behavior. But let's go. So Hover Air, cool. Um, this one, I can't remember what color I ended up buying. There are two colors available. Um, this one looks like it is— oh gosh, oh my gosh, here we go. Oh, I bought the black one. I almost got the white one, but this is it. It looks like this. It's pretty small, sits in the palm of your hand. Follows you around. You will see. So I'm headed to BUILD. I will be flying this puppy everywhere I go. It's going to record me asking questions of my friends and people in the— I'm going to be running the Expert Booth. So if you're at BUILD, come on over. I'll be at the Expert Booth. And I don't think it has— oh, maybe I got an extra battery. I don't think there's a battery in it. Let's see. No, I don't think so. I think I have to put the— oh, there's a light. Oh my gosh, I don't know what to do. Oh gosh, it's lighting up. So what are the dangers? Current mode hover. What? Please download Hover X1 app. Can you guys hear that? Connect and activate the device on the app. What app? So it said to connect and activate the device on the app, but it's going. And so you can see here there's a little camera, right? It's hard to like orient, but that's the little camera. It's got a little light. The audio is not terrible, but actually I think picks up on your phone. Um, so you can count on me. I did also get the, um, the AI pin. I did not get the rabbit. Any of you have the rabbit? Um, so my friend Wile has the rabbit. I decided not to buy it yet, but, um, but I got this one instead, and I'm super excited because this is gonna help me build more content. And I could use Hey Jen to create instant highlights of it all. It's gonna be amazing. So many of you usually ask me like, how do you do all the things? How do you, you know, like end up at all these events? How do you record all this content, publish this content? And I will tell you, it's because I get ideas like this from friends that I connect with in real life. And I use, I am constantly on the lookout for things that'll optimize my existence. The new ChatGPT 4.0 is going to be amazing, huge optimizer. Certainly being able to, oh, I'm gonna share with you now the, I'm gonna go ahead now that I know that it works, I'm gonna go ahead and share with you the OpenAI Studio. So in OpenAI Studio, when you go in there, this is it right here. You'll see here there's now this early access playground. Let me see if I can make it a little bigger. That's a little bit bigger, that's better. So there's a bunch of stuff in here. Assistants, that's the world of agents. It is the future. Multitask orchestration is where it's at. But if you look here, Early Access Playground will actually show you if you scroll, if I scroll down, oh boy, maybe I, oh dear, let's see. I have too much stuff on my desk. I can't see. But you can see down here the model that it's using is GPT-4o. So you can play with it right now, right now. So if you need the link, I re— I reposted Eric Boyd's link, but you could just go right into the Azure AI Studio, click on it. And again, the benefit of this is that, A, you have the ability to use it inside of Azure. Your Azure tenant is well architected, I hope. And you know, you've got the security configured accordingly for your tenant. So when you use this, you can use it safely. You can even add data sources. You can provide a little— you can have a little bit more flexibility in the type of use cases that you test out. But super exciting stuff. I'm going to go ahead and minimize that a little bit so you can see a little more of the environment. But over here you can see the setup. This leads me to the last thing I'm going to talk to you about, which is— oh, good, and we'll end right at 30 minutes, should be great. The last thing I'm going to show you over here is the new Azure Content Safety tools inside of— oh, that was the dinging— inside of the content, or sorry, Azure Content Safety. These things are called content filters. I'm going to create a new one just so you can see what it looks like. But these content filters— let me see if I can make them a little bigger for you. It's about as big as it gets. Hopefully you can see that okay. But inside these content filters, look at the things that you can detect on your prompts and your responses before customers see them. If you don't know that this exists, I mean, these are things that I've been talking about manually having to red team. You now get the benefit of all of the work that Microsoft has done with organizations like DHS and some of these high, high-profile, really risk-averse organizations. These are the types of things that have helped inform the development of these services. And you can follow Sarah Bird. She runs the Responsible AI team. She's out in New York City. You can find her and follow her on LinkedIn. But look, hate, sexual, self-harm, violence, and the threshold by default, it's set to medium. But if you are in, let's say, an educator building an application for a school system, like, whoosh, maybe you increase that threshold. Or maybe you're building something in a game and that game is like Grand Theft Auto. Not that I'd ever play that, but if maybe you're playing a game that like has blood and guts in it, maybe you have to think differently about some of these thresholds. So it allows you as the developer, the practitioner, the business leader to go in and, and know that these settings are available. Hey, Stephen, what's up? Stephen from way over yonder. Are you in Thailand? I think it was, um, really far away. Welcome to the party. Thanks everybody. Uh, this is awesome. So yeah, prompts versus completions. You have two different content filters for each. I'm just gonna flip through these settings real quick so you can see what else there is, because it's not just about content filtering, meaning What are people saying to your bot and how is your bot responding? Super important. But that's not it. We also have the ability to add what are called prompt shields. I did a whole article on this in the AI Leadership Journal, which is the newsletter I have on LinkedIn. If you're not subscribed, definitely subscribe. But if in that I talked about these prompt shields and I said, you know, all of us are superheroes and not all of us need capes, but we definitely need shields. And these prompt shields will save you so much time. They again have looked at some of the most common patterns of bad behavior around jailbreaking and cross-prompt injection and prompt injection and created the opportunity to create a shield which will basically alert you when these things happen. So amazing. Why wouldn't we do that? Why wouldn't we use AI to tell us when bad things are happening so that we as the human can make better choices? We should. But you know what? We're like, you now are part of like the 1% of the 1% who even knows these features exist. So share it with your friends. Tell everyone. You can also add a block list. So as people behave badly, you can block them. So that's kind of nice. And then of course, there's the opportunity to look at streaming mode. So one of the cool things I've built in HeyGen is actually a streaming Noelle, nearly Noelle avatar. And in the streaming nearly Noelle avatar, I'm streaming against GPT. So I would really like to make sure that those filters and protections are in place in streaming mode. It's not for every use case, but there are certainly some, especially like conversational chatbots or avatars that are always on and streaming, uh, question and answer back and forth between a client. Um, that's something that you might want to consider. Um, so anyway, that is it for me today. I'm so excited. I got to share with you all my fun new toys. I'm gonna just share with you a little bit about what I'm doing next. This coming week, actually tomorrow, I am headed to ADP. So ADP has a Pro Summit. I'm going to be their keynote on Thursday morning. So if you are going to be at ADP Pro, come on out. I'd love to see you. If you know someone who's at ADP Pro, tell them that I'll be there. I would love to meet them. Then I'm going to be on Friday going to a graduation. Yay, teenagers are graduating. And so I'll be doing that. And headed to Tampa because I owe some kids some roller coasters. So yeah, I'm hoping that I love them. I remember liking roller coasters when I was younger. I'm hoping I still love them now. If any of you are, you know, in my age group, meaning you were in this industry during Y2K, let me know if you hit a certain age and they were no longer fun. But I'm gonna do it. I'm gonna do it. Busch Gardens, like, how can you not? So ADP Pro is next week or is day after tomorrow. Next week is Microsoft Build. I will be streaming live every day. I'll talk about all the keynotes that I'm going to be going to and all that. Oh, thanks, Stephen, you deserve one. I might spend whatever it costs to get you one all the way where you are. But yeah, AI Leadership Institute in the house. Oh my gosh, that's a great announcement. But anyway, next week I'll be at Build. At the end of the week, I will be back in Orlando for the F3 Festival. Look it up. It's going to be amazing. I will be running Good Morning AI every morning covering all of the fun topics that we will be covering. And that is just like the next like 8 days. I will also be doing other things. SAP Sapphire is coming up and I will be hosting that virtually. I also have an opportunity to Um, support the, uh, Adobe launch of a couple of very cool AI features. Things that I built from scratch that I took a long time to build are now integrated in some of these products. So I cannot wait to show you. That's at the beginning of June. Um, I have a lot of events coming up, but for today, I'll just tell you if I'm coming to your city. If you're in the community, you can go to the calendar and see where I am. Um, and that way, if you're in that city, hopefully we can meet up, have coffee. I try to do coffee chats, like, you know, say, I'll be at the Starbucks for these 2 hours. And if you are around, I hope you'll join and meet up in real life because that is where the fun happens. Right? I love it. I love seeing everybody that I've only met online and finally getting a chance to meet people in person. I met someone a couple of days ago in DC and they were like 6'8" or something like huge. And of course on on video, you have no idea. So I walked in, I'm like, I did what you should never do. I looked up, I was like, oh my gosh, you're so tall. Anyway, I regretted it the second it left my mouth. But yes, oh, awesome. Yes, love it. So, oh no, oh no, Stephen. Stephen's giving me some insight into the roller coasters. But great point, actually. Stephen also mentioned there's a couple new federal regulations that have come out, if you're not aware of them. And then for real, I'm going to let you go because they're eating dinner right now. So I got to go. But NIST announced a new part of the risk framework for AI and the executive order. Kamala Harris, Vice President Harris, actually also released an amendment to that executive order all around AI governance and risk in the area of AI and innovation. Really great documents. I'd encourage you to read them. But guess what? If you don't want to read them, what can you do? That's right. You could download the PDF version and you can feed it into GPT and be like, hey, ChatGPT, tell me what I need to know about this. Give me the highlights or tell me a bedtime story in the context of this document. Oh, it's fun. Let's go have some fun. Make these tools your friends. I know we say this all the time, but You know, AI is not out here to take any jobs, replace any roles, but people who are using this technology, they will definitely be a threat if you are not one of them. So hop on board, learn, learn, just learn and be curious. Start asking and prompting these models for things you ask yourself every day. You don't have to use the answer, but just get used to asking the question. It's been my absolute pleasure to be with you today. Thank you for staying and thanks for asking your friends to join. And this will be available on replay. So if you see the replay and you want to repost it or share it, please do. You guys are awesome. You're my friends and it's been amazing. So, all right, I'm going to go. It's been fun. I'll see you next time. And yeah, I think I'll see you at the ADP Summit next. So catch my live with my new Hover Air. You'll catch my live then. Okay. Take care, everybody. Have a great day. Bye.