I ❤️ AI - AI Leadership Happy Hour

to learn from Noelle and soak up some of the positivity and energy with all, all of y'all. So that's me. Amazing. I'm hoping when this goes live, it's gonna stay muted, but sometimes it doesn't, and it creates this crazy echo. So we'll see it in 1 second. Oh, it looks okay. It looks great. Okay. I'm not gonna look at myself because that's very unnerving. Alright. Awesome. So we are live. If you wanna repeat any of this, of course, it'll be recorded and put inside the community. But it's also on my YouTube channel. So you're welcome to head on over there. I always like to start these calls with just a little bit of like what I'm learning, what I'm doing, how I'm playing around with things, invite you to play along with me, and then we'll just open it up. I ask you just ahead of that to be considerate of your neighbors. If you have a question, raise your hand. If you hear something like, feel free to just interject, but let's just be considerate of each other. Uh, it's always a challenge, right, when we get in these calls. I want everyone to be heard. Um, and then, but then the introverts don't feel like they want to like say anything. So I just like us to kind of like— that's why it's nice to be on camera. If we, we see somebody like, you know, wanting to say something, we can actually just invite them, uh, to, to say something. And if you feel like you're dominating the conversation, you probably are. So it's, it's okay to do that in, in certain circumstances, but also just be aware, like, if other people have their hand up, no, that's the benefit of raising your hand is that you set an intention like that you have something to say. If no one's talking, which does happen, then we'll just talk about the thing, whatever it is. So, but if you want some to, to interject or change the subject, just raise your hand, let us know that that's happening, and we will welcome those changes. And our community chat is always a great place to continue conversation. So I would love to see more of us just adding, you know, questions that we have, not necessarily that we have about, but maybe a question you solved and you wanna see how maybe somebody else would have solved it. It's a great place to learn and teach, and we've got really amazing people in our community with great deep expertise across the board. So this is a great place for that to happen. This is our free call we do every week. This is my huge joy and pleasure I prioritize this call because it's a chance to just hang out with my people. So almost all of you, I don't do any like advertising. I don't like, you know, have Google Ads or Instagram. I mean, technically, I guess Instagram will hear me talking on your phones, so you might get something. But I didn't do nothing. I have no ads running. It's literally word of mouth. So either you shared this with someone, or I got on a stage and And you heard about it, or I last night, every trick-or-treater got a card to join the AI Leader— the I Love AI community. They also got candy. Don't judge me. But they did get a little I Love AI sticker. They were like, what is this? I know. But the cool thing is, is like with some of our members, like, your kids actually will know more about this than you might initially. So you show a kid an I Love AI sticker, and they're like, what is that?. And then, um, yeah, I'm very big on the TikTok with 7th graders. So if you have a 7th grader, they might know me already. Um, that's like, I, you look at my demographics of how I, how I'm landing, uh, with my content on the internet. I'm like, oh yeah. So if you're born, if you're a 7th grader, 8th grader, I'm like, I'm speaking your language. I'm an 8th grader at heart. That's why I don't know if I have one handy right now, but I have like this cool glitter sticker that says, I love AI, but it's like literally a sticker I would have been so excited to have when I was in 8th grade. So I, I, I live and breathe it. Some updates for me. This week was an amazing week. We started off with our very first live bot building workshop we did in DC. So that was super fun. We had over 50 people in the room at one time. I don't know how many people actually paid to be there because we just kind of let people in. But But it was, yeah, it's amazing. Oh my gosh, Kirk's in the house. So we actually have people.

Okay.

Yeah.

So it's so fun. I mean, now he might look at you and be like, no, it was terrible. But no, it was really fun and exciting. We had a lot of people who had never experienced it before, and we had the news present. ABC News was there. So that was really— it was really fun. And one of the benefits of it is up until this point, we had— we've taught lots of these, but I've only ever done it live inside a company. So it's very different because everyone already has a culture, and they already think a certain way, and they're encouraged to participate. This is the first time we did it live and not knowing anything about who would show up. So I thought it was extremely exhilarating and very fun. Lots of things we'll do, you know, that I learned from doing it. But we had about 8 people complete the bot certification all like live, and another 8 people say that they were gonna do it. Later. We'll see. Um, but everyone had positive things to say because there's just not a lot of doing in the world that we're in right now, not a lot of chances to just put your fingers on a keyboard and test things out. So it was a huge success. That was at the Voice in AI, uh, Voice in AI Summit and the Gov AI Summit. I've been in the keynote with Pete Erickson and the MoDev team for like, I mean, ever, since he started the Voice in AI, uh, conference, and GovAI was started after that. And so I've always been a keynote on his stage. He's very gracious to give me time with the audience. Um, but he also can relax, like when I take the stage, like he doesn't have to be worried. So that's one of the— it's a gift I give to him, and that is very, uh, reciprocal. Uh, because he— yeah, it's awesome to be able to talk to those people. But both government and kind of voice AI people, I've watched the exposition hall evolve from, you know, like a handful of companies playing in this space to now like every— like there's so many— there was so many companies, 50, 60 companies that are building AI solutions, which is awesome. But I'm that person who walks around and I did do it and I create a little bit of angst. But I would ask the question like, great, you know, how's that going? What do you do to build responsibly? Like, how do you make sure you're not using data you don't own? Like, what happens when you use copyright? You know, I'm asking all these responsible questions. Now, if you know anything about booths, typically the people manning the booth are not the people that have these answers. So it's a little unfair. But I like the people who are doing the selling, which is usually who's there, right? Biz dev people, ex— uh, business development executives, uh, marketing people. I'm like, you should know because if you create messaging for your product and you actually build this into your product, people are gonna eventually start asking, and it's better that you have a story now than waiting until your customers ask it. So, yes, who was there? I would love to know. It was me, Matt. Matt spoke as well. Kirk was there, of course. And Sarah was there. If anyone else was there, throw your name in the chat. We'd— I mean, we should have met each other, but if we didn't, now's our chance. But it was super fun. So we did that bot boot camp. The next day I did, like, my keynote. So that was fun. I think I came home on Wednesday. I don't really remember. It's kind of a blur now. I think you did because— Oh, right. That's right, because I had classes to teach. So Monday night while I was in DC, I taught the AI Solutions Architect course, which is a course on Maven. And then on Wednesday night, we did the Document Intelligence workshop, which I did call an audible at the last minute and decided to broadcast it live. But it was available to all of you all as well. So I— you would have seen some, some posts in the community like, I'm live right now, come see me. It was very fun. We had about 200 people join us. And they— most people that went through it and actually completed it were like, I never even knew what Document Intelligence was until I saw it pop up on my feed. So it was really cool to, to expose that. And that is the mission. So for those of you who are brand new to our community, welcome. One of the things that we— that I— the reason I created this was because in February, I was going through a weird transition period in my career, and I was trying to figure out, do I go on my own? Do I go work for a big company? I've done both before. And so I decided— it was February, and I was like, you know what we're gonna do? It's gonna be called I Love AI Month. And I already had a t-shirt that I had been wearing for years that said I Love AI. And I'm like, that's it. It's I Love AI Month. We're gonna do an I Love AI challenge. I brought in 5 different speakers. We had 21,000 people come and listen to that sequence. It was like 10 hours. It was a lot to ask people anyway, but I was surprised at how much, how many people joined that I didn't know. Like, I did run ads and I like said, anyone who wants to learn AI, come learn from me. It went very well. So we created the community, this community on that day, and I just invited all the people that showed up to that event. And we got like 30 people out of 21,000. 30 people were like, I'll join your crazy community. But that's how it started. And now just last week we passed 1,000 members. We have pretty high levels of engagement. We're right around 40% engagement. But again, that's only because I only recently started doing cool stuff. I have one person who's a big part of our community told me point blank, like, it's kind of boring. Like, what am I going to do here? And I was like, we need to do more live lessons. That's how live learning was born. And like, like I tell everyone, there's not enough doers in the world of a bunch of talkers. Lots of people talk about AI, very few people build AI. And everything I build, I don't need to know how to code to build. So I was like, I should just show you what I do. So that's literally how this community was formed. It's why we bring people together. We have this call to just get to know each other and hang out. We'll open it up for Q&A, uh, or just like conversation. I usually like to pick on a couple people that I haven't seen before, uh, and you don't have to— you don't go off camera just because I said that, um, you don't have to say anything, but it's a chance because some of us know each other. We've been here, you know, week after week. Those of you who are new, I'm going to invite you to share first, but I'm not going to pick on you, like call you out. If you want to share and you're new, please raise your hand so I could just like invite you to say who you are, what you do, give us a little context so that the rest of the community can know you. Most of us have done that on a previous call, um, and you'll learn everyone else because as you get here, you'll notice you're gonna have knowledge that everybody wants to know. So there'll be opportunities to talk. I do a lot of talking sometimes because I'm passionate about stuff, but most of the time my hope is that we— it's a happy hour, like we talk to each other. Um, and I want to give you all an opportunity to do that. Today we do have a hard stop in 45 minutes. But, but next week, in case any of you are around, next week on Tuesday, I'm in Charlotte talking to the Landscapers Association of America. Landscapers are AI people too. As a matter of fact, it's very cool stuff. I'm talking about AR, VR overlays, like really, most importantly, companies that have most of their field workers are non-English speaking. Most of the leadership is English speaking. So how do you create from a workplace safety, education, sales, marketing, all that stuff becomes much easier when everyone can understand what's happening. So landscaping. And then on Wednesday, I'm at FICO, actually. I'll be talking to FICO. They do an executive leadership mastermind. 75 of their biggest customers come to play. And I think you all— everyone here has heard me speak, or at least knows what I do for the most part. But I own a company called the AI Leadership Institute. And we go into companies, and we go from the boardroom to the whiteboard to the keyboard.. But in order to get me into the boardroom, I have an agency that books me on stages all over the world because of my experience. And so I talk to these big stages and then at the end I'm like, by the way, if you would like me to come educate your board or teach you how to architect these solutions or enable your workforce, I'm your person. And it's not just me, actually everyone in our AI Leadership Institute I call on to help deliver these solutions as well. Um, so, uh, so it's super fun because there's there's, there's a lot of hype, but this ends up being much more tangible for, for customers when they can actually see you build something right in front of them. And that's the, the philosophy.

Yay.

So thanks, Matt. Yeah, it was a really fun week. I should have— I wanted to do more, like, in celebration, but I went to Sydney, Australia, then DC, then, like, Rhode Island, and then back to DC, and then it's just a lot. But we'll be celebrating surely as I go into my quiet travel months of November, December, and January. We'll be doing a lot more together. All right. Awesome. So, Bob, I just saw your hand, so I'm going to pick on you. I want Bob to introduce himself, say hi, and also, you know, anything you want to share for the community to know. And then I will head on up to Roxanna.

Right.

Thanks for having me.

Great to get to know you guys. A quick geography quiz. Anybody know where that bridge is? That's where I live.

Nope, I don't.

It's in the heart. It's in the middle of Texas. So I'm in Austin, Texas. That's the 360 Bridge.

I love Austin weird.

Yeah, there we go. So that's the 360 Bridge on the northwest side, and my house is just across to the right.

Anyway.

Oh, cool. Yeah, so I've been in technology my whole life. I'm a degreed electrical engineer. I've been in both large, very, very big companies and then I've been at 3 startups. I just finished, I just exited a startup that I started from scratch. I've been advising the last year to early stage startups and then Fortune 500 that's actually spinning out of Intel. And through that, I had a couple of light bulb moments with regards to AI. That I think fundamentally are gonna change my life anyway. And so here I am.

Yeah.

So that's a bit about me. And then the most important thing in my life, I'm a little bit ahead of you, Noelle Russell Russell. So I've, I've got 3 incredible kids. They're all older, but I've got, I've became a grandpa a year ago. So last night I took my, my grandbaby Brayden and walked her around my neighborhood where I walked my children for 10 years or 15 years. So that was super cool.

That was really, really cool. That is awesome. You're right, that, that day is coming, I hope.

Yeah. And I'm blessed, I'm blessed that all my children live in Austin and they want— they actually like being around me. So it could be because we, we have really cool ski boats and we all like to surf. But yeah, I'll just blame it on that. That's right. That's what I do.

I try to make it as nice as possible. So I'm like, really? I mean, do you— anywhere else is not going to be awesome.

So anyway, that's, that's a bit about me. Yeah, and there you go. So that's a bit about me.

Awesome.

Thanks, Bob.

You can tell my school affiliation I wear proudly, and I'm sad to say both of my oldest children went to A&M, but that's beside the point.

Amazing.

The Aggies will say I raised them right.

That's amazing. All right, Roxanna, your hand went down, but you still good to say something? Okay, great.

Yes.

Then I'll come up to you, Kristin.

So I've been with the same company for over 30 years. I work for a utility company, the only company that I work for. So I have an engineering degree and a law degree. So I am a lawyer. I've been selected to be on the AI committee for the law department. So we are providing guidance to the company as a whole, which is approximately 14,000 employees. So we really have to catch up very quickly. As an attorney about AI, right? So whether our outside counsel uses it, what we— what are we going to put in our contracts with others using it, using our data. But I'm also the family matters— we have ERGs, and I am the lead for high school/college kids. And so I promote things. So anything that you can help me with, whether to proceed with STEM, you know, that's something else. You might have seen my little notes to you saying about your AI gear. I don't know if you remember that. That was me. So we're quickly trying to acclimate. We use, which is very strange. I work for the utility company that I work for in New York. I apologize, I'm from New York, uh, Con Edison, um, which is the, the largest utility company here in New York. But in any event, um, um, so I, you know, we use Copilot, um, unbeknownst to me everybody has access to Copilot but only certain individuals have a license. I am one of the licensed individuals, so I have to be hush about it because, you know, you pay a price per month, and the company does not want to promote that, right? Because it's very expensive. So I have to teach people, one, how to use it, how to protect it, you know, when you record on Copilot, just all these guidelines that I really have to learn relatively quickly.

Yes. Oh my gosh. Well, I don't know if you know, but I have an AI, uh, a, a, um, Copilot, Microsoft Copilot super user training, so I can send it to you.

Yeah, oh, I would love it. So, um, if you have— so if you— maybe we could DM or email or phone. I'm not sure. What do you want me to do? Um, what would be the best way to communicate?

I'm gonna put my email in here.

Okay.

Um, .ai, and that way I can share with you the details. Absolutely right.

So if we want to explore maybe speaking, you know, just to get our department to use it really, because you really got to catch up on it quickly. And I kind of teach people, you know, what I can, but I need to be taught.

And even a train the trainer, right? We do— like, I often will teach the first one for not just trainers, but like for the first group. But in it, you seed like 5 or 6 people that'll then teach it on behalf. I'm a big believer for many reasons in teaching people how to fish as opposed to fishing for them, which is like, here's a prompt guide, just use the prompt guide. Like, there's, you know, it's, it's a process.

So, yeah, so I support you. You're female, obviously. You're Hispanic. Of course I got to support that.

And I'm from New York. I was born in Manhattan.

Right, right, right. You have 3 children, right?

6.

3 under the age of— under the age. I was going to say of 10, but my 10-year-old just turned 10. So 10 and under.

Yeah, no, I have 3, and luckily they're all in high school, so I'm so glad to get them out. Um, I can, so, but I know the others want to speak. I could speak forever, but, uh, nice to meet you in person, and I can't wait to chat more.

Okay, awesome. Thank you so much. Hi, Kristen, how are you? Am I pronouncing that correctly?

Yeah, you are. That's great. So, um, I saw Nicole— Noelle speak at the Grace Hopper Conference earlier the month of October. It was great. You you were on a panel. And I am from the state of Utah. We have a pretty strong, like, tech ecosystem here. I run a nonprofit called Women Tech Council. It's all about building community for women in tech. And I'm here mostly— I signed up, I was like, I love AI, you know, your, your promotions on LinkedIn. I heard you speak, and I was like, I just need to stay on top of trends, and I want to know more about AI. So so that I can bring kind of like those topics that are really trending and hot to our community. And similar to what Roxanna was saying, learning more and taking it to our— we have a high school program as well. So all the things I want to learn and grasp and hang on to the things that are discussed here so that I can kind of take it to my community as well.

Awesome. So good to meet you. Yeah, so there's one thing I meant to say before, but for both of you, You probably know this exists, but in the off chance that you don't, there's a resource. I should probably type it in and then copy and paste it, but it's called RAISE from MIT. Are you familiar with it? So it is really great content. I'll share. I'm just gonna type it in there. Oh, see, it came right up. But yeah, if I try to type it in for you guys, it'll— I'll mess it up. So I'm gonna paste it. But here is the the website. So this is an initiative started by Cynthia Breazeal. She is awesome and amazing. She cares a lot about what we're saying about raising the next generation. They created— if you go, I actually should have put you in the resources section, but if you go to the resources section at the top, they have a whole curriculum on responsible AI and a whole bunch of labs prebuilt. It was created for educators to teach. Like, it's got full-on courseware. It's amazing. But educators won't teach it because they're like, I don't know nothing about AI. I have no business teaching AI. So there's this gap with this incredible content that's created, and then the audience they built this for is not feeling like they can deliver it. So now it's not getting delivered. And it's super unfortunate because in the hands of people who know just a little bit, as you know, Roxanne, like, like, it doesn't— you just have to do it once and then you become more educated than everyone else. So a lot of this is really cool. They're very fun games. I've actually used this to teach adults because it's fun and engaging. I don't know why we stop being fun and engaging when we grow up, but, um, I think all content should be built this way. But it's like, uh, like superhero-themed stuff. And anyway, and I don't know if y'all saw who I was for Halloween, but I was in the multiverse in Spider-Man and dressed up as Gwen. If you know, you know, but it was cool. I got a wig and everything is— I was blonde. You might not recognize it on LinkedIn if you scrolled by, but I was like a blonde hair with this weird asymmetrical cut. It was so fun though. It was so fun. But that being said, this content is very relevant, very cool, very interactive. So I share that with you and I just tell people you could build an entire company right now. There is massive amounts for women-owned businesses, underrepresented businesses in government. You can get funding like not loans, grant money to start a company to just teach STEM because there are tens of millions of kids that are going to need this education and there aren't enough, like I said. And the educators are like, that's not me. It could be them. So you could also educate educators if you wanted to do that. But just a huge, huge opportunity. And I will also use this as a moment to share with you my favorite story of the week, which is a young woman stopped me at Voice Summit, and she was like, Noelle Russell, and I'm like, yes. She's like, you probably don't remember me, and she's a young woman. You probably don't remember me, but you spoke at an event and we took a picture. Apparently I take selfies with everybody, but we took a selfie together. And she's like, we spoke at this event, and you— it was the first time I saw someone who looks like me on, like, talking about data science and this career journey and how you ran into trouble, but you still did it anyway. And so she recapped, like, my story, and I was like, Yeah, of course. And of course, I'm like, of course I remember you, which I didn't remember. But she forgave me. But she said, like, that at that time, she was not foreseeing herself moving to a career in this area. And just seeing— that's why I say, like, all of us, you have no idea who's watching you and what you do. But she saw me and she's like, I could do this. And so today, she's now celebrating her first year as a data scientist at a company. I could get all cryy-eyed about it. But I was like, oh my gosh. She was like, you changed my life. Like, you changed my life just showing me that what you do is possible. And I was like, I've got— I've had this feedback 3 times in my life, and it's the best thing people can say about you. So anyway, I'm all crying now. Dang it. It was so good. I couldn't— I was like, I hugged her. She took a picture, and then later she took both our pictures, like 2 years ago, us, and then the next— it was so good. Oh, like where they were then and where you— yeah, like where we met, and I didn't know who she was, and then our picture together this time. And my husband told the person she was with, she goes, you know that this is like the highlight of all the things. Like, no matter what cool thing she does, this is going to be the moment. And I was like, you're so right. But that's like what— I don't know. Maybe not all of us live for this. But it literally is like a brown girl seeing herself in a role that she did not even imagine 2 years ago and now having gainful employment in this. Like, I was just like, yes, let's go. Screaming from the rooftops. So that's why I do what I do. That's why, you know, people say you don't take enough breaks and you don't take enough vacation. I'm like, I'm changing lives, you know? Like, I mean, I'm not going to burn myself out. Like, you know, don't worry about me in that way. But I do work hard because this is serious business. So that's really cool. Anyway, awesome. So thank you for sharing that. And hopefully this resource is helpful. And then definitely— so my email is there. You're welcome to email me. And if you want me to read it like urgently, you should definitely mark it as urgent. I'm just telling you. Okay. And I don't say that to everyone. You all are in my like little circle here. So I share that with you so you know how I read my emails. Awesome. Anything else, Kristen, or is that good?

No, that's good.

Awesome.

Nice to meet you.

Appreciate the resources already.

I mean, you've been here before, but we didn't get a chance to like just talk to you directly. So I'm glad you got a chance to introduce yourself this time. Thank you. All right, Jason, you're up.

Well, thank you. My name is Jason Anton, and I call myself a digital janitor. And the 2 questions I get the most in my life is, one, who are you and why are you in my yard? And then two is, what's a digital janitor? So the first one is— there's a lot of answers, but the second one It's mostly just, you know, I've been in— I've been an engineer of some kind for, you know, nearly 20 years. So I've worn many hats. I've done the ops, the coding, the, you know, pen testing, all sorts of different things, sort of like everything. So the easiest way to explain that is just, you know, I clean up messes, I fix things, I build what needs to be built, and that's, that's a janitor. So That's what I do. I'm here because I've— I'm in a sort of similar situation as you are, Noelle. I was like, I've been— I've never really done a startup thing before, but I'm doing it now. And I've been working in, you know, like I said, various capacities for 20 years. I'm still consulting as an AI consultant for the federal government, for the Center for Medicaid Medicare Services, trying to help them how to figure out how to do AI responsibly with this sensitive data that they possess. Considering the nature of how some of this technology works and the way government works, that's become a bit of a challenge at times, but not an insurmountable one. And my startup is building an AI platform that focuses on that data security and isolation and, you know, that sort of thing from the, from the beginning. So there was a really good sort of like intersection of what the government was looking for and what I was already building. Um, so that's what I do. But what I really wanted to sort of introduce myself with here, and what I really wanted to bring up, was, uh, yesterday was Halloween, as Noelle, you pointed out. But I didn't get to really go out and show my costume, and I kind of came up with my costume at the last minute. I was really proud of it, so I want to share a link to my costume post from LinkedIn.

Yeah, you go!

That's right, you go!

And then I want to explain what I did.

And, uh, uh, careful, Jason, like, this is a moment you could either be banned for life or—

uh, I'm ready to go. All those likes, let's go!

Yeah, we are, we're on YouTube.

YouTube, we're on YouTube.

Okay, like, this could go any direction. I'm a little bit nervous.

All right, so I sent the message. I was a chatbot for Halloween.

Is what I want.

What I did was I took— I have like an ROG Ally, which is like a Steam Deck kind of thing, and I just threw up a simple like HTML JavaScript thing because that's such a small device. And I have a, like, a little like knapsack. It's not a fanny pack, it's not a fanny pack, but it's kind of like a fanny pack. I put that on. I had like a small messenger bag, like a mobile monitor that I just duct taped to my chest plug that into the Ally, had a wireless keyboard. When you typed in and hit enter, it played on my bone conducting earbuds what you said, and then I would answer it.

Ah, that's cool.

I was a chatbot, and I'm a big fan of like low maintenance costumes but that are high concept. So I was proud of it, but like I'll never be able to do it again because I think the moment's passed, and like 4 people saw it in real life. So I wanted to share it profile picture too.

I love it.

Oh yeah, I got you. I mean, you get— I gotta get as many miles out of that.

That's right, 100%, as you should, as you should. Did you get video?

Wait, what?

Did you get video of anyone using you?

No, because only like 2 people. Like, it was an office happy hour. It's the only place— I live in a town that does Halloween really weird, so like, that was over a week ago that they did the trick-or-treating, and my kids are all grown, so like, I didn't really get to go out and do anything.

And I have to go to the office. You've inspired me, though. We should totally do Halloween something next year. Like on, you know, Halloween.

Yes, that would be fun.

We should totally do something because, I mean, I would have been Gwen, Spider Gwen, but this is awesome.

Yeah.

Well, I still have little kids, so—

Sorry, go ahead. I'm not breaking up.

No, no, no. Go ahead, Jason. This is awesome.

Well, 10 years ago, I went as Facebook. And the way I did that one is I printed— I took my profile, I really carefully organized it, and then printed it out and put it on like poster board, cutting out where the picture is to put my face in, and then just had a small like whiteboard there that people could write on the wall when the wall was still— I don't know if that's still a thing. I'm not on it anymore, but I feel like this is a really good spiritual successor to that concept.

I agree.

Very like an evolution of costume. I love it. I think it's awesome. I'm glad you got a chance to share it. You should totally post it in the community.

I was disappointed I couldn't share it more.

Yes, you should totally post it in the community, like in a post, and be like, what were you for Halloween? And then if anyone else has pictures, they can throw theirs in too. It'll be hard to compete though. But I just like— it's actually quite innovative. Like, you know, like I'm just thinking, wow, what could I do with a, you know, a pocket? I used to use, um, the— I built an Alexa Pi back in the day before we launched our product. I used a Raspberry Pi and then used the Alexa voice service in it and then literally connected it to— I don't think I have one handy, but like headphones. And so Alexa. Yeah, exactly. And so it was very cool. Eventually the device is cooler, but it was— I was like, this is awesome. But I never thought I could now be— I could do the same thing.

Oh my gosh.

I mean, I still mess around with like a similar sort of thing. So I take like the Raspberry Pi Zero or whatever Pi I have. And I've been— so I'm, you know, diagnosed autistic. So there's times where, like, I miss out on sentiment and things. So I've messed around with a similar idea where I have this and I sort of do something, you know, the Pi is just a relay, but do some sentiment analysis over what I've said or maybe what it's heard and then get either in text or sort of here. And it's not, you know, great, but messing around with that kind of stuff, like, you know, Pis are great for that. But with this, after I did it, like, the monitor is the problem. So I was like, oh, what if I could just get like a flexible LED and then like I could go play at like conventions and do it? And it's like, wouldn't people just talk to you instead of type to you though? Like, it's not really very useful. It's funny, but it's not actually useful, you know? But yeah.

I think that's funny. No, I mean, I don't know if you know, but they do have these like wearable devices now that have screens on them. Like it's, ah, We're not smart. Yes, here we come. Very cool. Thanks for sharing that, Jason. It's awesome. I love it.

Thank you.

All right. Awesome. Margaret, you're next. Nice to meet you. So glad you're here.

Hello, everyone. My name is Margaret Farrow. I was not an AI for Halloween. I was a giraffe. I am currently the sponsorships lead at a nonprofit called Tectonica. We help women and non-binary adults who are seeking economic empowerment through software development roles get into that first position. So we've got a 6-month fully open source project-based curriculum. And then at the end of that, we try to place them with a position at a partner company. That's generally a 6-month paid internship, but there's a little bit of flexibility in terms of length. And my current role is finding those companies to match people up. But this is a residency-style position. And outside of this, I've historically been in much more technical roles. I was a technical writer for 15 years. I've done a bunch of other technology role types. More recently, I was working in security. First, I was the security lead for an AI-powered gaming company at the time when we were one of the GPT-3 Alpha test customers. I stayed there through a couple of months after GA, and then I went to a consultancy where I was the head of corporate security. We focused on helping startups kick off their security function. A lot of what I was doing there also related to AI governance, AI safety, and setting the policies that were going to outline how those companies were going to use AI in the future.

Wow, that's amazing. I was like, I know, very excited to talk about any of these things at any time, to like eat lunch or I don't know. So you're in DC, you said?

No, sorry, I think I forgot to say I'm in Oakland, California.

Oh, Oakland, awesome. I'm actually going to be in Oakland eventually because I'm working with dream.org, which is headquartered out of there, another nonprofit, but in the advocacy Space. So if— when I get there, which will be sometime, I think it might be early January, but when I'm there, I'll definitely come and say hello.

I'll be around. Please do.

Oh my gosh, so good. Okay, so, uh, just so you all realize, hopefully if you don't realize yet, uh, this is a great place to meet cool people, and some of them are even willing to like have a separate one-on-one and talk about what they do. Um, not that I'm volunteering that for anyone. Set boundaries. That's another life lesson I've learned. But if you are willing to connect, you'll see Mauricio and Jason and some others that I haven't scrolled up for have thrown their LinkedIn profile. So not all of us are friends yet. This is a pretty small world, especially as we begin to really understand the difference between the hype and like reality. It's great to know some people in this space. And so I'd encourage you to throw your link in and actually click on them. While we're talking and connect. Don't just follow. That is my secret, my secret sauce. All right. And it's great stuff. Good to see you. Welcome back. I know it's like all day Noelle day. We might change that one day, but right now it's just all me all day. Welcome back. Say hello to everyone.

Hi. Thank you. Hi. Hi, everyone. I'm Stephanie Namaku. Right now I am like looking for a new role, but previously I was a prod— so I would, I guess I'll start back with my interest in AI. So it actually goes back to when I was in college. So I doubled in computer science and communication. And so in my communication classes, I was doing intergroup communication and organization behavior, and then like realized, oh my God, I am literally like the little node in all the research the communication classes are talking about within tech and women in tech and like knowing your percentage and then feeling like it's like I am the statistic in like in that space. And it's a really weird place to be, like a little loop. You're like, what's going on? And it's like, oh, like this is why this is happening to me. And so basically, like in my communications classes, I started analyzing at least how things were going on in my computer science classes. And then also like as the interview process was going, like forward and like having to do like the technical challenges, technical interviews, being flown out, all that stuff. And knowing that like, first off, a lot of women don't know about doing that, and we have a perfectionism thing and all that. So that's what got me passionate in bringing more women into computer science. And I was actually a co-founder for Ladies of Computer Science at University of Maryland. So trying to bring, like, retain, I guess, like, people in computer science and also lift each other up. Oh, did someone say they're also at University of Maryland?— that went there? Oh, sorry. But yes, so in college, I actually wrote a paper about, or like, pushed a— wanted to push a policy of our university to ensure that an ethics course is mandatory for computer science majors. They are for info science majors, but not for computer science majors, not people actually writing the code, which is crazy to me. And so That was something that I was pushing for because it's really interesting. And then when I was studying abroad in Madrid, I was taking an AI class and I was dating this guy who was taking an AI responsibility class. So he's like pre-law and I was computer science and kind of talk together and show each other our homeworks and kind of like kind of discuss the future of what AI is. And it's kind of weird that it's actually like coming to what we've been talking about, which feels really weird. He's also still like He's going to be a lawyer soon too, and we're still in contact. He's in Ireland. And so it's just really cool to see all these people seeing the same thing. And I feel like I'm just a big nerd from like the consciousness level with AI. And I don't know if anyone's read The Golden Braid, but that stuff is really cool. Just the intersection of art, math, and music. And then I'm taking an ink meditation class and it's just So crazy how like one little drop can like impact all the water like that you're doing.

Like a—

yes, yes. That's like life lesson.

Yeah.

No, if you want to, you, you could unblur your background so we could see it.

Yes.

I know it's extra effort, but it's so worth it. Uh, but you're so right. Like the butterfly effect, ripple effects, like it's so, so cool. And your experience, like, oh my gosh, cool. You can pin, uh, if you want, or double-click on it.

Oh, wow.

Yeah, it's an ink meditation class in DC. Um, it's Washington Studio School, if anyone—

I've never even heard of that before. I'm in Arlington, Virginia, DC area.

So yeah, I really like the course.

Yeah, I love it. That's a great idea. Um, that's so cool. So yeah, I mean, I also feel like there's such a It's interesting when like these things happen and then you're like, wow, I'm like completely set up to do a thing. And like being double degreed in comp sci and communications in a world that becomes generative, like those are literally the two top skills you need to have. That's pretty cool. And then yeah, also around, of course, the concept of kind of ethics and engineering. I've also been there where I was like, wait, what? Like I went to Embry-Riddle Aeronautical University, which is a of course, highly engi— like all engineering, not liberal arts. I didn't even know because it was my— astronauts go.

Yeah.

Yes.

Lots of rocket science. Like if you—

yeah, I have friends. I've also dated someone who wants to be an astronaut.

So yeah, but nobody— there were no ethics or what's worse in these technical colleges. Not only is there no ethics, there's like very little psychology, very little human interaction, communicate. There's no communications classes. So it's fascinating to me that we're breeding these technologists, hyper-technologists with no liberal arts experience whatsoever. And now, of course, our models are built like with language. Ironic. So that's very cool that you had that experience. Really, really exciting. And yeah, I have read The Golden Braid. I'd encourage others to as well. I also think that that's very like telling of where we are right now, where all of these disciplines are intertwining together. At least that's been my experience. And I think it's awesome because I see when I get on a stage, like, I can literally see a place for everyone, like everyone's unique experience, everyone's lived experience, like people who are creative, people who are musical, people who are artistic, but also on the other side, people who are very like me. I'm a, I'm a creative, but not in like a creative sense. Like I create things all the time, but I'm very like pragmatic in my approach. However, I'm a quick start, they call it. Like, I like to build things in the beginning. I don't really like dotting i's, crossing t's, shining things, repairing things. I appreciate good operations, but I don't really like caring about that either. I just like building. Like, if I could just build. And so AI teams, typically, it's exactly like this kind of set of skill sets. And, and because AI is constantly iterating, I get to constantly build. So I finally found like a industry where I could literally just do the thing I care about, surround myself with really smart people who could manage my LLOps and people who could help me do user research and who— people who could help me round out my customer experience. But I could just build and build and build. And that, I mean, I've never met a model I couldn't make better. So, so it's like infinite job security. So I really like that, that idea. So thanks for sharing that too. Alright, awesome. I don't see any more hands up. So it's a free-for-all. If anyone, let's see, what time is it? Oh yeah, we got 10 minutes. 10 minutes. Otherwise, I did write down some notes from earlier that I want to share with you about.

Oh, we have a hand.

Oh, good, great. Oh, of course, Daniel. I was wondering, I didn't see you on there. I'm glad you joined us.

I know.

No, that's all right. I'm glad you're back.

And just a quick question that I didn't have a chance to ask in the previous one. So specifically on chatbot building, right? You mentioned at some point that you had a whole session about people who wanted to start a business out of selling chatbot building, right? Would you give us more details where to find that session? And then the second question that I have is, So I built this chatbot and I'm about to pitch that to my daughter's childcare center, right? So the thing is, it's a big chain of childcare centers across the entire country. So they have roughly around 100 different centers and each one of them, they have information about the different branches, right? So what's the pricing strategy that you advise in that specific situation? Like one big chatbot to cover all of these? Or specific and separate ones to cover every single branch?

Yeah, so the way I like to approach these types of questions is around like the way we did things in other production systems. So at Amazon Alexa, we went through a very similar moment where we had to decide, are we gonna build one big brain that speaks one, like speaks a bunch of languages? So language was the biggest area. You're talking about locations, but location, language, domain, doesn't matter. But we had to choose, are we gonna build one gigantic model that'll do all the things, or do we build individual models? 10 years ago, we made a very specific, long, drawn-out, like, conversation, but very specific decision to go multimodal. And the reason for that, there was a bunch, but some of those reasons, number one is it's easier to maintain smaller, tightly scoped models because the bigger the brain gets, the harder it is to, especially synaptic like CNN type of networks, it's very hard to control how those synapse start firing as the model scales, which is exactly what we see in most AI models at scale is that eventually they get to this point where they start making, they create associations that you never dreamed that they would create because it's, that's its job. Anyway, so, so that's the challenge is that you make that decision We chose to build one for each to get tighter control over the accuracy of the model. But as a result, we massively grew the complexity of managing the model. So our MLOps was, like, awful. We had to have entire teams that built in languages. In your case, you'd have a separate team per location. So the infrastructure will expand in alignment with kind of the smallness of the model because you're gonna have to build one for everyone in that case. However, at Microsoft, we have— I've also built solutions where we just had one model, one model to rule them all. My— what I do now is when I go to a client and they're having this decision, I encourage them to build a single brain to identify and scope the context of what would be considered grounded knowledge. What is the in— What is the information that every one of those organizations need? And they will all need like home, about us, you know, all the stuff that's on like the brochures.

They all have their own knowledge set.

Yeah.

Does that make sense? Right? A shared knowledge set. And then you could create like, so that's the big model. And then it becomes like an agentic implementation, which is what we're calling it today. But back in the day, it was just like you could orchestrate smaller models to just use that model as its baseline and then customize. And this is what, like, Chat App Builder allows you to do. Like, you create a—

with different flows.

Yeah. You have a big—

oh, fair enough.

Yeah. And then you create different flows that trigger a new smaller model for the delta in knowledge for each specific location.

So you can have different—

like a one way or the other.

So you can have different prompts and different models per flow. Can you?

Right. Okay.

All right. That helps.

And even inside a flow. So inside a flow, you might have a decision tree, right? So if a decision tree breaks, you can create one OpenAI, like, one model that has a certain prompt engine, like, prompt, system prompt that governs a certain type of decision. And then in the same flow, have another model that says, well, if you didn't choose that, you're gonna choose this and have a totally separate system prompt for that subject.

Oh, that helps. No, thanks. That's so helpful. Right? So maybe I can have the, the first one being the kind of a router, and then I branch out to one flow per branch, and then I have the specific prompts on that. That helps. That's helpful. Thanks.

Great. And then when you go into that flow, just remember, like, the first node is going to be in OpenAI mode. So most people don't realize that that's not— like, they make an assumption that that OpenAI is like the main one, but it's actually a brand new one. So you have a whole new system prompt area that you can train just that model on making decisions.

Great point. Yeah, great point.

Awesome.

So just, just one more thing, Noelle. Can you please just remember to get me on the Maven course? Oh, yeah.

Write it down because I was afraid I was gonna forget.

Yes.

Do I have a form for that?

And also, it's a great segue to we also are running an AI incubator. So we're gonna run this virtually. I have an in-person one. So we're gonna run it in person in DC for the first 6 weeks of the month. This is like if you happen to be in DC, if you are in Arlington, The state has funded us to launch this. It is a super trivial, like, $3,500, but we go from, like, nothing to customers. So my goal is to help companies actually fund themselves through customer acquisition as opposed to the hope and dream of getting VC funding. Some of you will be able to get VC funding very easily. Some of us will struggle. And I think the best way to get VC funding, if you want it, is to actually build a customer, build a product you can sell like that actually people want and that they're willing to pay for. And so that is my philosophy in this course. But I'm not the only one delivering it. So I'm co— I'm partnering with a company called Unstuck Labs. So if you're in DC and you're launching a startup and you want to do it safely, responsibly, and with my help, we are launching it at Unstuck Labs. You can go to unstucklabs.com and just— you'll see it's called Hyper Accelerator. The other option though is to go virtual, and the virtual one, it's the same except you won't get to come into the office. I will tell you, non-virtual accelerators are harder because it's all on you. The discipline is on you, but I'm disciplined, so hopefully I'll attract some disciplined people. But we do all the things. It's like fingers on keyboard, building your business, acquiring your customers, talking to customers. Eric just joined the call, so hi Eric. And your hands up, so I will come to you right after Jason.

I don't know if we have time though.

Just watch the time. Oh, 4 minutes. Okay, so I'll stop talking so we can—

okay, we're gonna take Jason.

Jason will be our last ones.

Well, Jason also raised his hand, but just Bill.

Yeah, so I'll go to Eric first and then Jason will finish up with you since you asked earlier. Is that okay?

No, no, no, go to Jason. I'm okay. I feel like I've got like a loaded question anyway, so it's— it's going to take a while. I'm very interested in the thing you were just talking about. I think I might have signed up for it, but I'm going to look into it again. I mean, I'm literally doing sort of the thing that you were talking about with where the decision you made at Alexa with like smaller models and the scaling. And in my opinion, like coming from the perspective of like all the things, like the ops is like the hardest part to me to get right because there's so many moving parts. It's like that balance. And I'm, you know, it's all— I'm doing it all Kubernetes and everything. So like just that upward, that horizontal scaling and sort of serving models as a service within the cluster and all that kind of stuff. And my question was really like, are there— what got you as their life lessons? Could you warn me that I'm going to lose sleep over or that, you know, I'm going to pull my hair out? And again, that is a loaded question, so maybe I'll ask it again another day.

And do you mean specifically like on ops or just in general?

No, because like I'm— because yeah, like the ops to me is the hardest part to balance it right. But it's also like you get, you want to make, you want to fine-tune your models to a spot where like it performs well for what you're trying to do in the specific use cases. And you know, the way that I'm building the microservices, the idea is to be able to combine them to make different tools.

That's right.

Multi-use and reuse and reuse, right? So like there's a lot there that's, like I said, loaded question and I don't think we got.

Yeah, so I'll give you a sliver of it and then we could definitely follow this up in like a conversation on the chat as well. And I think there's a lot of us that would be able to provide some, some different perspectives. But the one thing I would share with you is, um, just overcomplexity. And that's what we suffered with at Alexa, is that our model became so complex. We didn't think about— so at chat, I was a beta tester on the ChatGPT models, and we also had this decision we had to make around multimodal, multi-head, single model. And again, we led towards build small models and then condense them into a single model interface. So ChatGPT became the interface, but it talked to 8 different models. Codex was one of them. It actually became the genesis of GPT-4 Omni. But like thinking about, I always say like, look at whatever you're doing and then think like, is this the simplest that we could possibly do it? Because so many times, especially with these, especially that's why I was asking you about ops, When it comes to operations, like, it's gonna get very difficult to test these models the bigger that they get. Because you're gonna have behaviors that will be only be exposed by people who know the right question to ask. The more narrowly you define its use case. And Sarah and I were just talking about this earlier. We are moving into a world where you're gonna wanna build these very small, like, purpose-driven, purpose-built models. Because we're then gonna use, right, agentic, right? Is that Monique? Oh, no, Marion. Hi, Marian. I just heard your voice, so I didn't see you, but I see you now. I'm on mute.

Hello.

Building these purpose-built solutions, the tighter the scope, the more control you have, the easier it is to follow a lot of your other experience, which is software engineering, unit testing, functional testing. You can borrow from that. But the bigger the model gets, the harder it is to rein in. So I always encourage people to really, especially at the beginning when you're building your systems and processes, to kind of really scope carefully. And that way you can really get a handle on testing and deploying responsibly before you go bigger. And then that, you'll just have to make that decision when you come to it and weigh all of the tech debt plus, because that's the other thing is that these models, I mean, Alexa just is entering its 10th year and the 4 models I built are already, they're obsolete. So we, my team, we had 2,000 people. All that work is obsolete now because we now have a multi-headed, multilingual model. So that decision we made to build a bunch of little models, how much work was that that we did just replicating information?

Anyway. That is—

All right.

Close it out today.

It's time to go.

Yes. As we close out, I'm just gonna let Eric say something really quick. Good to see you again. Welcome back.

Yeah.

No.

We can—

we can all go. But the one thing I just wanted to call out is we've started an interesting thread on School, okay, where I'm asking people to just discuss their, um, different examples of use cases that they're using for multiple agents. Amazing. Well, agents, I'm gonna put some really great stuff in there too, so it's been—

I'm gonna pin that, uh, to the top so that people can see it. So thank you for pulling that because there's a lot of stuff in there, so I'm glad.

Yeah, there's a bunch of good stuff. I even— I learned some great things already, so it's good.

Amazing.

Oh my gosh, so good. That's a great way for us to end. So check the community. I'll pin the one on use cases. That's one of my— we have a few threads like that over the last few months. They're my favorite because it's like active, people are using it, what they're using it for. And maybe next week we'll do a prompt example. So what's your favorite, like, prompt of the week? What's your favorite prompt you're using? And that way we can learn together. And it can be any prompt in any context, you know, within crazy reason. I'm telling you, Jason, you freaked me out with your, like, go to this random website and look at a picture of me.

Yeah.

But I'm glad it all worked out.

I mean, I appreciate the skepticism because you— I mean, the next time we'll see.

Yeah, yeah.

Well, I'm gonna keep, I'm gonna keep testing that.

Pushing the envelope. I appreciate you all coming. Happy weekend. Uh, enjoy. It's the first week of November, so please take a little bit of time to set your intention for what you're gonna do with AI in the month of November. I'll be here every week. I'll see you all next Friday. Thank you for joining us.

Bye.

Thank you.

Thank you, Noelle.

Thank you, Sarah.

Thank you. Have a great weekend.