I ❤️ AI Q&A: How To Build DIY AI Solutions and Become an AI Architect?

Hello, good friend, good to see you. Good to see you too, thank you. Hello everyone, what, some new faces? Glad you're here, awesome, amazing, welcome in. Hello David, who's Lenovo ThinkPad Yoga?

That's me. Oh, okay.

Hi. Hello. Good to see you.

So nice to finally meet you. I heard wonderful things about you from—

I met— Yuvika and Bhawini.

Yeah, Yuvika at the Microsoft AI conference.

Amazing. I just posted about them today. Yeah, I wasn't able to be there. I think I was at a different event at the same time, but they took lots of pictures and video. It looked like a really good time and lots of people, and they did an amazing job, I heard. Yes, they did an amazing job and I am so interested in joining your team. I want to present also. I love presenting.

Oh good.

Oh, I'm so glad. I'm so glad. Well, welcome. This is definitely how to do it. Many, many opportunities for those of us who choose to take up the challenge. Yeah, yeah, that's amazing. Awesome. Alright, so as you, some of you are new, but well, people join over the next few minutes, but this is live broadcasted. So we have a couple different sessions that we do on Fridays. One is for our AI Leadership Institute membership that is done at 2 o'clock, but that is for members only, and they get like secret behind the scenes of all the things that I'm building. And then this one is just I Love AI, so it's a community call. I'm gonna talk a little bit about whatever interesting technology I am focused on. So I'll be sharing with you all that today and a chance for you to go and play on your own. Um, and, but it is live broadcasted, so I always like to tell everyone as we begin to start this process that we are live right now. I, it might even be live already. I think it is. Welcome to the I Love AI community call. Super excited to have y'all here. If you are in the Zoom, that means that you get to ask questions and you could be on camera. You don't have to be on camera. It's totally up to you. I personally like it because it makes me feel like I'm not alone in the world, but you're not required. But it does also allow you to unmute and ask questions at any time. We'll specifically be asking questions of you. So we end the call with about 10 to 15 minutes of just pure networking time, a chance to hop into a break room with 3 or 4 people that are in the community and just find out who they are. So that's kind of my favorite part. But I do use this time to update you on all things AI, AI leadership. And whatever craziness that I'm working on at the moment. So I'm gonna go ahead and jump into that. I do have some robots to welcome into the room, so let me do that. Good to see you, Kim, what's up? Awesome, and Diana's No Takers here. I can't tell if Diana's in the house yet, but let me see, let me see. Oh, Joe's here. Joe, you'll be very happy to know that your book is finally on the way. So you'll get it tomorrow, yay! And hopefully you'll enjoy every moment of that. Let's see, Tara's in the house. All right, so typically I kick out note takers, um, within the first 10 minutes if the human does not show up to actually be here, because the whole point of a note taker is to help you stay present while it takes notes, not to replace you as a human on the call. So not to punish anybody, but I do like to make sure that we're all represented. Um, but yay, there you are! Good to see you. Okay, cool. Um, all right, awesome. So welcome in. Uh, I'm gonna start with a little bit of an update. One, uh, last week— I mean, we just talked on Monday, so it's been a very short week so far— but last week I was over at NDC at the Cyber Risk, uh, International Cyber Risk Summit and got to speak, uh, to a bunch of people, but also got to demonstrate a workshop that I do from stage on deep dives. Uh, it's a deep dive into AI solutions design. Um, but here's the cool thing is that we actually did— we recorded a version of it and I just got it back out of editing. And so that's going to be available to you all. Uh, if you are a member of the I Love AI community, you'll get access to it free. Um, it also has a set of resources that come with it which will be available in the Institute. So that I'm going to try and do before the weekend, uh, because I'm excited. I hope you like it. I don't know, I'll just— yeah, I'm, I'm, I'm gonna let you see it. It's— I I think 7 or 10 sessions that are like 4 or 5 minutes apiece, um, but it was a whiteboard session. So I always tell people— I don't know if I have one here— but I always tell people like, record everything, uh, everything you do, record it. Like, I will record things on my phone. Oh, I don't know if you all remember, but I had a Pixel broken on my phone last week, but it's fixed now. Um, but I'll just turn on voice memos on my phone anytime I'm about to take a stage or do a podcast because today there's the ability for you to repurpose that all over the place. So I'll share with you some of those tools that I use, but, but I always encourage you to do that. So I did that. I put up a GoPro in the back of the room of this whiteboard session that I did, and so we are repurposing that. So when you get it, I hope you— hopefully you will take a look at it, and if you enjoy it, please comment in the, in the course comments. Let me know what you think. But it's just, you know, I'm just trying to like help you all be in the rooms that I'm in and get to learn from the things that I'm doing. Um, great, cool, good to see you. Thank you for joining us. Hello everyone, what's up? Derek and Joe, thanks for joining the video. What's up? Um, I keep getting, uh, robots asking me for permission, so I am giving permission to the robots. I think we're all accounted for. All right, awesome. So first thing, uh, that conference was great, but it actually led to a couple really interesting conversations. Um, the AI leadership call we just finished, I focused our attention on autonomous vehicles for a couple reasons. One, there was a recent acquisition by Amazon of a robotaxi service that's owned by a woman of color who I'm super excited about. So I would encourage you, I think the, the company's called Zoox, if I'm pronouncing it correctly, Z-O-O-X. So go take a look at it, celebrate some success. Obviously, I like its origins and the team that created it, but it's also actually just cool tech. I think it's interesting to see these robotaxi companies get started, get acquired. Amazon acquired them. And, you know, we've got the whole Waymo situation happening in 3 cities right now. It's still pretty new, but it, it's an interesting problem to solve. So I, of course, am a solution architect at heart. When I see a problem, I try to solve it specifically with artificial intelligence. But before that, with cloud. In data, and before that with enterprise Java. So I've always been kind of that person who's like, if I see a problem, whatever tech I know at the time, I try to use it to solve that. And hopefully, I think I share that with many of you that are here. So this autonomous thing, this autonomous vehicle thing, I think is an interesting problem. So what I'm going to share with you today is a project that I shared with our leadership organization, but I wanted to give you a chance to actually build this on your own, DIY. So that's what this call typically is, is a lot of things that I will resources and tools that I'll give you, and that's kind of do-it-yourself. You can go and do it and follow it. I'll give you some guidance in the community. Um, inside, uh, our school community, there's an opportunity for you to like take classes and step through the process, but it's mostly do-it-yourself. It's kind of like, okay, Noelle said a cool thing, I'm gonna go research it and do it on my own. But hopefully over the, over the weeks that you spend with us, you will find that there are some things that I will encourage you to do yourself that you maybe would have never thought to do on your own.. But even more importantly, if you've already done the things that I'm encouraging other people to do, you get to step into that mentorship role. Like, this is not just the Noelle Russell Russell Show. This is kind of like, oh my gosh, I did that and it totally failed, or it totally worked. And it becomes so much more fun, uh, at towards the end of this call when we're all kind of sharing our ideas together. Um, so we have a couple announcements in the AI, uh, in the I Love AI community. One is that we are doing a live learning lab. On Tuesday. Live Learning Labs are the fingers-on-keyboard activities. You all will get to see the replay because you're in I Love AI, but if you want to be there in person, let me know. But it is around Gamma, asset development in Gamma. So what I, I do, and so I'm going to show you some of the resources that I've built in Gamma. If you took the challenge, the Rise Up AI Challenge, the assessment scorecard was in Gamma. For example, I'm gonna try and find my chat. Oh, there you are. Hello everyone. I don't know, Sherilyn might beat me to it. Let's see. Oh no. Good, good, good. Oh, awesome. Sherilyn's on it with the robo taxi links. Fantastic. So I'm gonna share this link with you first. Oopsie. My keyboard is Bluetooth and takes a second to connect. Okay. So this one is the link that I gave. It's inside our community. If you go into the Rise Up AI Challenge content, It is the scorecard that I created. So the first question from the community was like, hey, teach me how to build a scorecard. And I was like, sure, I'll do that. So I want to show you what that looks like. I think I called it gamma.upskilling. I can't talk and type at the same time, so forgive me for a second. And I think I called it scorecard, but the link is in the community. There we go. I'm gonna click on it and see if it works. So I use this to build a scorecard. I iterated with it, right? Vibe Coding at its best, where I'm like, down, I wanna be able to download a sample report. I wanna be able to let people take the assessment. This was an illustration in the Rise Up community of how to Vibe Code your way to an asset you can use to value your business. Ideally, you would actually build this assessment to dump that content into a Google Form and use that to inform conversations you have with clients if you're a consultant. I didn't do all that, but what— that's what we're going to talk about in the live learning lab that we do on Tuesday. It's in the middle of the day. So if you were able to see that, that it— that was built in Gamma. It's hosted on Gamma. And then I also had a request for our catalog. So for those of you not familiar, AI Leadership Institute, which celebrated 10 years this week, I'm super excited. We started with 3 classes at the very beginning. I have this mantra. We go from the boardroom to the whiteboard to the keyboard. For 7 years, we taught a boardroom education class, a senior technical leadership workshop, and an AI for everyone employee class so that everyone in the company got a level of AI literacy. Only recently have we been asked to push the envelope, right? People are like, I mean, before they're even ready, they're like, okay, what is GenAI and what is, um, agentic and what is MCP? What's A2A? I'm like, you don't need to know all that, but I, I solve, you know, I help them help themselves. So I, uh, started creating lots and lots of courses as a result of that. Um, and so I'm going to share with you this. Thank you, right? It's exciting. I can't believe it's been 10 years already, which means I've been out of Alexa. So my journey, many of you know, uh, into AI started with me saying yes to joining the Alexa team, and that means that was like 12 years ago, which is crazy, crazy that 12 years already. Um, all right, so this next link is— I have to focus because it's long— aileadershipinstitute.com/catalog. Let's hope I spelled that right. All right, there you go. So this is another Gamma, uh, resource that I built. So this one I'm going to walk through, but I think, you know, like, time to market is everything now. And one of my philosophies some of you may be familiar with, is to build— anyone can build a viable product today. If you have an idea, anyone can build for viability. What we will do as responsible, but more importantly, AI solution-driven leaders is we should build for remarkableness. How do we make it cool? The remarkable part of this is that I get to use this as a story to a team to say it took me 45 minutes to build this catalog. It becomes part of the story, being a doer, I'm a doer, not just a talker, not using a PowerPoint presentation when I can use AI to generate something. Because as a solution designer, I wanna show them that my words become my solutions, and that's the world we're living in right now. So, so it's kinda cool. Even if you don't really want to use Gamma, you don't have to use Gamma. You could use Beautiful AI. But leveraging AI to be the asset you hand to a client or you hand to, you know, if you're in HR, you hand to an employee. It is a demonstration of like this symbiosis, right, that's starting to happen and that it's okay to use AI to do these things. So I find it's really valuable. A lot of people, of course, end up coming back and be like, how'd you build that? How did that work? But it also, of course, provides a bunch of information to companies about what we do. So, so that's what we're gonna do on Tuesday. This coming, I guess, August 6th, so I guess 2 weeks from now, Maybe, yeah, 2 weeks from now, August 6th, I'm hosting a 30-minute, what is it called, lightning session on Maven on how to become an AI Solutions Architect. It's 100% for free, but it is a lead magnet, if you will, or an intro session to a class we just released on Maven, which is how to become a certified AI Solutions Architect. It is a 5-day crash course Maven course. We did it in 5 weeks. I'm a believer that if you are dedicated, you can knock it out, all the work you can do every day. I guess I am ADD, and things that take 5 weeks, honestly, I'm like, I lose interest by the time I get to the end. So I was like, what will this look like? So it's a bit of an experiment. We're gonna do everything we did in 5 weeks, but instead of giving you a week to do it, we're going to give you an hour in class time to do the, the required, uh, work. And it's not super— some of it's gonna be easy for some of you who are prone to being able to have conversations with clients. Most of us, we struggle with translating engineering work into business value, and that's what you're gonna get for the most part as an AI Solutions Architect. That's your job is to take any number of AI components, capabilities, technologies, and turn it into business value for a company. So that's what we'll do. We'll take you through the different stages. I have a book, called the 21 Laws of AI Solutions Architecture. They're literally how I have generated, at Microsoft, I influenced over a couple billion, I think it was $1.8 billion in AI sales in 2018 and 2019. This was before AI was cool. And so there were patterns in the work that I did that got customers to say yes, even if they didn't fully understand what I was talking about. So that's what I bottled up. Yeah, yeah, well, I would say, Tara, to that question around PowerPoint, I would say that they— many HR people and certainly hiring managers have a very bad habit of copying and pasting old job descriptions into new job descriptions, which becomes even more evident as we move forward. But I would also say that I'm a bit of an anomaly in— and I actually like it. I hope everyone keeps using PowerPoint because when I come in, and don't use PowerPoint, people are like, oh, what is this? You're going to whiteboard? What is this? You're going to use an AI-generated visual? Like, it's nice. If everybody was doing that, it wouldn't have a cool— it wouldn't be as remarkable as it is right now. So I don't mind it. But yeah, I would agree that I would be shocked. But a lot of people, I think, are just copying old stuff, old bullet points into new job descriptions. All right, cool. So lots of links in here. Don't forget, to also— I don't know if I see any in here, but this is a great opportunity to get to know each other. We had a call earlier this week, but typically, you know, we have lots of people join these calls, different people all the time. We have entrepreneurs, business owners. This is a chance to get to know each other. If nothing else, at least connect with each other. I will tell you from experience, you never know what the next 6 months will do for someone's career or business, and you never know who they're going to be or who they're going to know. Uh, which leads me as a nice segue to the next topic. But before I do that, now's your chance to kind of throw into the chat your LinkedIn profile, if you will, if you are interested, just so that people can connect with you. Many conversations have been had offline, like where, you know, you guys connect on LinkedIn, you're like, oh my gosh, we're in the same space, or you're in healthcare, I'm in finance, I would love to learn what you're doing. And these coffee chats, who knows, turns into business partnerships, turns into affiliate relationships. You never know. So I always say, you know, more is more. So definitely do that. And as you see these links come in, what I do when I'm on these calls and people are sharing, I just click on the button and what'll happen is it'll open up a browser tab and then after the call you can just go through each one. I'd encourage you not to blindly connect to people. I don't. I try not to. It's better if you like actually kind of look at them and be like, oh, do we have anything in common? Do we know the same people? And even better if you can write them an actual personalized message. I know, scale on the unscalable, it's a thing. But I, I do think in this world that it's really important. And a lot of my best, best friends, honestly, I met on LinkedIn. So, so I think you have a chance to meet some friends here as well. And then in the chat box, when you both take it out, there's 3 little dots in there. It allows you to save the chat. So that's also if you don't have time to like go through and click all the links. You can just at the end of the call, I'll remind you again at the end, but at the end of the call you can click the three dots, save the chat. I think I'm making sure. Yeah, yeah, so you hit save chat and then that way you can get all of the links that were provided by Sherilyn and anyone else who throws links in and also everyone's LinkedIn profile. So great opportunity to grow your network and like I said, you never know who you are going to meet in here, but more importantly, you just don't know who people are going to become from here. That has been my most delightful thing is to watch watch people grow and scale successfully. You know, once I meet them, I'm like, holy mama, I'm so glad I know you now and that I knew you before you were you. So one of those people is, many of you know, Ali Miller. We were friends when she was at Amazon. And yeah, I was like, now, and now she's like got a million followers and is all over the interwebs. I'm like, good for you. I'm glad I knew you when you were little. Didn't know anyone. So awesome. Okay, cool. Great, great, great.

Great.

Um, all right, so I, that was gonna segue myself into a conversation that now I can't remember. Um, what was my other announcement? All right, so I told you about the class. Um, I told you about AI Solutions Architect. I guess that's it. All right, so there was something else, and when it hits me, as many of you know, I will just go back to it, uh, because I can't remember it right now. I should have written it down. Um, what I wanted to do today was share with you a couple different technologies that I'm playing around with that just have my attention for some reason. Um, one is autonomous vehicles, uh, because of this acquisition that just happened. That means Amazon— Amazon has a tendency of acquiring these kind of ideas. I don't know how many of you know, actually, I just got my, uh, Prime Day purchase. I, I try to only buy one thing on Prime Day— days now. They used to be just one day, now it's like 4 days. Um, but I bought the newest Blink camera But fun fact, I onboarded Blink the day they were acquired. And so there's a whole team, just to give you some, you know, inside baseball. There's a whole team that onboards these acquisitions when you come in and they all sit in a room and they have to learn how to be Amazonians and they didn't sign up for that, right? Like most of us that get hired, like we are like, I know who Amazon is and I want to be there. With acquisitions, like whether you like it or not, you're going. And oftentimes with the timeframe. However, in most cases, they're pretty excited. So when I onboarded, I did the onboarding session for Blink, they, they had just, you know, got all this equity. So they were extremely excited. They were like celebrating and having a party. And I came in to like teach them how to be Amazonian, which was my specialty, how to bleed orange like the rest of us. And so it was really cool to see You know, and now I just— so I just got a Blink camera, um, as my one purchase, and I just love it. But as a result, I know that this team is not going to go— you know, this, um, Zouk's team is going to go through the same experience where, you know, it's one thing to work in a startup, it's another thing to work for Amazon. Like, it's just a different experience. Many of us know the difference between an entrepreneurial world and, and a big corporation world, but I feel like Amazon's a little different in the sense that they really do like to preserve these companies inside of Amazon. So like Blink is still its own little team. Um, I don't know how many of you remember like the Zappos team, right? Like they're part of Amazon but not exactly. A16N, like all these acquisitions, some of them get completely immersed, some of them get to kind of stay their own. Uh, and Blink has maintained it. So I would imagine RoboTax—

medical—

one medical— good, one medical. Exactly. Yeah, good point, Eve. Yes. Um, did you— did I say yes to you? Did I reply back to you? Because if not, it's in draft. Oh, I did. Okay, because I wrote it. I'm sorry, sorry, sorry, a little side business. Uh, but yes, great, good, I'm glad. Um, but yeah, One Medical, like all of these areas, I just think it's interesting, and I, I do enjoy watching Amazon's growth trajectory, uh, from a business perspective. Amazon, probably one of the first coming. I mean, granted, I was there before Alexa was born. But they were always pushing the edge. I remember one of the things that they were doing was deploying on a single— like, if you go to amazon.com on your mobile or on your desktop, you are looking at about 200 web services that are all deploying in autonomous teams to the same website at the same time. And they're all on different— for those of you who are in engineering, like, they're all on different code pipelines. They all merge. To a branch. Like, it's, it's pretty impressive that not more than one part of that page isn't failing at one time. But I think that's the magic of knowing how to scale, and definitely, as we know, Amazon's really good at that. So this Robotaxi thing got me thinking, um, you know, I've also, uh, years ago I was in Arizona, in Scottsdale and Phoenix, when Waymo first entered the scene there. I don't know how many of you have been in one of these Waymo taxis. They're now in, I think, 3 cities, maybe 4 cities. But we're going to start to see more of them. I also think it's really interesting to think about like the autonomous capabilities in our vehicles. So Ford recently did this whole campaign and it goes to kind of the point of why I think this is an interesting time is Ford did this whole campaign around drivers who are not technical getting in the car and then the cameraman, like the camera's here. And then there's an interview and they're like, turn on autonomous driving. And they're like telling the driver to turn on the thing. And then they go turn it on and they're like— and I think the whole goal is us look at their face and not see terror or fear. And, and all— I mean, who knows how many of these they did, but the 4 that I saw, they were like, oh, oh, well, this isn't, this isn't terrible. And they were all different people, right? There was like, you know, uh, inner city person, there was a person from the clearly from like the country with the plaid and the like, you know, trucker hat. And like there were all different personas to try and demystify how, how this stuff works and that it's not scary and to go ahead and turn on these features. I think what we're seeing, which happens with most AI models at some point, is that the only way AI systems are successful at scale is when you get massive adoption, right? And we can, We can like just look to ChatGPT as a clear example, right? And I don't know how many of you know the origin, you know, the origin story of how ChatGPT came to market, but it started off as a tool for data scientists. And it's what we use actually in AI Foundry today in Microsoft AI Foundry. Like it's just a playground for me to test out things before and without like contaminating a model's training data, right? It's just a playground, a lightweight version to kind of test out certain, language constructs. And this playground that they built in OpenAI, the engineers would spend more time in the playground than actually doing, like, the job. And Sam Altman talks about this. He actually just did a really great interview in Y Combinator. If you look it up, it's only like a week and a half old, maybe 2 weeks old. But he was talking about this a little bit, but it'll lead you down a rabbit hole of other content that talks about this in depth. Uh, but it's just really cool. I like these origin stories. Jeff Bezos's origin story around like those early days. Um, but realizing that this playground, they were like, wow, people really like just talking to this LLM, like for no reason, like not as a product, not as a service. So they, that's how, uh, 3.35, 4, like that's how that product became something that everyone, you know, now there's 300 million users a week, maybe more now. That was like 2 weeks ago data. But that's a lot of people, um, a week. Like, that's, it's a, it's a pretty interesting adoption like curve. And we all, I mean, we talk about it a lot about how fast ChatGPT worked, but I think it's interesting to see that models only, like they are only successful when they achieve this adoption layer. Like you have to have people who are willing to use it in order, like you could have the best tech in the world. Google Lens, I do think it was good tech. The— oh, do I have it here? Um, the Fire Phone, right? The Amazon Fire Phone. I don't know how many of you remember. It was good tech, like the hardware was awesome. It was awesome. I loved it. I still have it. I look at it lovingly, um, every so often. I'm like, oh, good stuff, man. Uh, but it did not get adopted, so nobody used it. Good tech gone to die on a shelf. I always walk through like, um, you know, in these stores like Ross or TJ Maxx, these are like US stores. You walk and there's like a technology area and it's like where all good tech that has bad marketing goes to die, like the never got adoption. And it's cool stuff you could see in there, like really interesting, like devices, sensors, and you're like, huh, I wonder why it's on the dollar shelf. Because nobody used it, nobody bought it. So similarly, fast forward to now, Robotaxi, we're in this moment where, you know, adopt or die, right? I just got a book from somebody who mailed it to me called Pivot or Die, but we're in an adoption moment where if people do not get in these cars at scale, like it's extremely cost prohibitive, right, to be in this business. So, so it'll be interesting to see how, uh, Elon, I mean, I don't know if you've seen his robotaxi dreams and visions, but there's a lot of people very interested in this space. It'll be interesting to see how they convince average people from Kansas and Ohio and northern Florida to get in these cars, because everybody I talked to is like, I'm not getting in that thing. And even AI people are like, I'm not getting in that thing. So one of the— this is how I solve all my problems. I'm a doer, not a talker, right? So I talk a lot in this call, but I do do more often. Outside of this call. But I like practicing. So there is a project at Amazon that I think is really, really fun, and I'm a Microsoft person for the most part, but I grew up at Amazon, so I get to claim a multi-cloud, right, cloud agnostic profession. And at Amazon there is this project called DeepRacer. I don't know if you've ever heard of it. It's very cool, so I'm going to show it to you. Really quickly. You could just Google or Perplexity this term DeepRacer AWS. It is an AWS project, but it started actually in the training organization, and the training organization was asked by the team, like by the executive and the leadership team, like we need a fun, engaging way to like get people to understand autonomous driving, but more importantly machine learning. It was a tool to use to accelerate people's understanding of machine learning. But what we would do in those early days is that we would— you would train this model basically in a 2D environment. You would, old school machine learning, train the model on your desktop, on your laptop, and then you would feed that model into a remote control car, and then you would put that remote control car on a track and you would race. And I don't know how many of you remember Grease, but our team would race for pink slips. Like, you would— if you lost, you got to own that person's car. And every week you got to bring it back, right? And you could either— you could keep it. Some people were really not cool and like would host other people's won, like, cars on their desk and be like, I own it. We had one car that like a lot of people beat, and so it would rotate around the office. Um, but it was really fun. But what I realized was the outcome of learning how to build an autonomous vehicle is that you, one, get empathy for what it takes to do that work, but two, you actually understand how the decisions are made and that it's not just like, you know, I think a lot of people think it's this big robot in the sky that has red eyes that's like, I'm controlling your vehicle and like driving you around. I heard an NPR podcast recently, and I was head of engineering at NPR during COVID and an interesting election year. And just recently I heard a podcast and they were talking about the perception people have about artificial intelligence and that we're right now seeing, you know, if you ask an average person on the street that their perception is very close to like Asimov's perception, right? Like an iRobot, artificial intelligence, the movie with the I see dead people kid. Like that's the perception that a lot of people have. And so obviously if we don't do something significant to shift that perception, they're not going to get in a car that looks pretty alien. So that's the one thing I think, uh, Robo, the Zoox, uh, the Zoox cars have going for them that Waymo doesn't, is that they actually do have a form factor that's pretty appealing. Like, they don't look crazy. Um, the Waymo ones look a little weird. They look like spaceships. So I think it's cool, but most people don't. Anyone here been in one of these things? Um, what's up, Joe? Uh, I, yeah, I'm always curious. So let me go over here and share my screen. Here we go. Hopefully the resolution will be awful. There we go. Um, okay, and then I'm just going to type in AWS DeepRacer. This is kind of dangerous because if I accidentally type it in incorrectly, I could end up on some weird deep racer, like, anyway, um, a AWS. I was just there recently. There it is. So this is the project itself. Now here's the cool thing. If you're gonna go to AWS re:Invent, they do this live on site. You don't get to race for pinks, but you do get to race. Oh, I should have brought it down. I didn't know I was gonna talk about this today, but you get a racer jacket if you win. And so I have this really cool racer jacket. I actually won it the year that AWS re:Invent turned 10. So it was really cool. Anyway, so you can see here you can build a model. So in the AI Leadership Institute, I have committed today, I was like, oh, this is cool. I'm gonna totally, we're gonna do a hackathon with Deep Racing. But you could go do this yourself too. It's, it's just as fun. You actually get to race in a community with strangers and beat other people, or you'd be surprised how difficult it is. So you might, it's like data scientists. It's so good because if you're not a data scientist, that's okay. You do not need to have data science experience to do this project. But what you will realize is that you will make assumptions, as most data scientists do, that will change the behavior of your car, and you won't know it until it's in production, meaning on the track. And you're like, no, but, oh, oh, I meant to— oh yeah, I didn't, I didn't think about that. So you will see people run off the track. You will see people not be able to get above 35 miles an hour. You will see people, like, be completely governed because there are default thresholds that you have to change. But it's good because it forces you to think like, what happens if you are about to hit a tree? What happens if you're about to hit another car? Are you going to support— like, what is that in like Formula 1 racing where like they have this concept of like rubbing other cars or bumping up against other cars? Anyway, I watch it. I don't know the lingo, but it's kind of neat to take that concept of something we all know. Most of us know how to drive a car, or at least the concept of seeing cars driven, and then build a model to support it. So, so worthwhile. There are events. I definitely have run these events in AWS Essentials. Like, they do these big, you know, conferences in regional settings, but definitely something I think you could— the other thing I was saying is any of you who are close to like a school and you want to just pay it forward, like, this is the kind of stuff that gets kids very excited about it. So this is what the Debracer looks like. It's a pretty cool car. You'll notice it has sensors and it's not cheap though. I will tell you that. But the good news is that you can build your model. So I'm just going to click on the Get Started. Oh no, I can't because I'd have to sign in. I want to do that yet, but we're going to do this together and I'll record it so you all can see. You can see the recording, but only members will get to do it live. But it's pretty cool. So I encourage you, if this is interesting, I do think it will demystify it. But the most important thing is once you learn to do this, you could teach other people to do it through a really cool vehicle. So some of us are teachers.

Yeah.

So you could see this is the 2D experience. Look at that joy on that guy's face. He's like, this is awesome. So I also have used it in executive retreats. So fun. These people are ridiculously competitive. So yeah, you could host a live race. I guess really cool that this was last year. So at re:Invent, they hold champions. Yeah, see, those are the jackets. So you get a jacket. Anyway, it's pretty cool. So, so I just thought I'd share that with you because you might not have even known that you could do that. You used to be able to actually go in. Let me go to pricing. Yeah, so this is the, the pricing for infrastructure. It's not very expensive, but it does cost. However, there is free tier support. So you'll see here you can actually get started with DeepRacer. You get 5 gigs, like it allows you to use it for 10 hours, which is a lot of training time, enough to get a model running, enough for a hackathon or, um, something like that. But if you get serious or like me, a little addicted, you probably will open up account, but it's relatively inexpensive. You'll just wanna make sure you build things for those of you who are engineers, um, in cloud services, you wanna build things in a resource group and then set an alarm. To turn that resource group off when you leave. Otherwise you'll wake up to a— I don't know how many of you have ever experienced waking up to a surprise cloud bill, but it only happens to the best of us is what I say. And I've had some doozies where I think it was Elasticsearch that I kept running and I was— it was thousands of dollars out of my own pocket and I was like, no. So I didn't make that mistake again. But anyway, so I thought I would share that with you. There is lots of training, and you, like I said— oh, I was going to show you on Amazon. When you search for this on amazon.com, uh, another very scary thing to look up. Uh, all right, if I go to amazon.com and type in DeepRacer AWS, you'll see you can actually buy a track. That's kind of cool. So I'm just giving it to y'all so that you have like just little seeds planted in case this is something you decide to do. But it's kind of cool. So I'm always looking for ways to kind of learn and be excited. That being said, I also want to share with you some foundational courses that are available right now. There's a brand new free course. Now I am going to— actually, let me go back and share with you what I'm talking about. I hope I can get there from here. But this is in GitHub, so I'm a believer if you are inside the I Love AI community, if you scroll through the classroom, you're gonna see a bunch of resources. It includes like my GitHub repos and all the things, all the GitHub repos that I use to either teach myself or teach others. So I'm gonna see if I can get there from here without having to sign in. Yay, I'm already in. Yay. Alright, so I'm gonna go to Agentic AI for Beginners, but Microsoft has created a series of beginner classes, and many of these beginner classes actually have video content to go along with them. For those of you, uh, who know, we actually delivered a live version of this Agentic AI for Beginners course on Maven a month ago. It went very well. We had over 150 people join, and as a result of completing our class, uh, members got an Agentic AI, uh, practitioner certification from AI Leadership Institute. You won't get a fancy cert for this, but you will get knowledge, which is very important. Let me throw this link into the chat for you so you can go grab it yourself. There you go. Oh, cool. Thank you. Thank you. Okay, awesome. All right. And so what I'm going to share with you here is it's got about 11 topics, but in each topic, and it's also translated, which is kind of nice. Um, it encourages you to build. Now, when I first started using this, I think, I don't know, there was less than a couple hundred people that had seen it. Now it's a little bit more popular because Agentek is taking over the world. But if, again, I'm, I'm, I'm a big believer in DIY. It's how I learned everything. I always say I graduated from the University of Barnes Noble. I would just go to Barnes Noble, grab a technical book, and like read it and do all the exercises. Today I'm facilitating that a little bit with this conversation, but also giving you a couple homework assignments if you choose to go and DIY your way to agentic awareness. But if you look here, I'm going to scroll down. These are the lessons which we teach, but Microsoft also provided videos as well. So introduction to agentic AI frameworks, design patterns, tool use. Agentic RAG, Trustworthy Agents, Planning, which is another design pattern, Multi-Agent Design Pattern, Metacognition, probably my favorite, and How to Use AI in Production. And then they just recently added MCP, which is why there's no video yet. But we— I have videos for all these too, but I'm just letting you know that this is— and this is just the beginning because you'll notice here this magic word, other courses. So they have 6 other courses. They also just released an MCP course, which I will be teaching. Because I like to teach and I find it fun. And yes, just the beginning, exactly. Um, and I just find sometimes— no offense to any trainers at all, everyone's got their own style— but sometimes I go to learn something and it's so boring, like I can't get through it. And I just— anyway, I'm like, I need to put Noelle's vibe on this. So it's not that the content doesn't exist. You can't— I always tell people, here's how to go do it for free. Um, but I just believe that there's got to be an exciting, fun way to learn this, and there is. And so, and, and so I'm doing my part and I hope you will do yours. But these are all the other courses. What I— so MCP GenAI for Beginners, ML for Beginners, Data Science for Beginners, AI for Beginners, Cybersecurity, Web Dev, IoT, XR, like a goldmine. And what I like about it is that they all contain text and code so you can learn by doing, not just reading, and they also include presentations. So if you go into this, like, let's say, Intro to AI Agents, it's got a README full of content that tells you what to do. So, and many of you probably use today ChatGPT or Claude or, or Gemini or some model to help you with all the things. I hope you do. And if you do, you can literally copy and paste this code into one of those models and say, teach this to me like I'm in 5th grade or 2nd grade, or As if I was a Star Wars character. Um, you can have it put it in words that you want. That's what I try to do with my own personal approach to teaching. But this is such— there's a whole team putting this content together. It changes all the time because, well, the tech changes all the time, but it's really, really cool. So yeah, you're right. Yeah, we just need— now we need to figure out a way to clone ourselves so we can consume all this information. I also am a believer in using like NotebookLM, right? Like putting this content into the podcast. Where was I? Oh, Google. So I've been a beta tester on some Google features recently for the last 6 months or so. And Google just released new AI mode features that are pretty cool. One, I get to add at the tool use pattern I just showed you. They've embedded the tool use pattern, so now I can tell it what tools I want it to use to to solve questions I ask every day. That's new. Very NCP-like. Oh gosh, I'm gonna lock myself out of my own phone. I'm gonna try. Let's see here. Google. So when you go to Google on a desktop or on your phone and you are in search mode, there's a little— it looks like Copilot, honestly. It's a little rainbow. I don't know why that's a thing now, but a little rainbow-colored magnifying glass, and that will put you in AI mode. And once you're in AI mode, you can basically add tools like your own calendar, your own inbox, your own, um, Google Drive and all the things, Slides. Now everything becomes accessible. So when you start asking for things, it's much more contextual. It's not great for me because I have so much content and not all of it's great and a lot of it's competing. So I ask it to triage my inbox and it can't tell junk from not junk. And, and honestly, I can't either sometimes, so I get it. But the fact that these tools are now integrated, I have the same thing is true in OpenAI's ChatGPT. They just recently announced their MCP integration of all, and I mean, I think there's like 200 or 300 connectors now that you can integrate with. So I'd encourage you to at least start doing that. Similarly in that I mentioned earlier the Y Combinator talk that Sam Altman mentioned. He also talked about in that, that like, imagine a world— and I, I'm experiencing this now because I've been using OpenAI's models for 3 and a half years since— I, I mean, technically I was in the beta during COVID but I didn't think to like maintain the same account or anything like that. So I got reacquainted with it in 2022, but I've been using it, the same account, since then. So it's I've been managing all my custom instructions, managing memory. I was very intentional compartmentalizing memory at first, ingesting memory through source files that I was basically agentic or RAG, retrieval augmented generation, with my own content. But now I don't have to do any of that. Like all that stuff is getting easier and easier. All I have to do is point it to Google now or point it to Microsoft and it automatically has this context. But what he talked about, what Sam Altman was talking about on the call or on his, um, in his session, he was being interviewed and he was saying that the level of value you get from a model is gonna correlate directly to how much it knows about who you are and what you want and how clearly the model can identify what's valuable to you. And so my results when I ask a mo— my models, 4 things are very specific. They sound like me. They're in alignment with my core values. They even question things that I'm doing in alignment with the things that I've always done in the past. And I think it's interesting because I now have to create challenger models that like challenge the way that I think, because I don't want to be who I was 5 years ago, and I don't want a model that's going to gently, right, course correct me to stay who I used to be, right? Like, so anyway, so I'm going through some interesting evolutions with these models, but it is extremely valuable that it knows what I've done, that it knows the books that I've written, it knows the companies I've worked for, it knows my OKRs, it knows the, the objectives and goals I've set year over year and how those have changed and how my priorities have changed. So I think it's important to do that. Now, I will tell you very importantly that I, in all of my accounts, For example, on OpenAI, I have an enterprise account. No, yes, I think it's an enterprise account for my company, and I have a team account for my family. And so at the bottom of ChatGPT, if you're not careful, maybe I'll go take a look at it. Oh, at the bottom of ChatGPT, I'll just tell you and then I'll show you in a second. It will say something like it could be wrong, and there's a setting in your personal settings that allows the data you're typing in and getting back to be used to train OpenAI's models. In my family, especially cuz my kids have accounts, I don't want that to be true. So I've locked down that access. I could lock it even further and create like a machine in my house that hosts a local LLM that then they get access to, but I didn't do all that. Um, but I did take the step of at least privatizing that model and every company should do that, right? I think a lot of companies are doing that, but If you're not, that's a first step is just if you're going to give it information about your healthcare and your finances, which makes me very nervous, make sure that you're not training public models with the information that you're providing and that there's some indemnification, like there's some responsibility that that company has. I was just talking to some, someone recently in a podcast I was interviewed on and they were sharing, they were asking me, you know, if I'm concerned about anything. And I would say my— probably my biggest concern is that the average user doesn't have any idea about the consequences of use of systems like this, right? That they don't recognize— number one, they don't even know that they can store memories so that it knows about them. But they also don't know that they can make those memories only theirs, that there is a privacy setting. And we don't make it easy. And it, you know, unless you know to ask the question, you won't even know that there's an answer. And most people don't know to ask the question. So that makes me worried. That's why I do what I do. That's why I always go live on LinkedIn. Like, the more people that can hear this, the more people that go, oh, I should privatize that. I should pay $20 and get my own private version of this model. I should build my own tenant. I should create my own, you know, infrastructure to host this. So something to be thinking about.

Hey, Noelle.

Yes.

I was listening to Matt Wolfe being interviewed earlier this week on a podcast. He says something he's gotten in the habit of doing on ChatGPT is a verbal journal every day that talks about like what he saw the day before that excited him, what was worrisome for him, the books that he read, articles he read, and his summary of them. He just keeps talking to it so his ChatGPT really gets to know who he is both personally and professionally. To help him with stuff down the road. And I would never have thought of doing that.

So I love that, Tara. Yeah. And I, I've called that like the captain's log, right? For those of you.

I love it.

Right? Like just getting in the habit of every day kind of logging. I mean, that's literally, except now we have an LLM that's going to interpret that over time that we can go back and be like, hey, I mean, one of the, my favorite experiments that I've run is just asking it what it, I ask it who it thinks I am. And that was another question I got at the— in this podcast was what I call my ChatGPT. And I think my answer surprised the host because a lot of people have nicknames for their chats, like Chat or Chat, or I don't know, all these names. But I don't anthropomorphize my AI systems. Like, I typically will call them by their name, like whatever the name is, even if it takes longer. I also don't typically add a gender like he or she. And I think this came from my early days at Alexa where one, I was always asked why Alexa was a woman and I would always have to correct them and be like, Alexa is not a he or she, it's an it. Like it's not anthropomorphic at all. Anything that you feel is comes from the data scientists that help build it. And I think, So that's, it's an interesting behavior change that I saw, but I actually know from listening, many of you know, I take care of my dad. He's actually coming home tomorrow. For those of you remember, he was in the hospital. He comes home tomorrow. Thank goodness. Uh, but he lives right on the other side of this wall and I hear him talk to Alexa in a very like, like he's like best friends with Alexa. He's like, I love you, Alexa. I'm so grateful for you. Thank you for telling me that. I couldn't remember that. Like he talks to her like a friend. And there's benefit to the anthropomorphization, right, of these things. But to a point, right? It's such a slippery slope, these conversations, right? Because now I'm like, we are now seeing massive adoption trends in the AI companion space, but specifically in mental wellness. That in the early days of AI companionship, it actually improves mental wellness. But then there's a tipping point, which we don't fully understand, where it starts to degrade mental wellness. Interesting, right? Like, oh my goodness. But good news, there's lots of research happening around this. Lots of people care about this problem. Maybe you care about this problem. But it makes me a little, you know, makes me worried because when is that moment? I would— oh, another thing that we talked about on the podcast is I was like, you know, she asked me what my perspective was on my children interacting with AI. And I was like, oh, and I told her that we have our own environment. We have a philosophy in our house, be a producer before you're a consumer, which means we don't go to AI systems with a blank page. Like we don't go, hey, I don't know what I want, help me figure that out. We start first with the synapse, you know, the synaptic activity that is idea creation and innovation and invention. And then we go to a machine and say, hey, I've got this idea, help me with it. But we don't start with nothing. We don't go like, hey, I wanna sell stuff to people. How do— how should I sell it? And, or even worse, what should I sell? Or what should I even do? I've kind of encouraged them to think through that first draft before they go to a system. And this is my kids, mind you. But she, she was like, well, I think we're, we're moving towards— I don't know if you saw OpenAI and Mattel just had a agreement recently, right? OpenAI and Mattel have a relationship now where they're going to infuse OpenAI large language models into toys. And I can't remember if it was Sam Altman or the CEO of Mattel, but one of them said that this will become your child's best friend. This, and in a good way, right? Like they weren't like, they were very much like, it will be so good 'cause your child who maybe feels alone and doesn't have friends or is bullied at school will have someone to talk to. And in my house, this is not how we roll. Like, I'm like, you will— you can have a toy that you could talk to. I mean, I don't know if you could see, like, I have lots of robots. Um, half of the robots in this office you can talk to, but not to serve your needs for human connection. Like, that's what the dinner table's for. That's what— when I stop and you get off your phone and you look me in the eye and we have a conversation. Like, I don't know. So something to think about. Is like the impact that the evolution of this technology will have on human connection, right? Like I'm always saying as a philosophy that I want AI to amplify human humanity, amplify human connection, give us more space to become more human. But as we start creating these relationships, right, in these companies that create opportunity, I don't know, I would love for you all to take a few minutes and talk about this. We are at that time in our where we pontificate together and learn from each other. I encourage you to, uh, we're gonna break out for about 10 minutes. If, uh, we're gonna set up little breakout rooms. If you've never been in this call before, you're gonna get a breakout room with random strangers. Uh, I encourage you to go around once, uh, and spend less than a minute just saying, we call it the I help statement. This is your chance to say, um, what you do. And we have a phrase that we use. I'll give you the template. If you can adopt it, great. If not, just say, I'm Noelle Russell Russell and this is what I do. We're trying to keep it short though, so everyone gets a chance to not only talk about that, but also give their perspective on kind of the future of AI and, you know, mental health. You can either choose mental health or you can choose like, you know, what happens if toys have bar— like if Barbies become your best friend and can talk to you like an LLM? Like, are you worried about this? Do you think it's awesome? I would love to just have you guys talk about that, um, as an opportuni— yeah, VTEC. Exactly. Yes. VTEC.

Exactly.

Um, yes, so many values. So what do you think the good things are? What do you think the dangers are? Good opportunity to have a conversation. So your I help statement is, I help whoever you focus on. So if you are in a company that's in financial services, you likely have a customer that you help. Maybe you're a consultant. Who do you serve? Maybe you are in a nonprofit. What is your cause? So what do you do and how do you, what problem do you solve and how do you solve it? So oftentimes I will say things like, I'm Noelle, I, you know, I help career professionals in transition learn about artificial intelligence, and I do that through my I Love AI community. And that's it. Takes less than, what, 15 seconds to say? But people, if they are a person in transition and if they are interested in learning AI, when they hear me say that sentence, they're gonna be like, tell me more about that. Maybe not immediately, but they now know this is who you come to me for. So we use these times together to practice telling people what we're, you know, useful for in their lives, how we can help them in less than 30 seconds. So take your time, practice, make sure everyone gets a chance to talk. You'll hear me, because it's super fun, uh, come over the airwaves and let you know when we have 2 minutes left. That's a chance for anyone who hasn't had a chance to share to take the mic and, uh, and share themselves. So let me— usually I do a better job of, um, setting up the breakout rooms while I'm talking, but I didn't. So here we go. Hold on. And robots, if you get a robot in your room, I will be moving them. I'm going to try and move them right now. 1, 2, 3. Looks good. Looks good. Okay. Oh, nope. That's a note taker. Okay. I'm going to put note takers in room 2. All right. Let's move you to room 1. All right. 1, 2, 3, 4, 5. Okay. Let me move you to room 5. Uh, oh darn, I moved somebody I shouldn't move. Moving you back, hold on. Okay, hold on. Uh, there we go. Okay, let's see here. I'm just checking for robots. Um, where did I put robots? Oh yeah, in room 2. There's a lot of robots. Okay, um, room 2. Let's see. Thanks for working with me. 1, 2, 3, 4. 1, 2. Okay, 1, 2, 3, 4. Oh, there's another one. Um, Room 2. Ban to Room 2. Okay, I'm like, Lenovo, I got you. Um, okay, Tara, I'm gonna move you over here. There we go. I think we're ready. When you get the invite, please just, um, click the little button that says welcome in, and hopefully— hold on, last one. Okay, I think— oh, you didn't move. 4. Okay, I think I got you all. Ready, set. If you have a robot in there, I will move it, but I don't. I think we're good. All right, um, so I will also message what to talk about, but hopefully one of you in the room was paying attention. Okay, go have fun, and I'll see you in a second. Javier, I'll sign you. Hola. I don't know if I got some time to a room. I don't think so. I don't see you, so I'm gonna put you in this room. Hold on one second. There you go.

Thank you. You're welcome.

All right, let's see, let's see. All right, I think I got everybody. Oh my gosh, Sky, I'm gonna move you to room 6. Um, sorry, she might reject me. All right, everyone, so They're in breakout rooms. I'm going to like say some stuff to you all right now while they're in breakout rooms. Super fun conversations though. So if you ever want to be here live where you get to meet with each other and like hang out, I hope you will take the opportunity. All you have to do is join us. You can go to myname.ainoelle.ai. That's a website. It'll take you to our free community. You'll also see some of the other ways you can work with me. There as well. But I Love AI is the community. You just click on the link and you join us. And then you can join us on these calls that happen every Friday unless I'm traveling or it's my husband's birthday. And then I will, will move them usually to a similar area. So yeah, join us for the next one. These Zooms are super fun. We get to know each other on a personal level. We get to talk about important topics and share what we're experiencing. With everyone. So hopefully you've enjoyed the session today, and, uh, I will see you all soon.

Bye!

All right, I'm just turning off live streaming. See y'all later. Bye!