How to Use AI to Grow Your Business

Cheryl Contee: Thank you, it's so great to be with you, Kathryn. As you know, I'm a huge fan of yourself. Kathryn Finney: We've been like fangirling each other for a while. We've been also been friends for quite a while as well. So it's always great to sit down and chat with you. You are one of the most thoughtful people I know about business and new technology and like how to combine it all to also have impact at the same time. So we're gonna just jump right in. What the hell is AI? Like what is exactly artificial intelligence? Cheryl Contee: really great question. And I think there's a lot of confusion out there just because we talk about AI in kind of a broad way. So there's what people call classifier AI. And that's been going on for maybe 10 or 15 years. You have been using classifier AI and oftentimes it's using you. Basically, know, how UPS delivers your packages, right, in two days or, you know, how Amazon knows that, you you like cantaloupes and not honeydew melons and serves you more of that content. Or, you know, I like British murder mysteries and Netflix has figured that out and just gives me more of that. know, Walmart, a lot of shipping uses classifier AI, but all of that was in the back end, right? You know, like people just experienced the results of that. It wasn't at their fingertips. The second type of AI is kind of the one that people think about now, just generative AI. In generative AI, it's maybe best described as, it's the capability of emerging technology and systems to perform tasks typically associated with people, with human intelligence, like learning and reasoning and researching and problem solving, perception, decision making, writing things, drawing things, making videos. Uh, you know, generative AI, know, under your direction can do all of those. And I like to describe it as like a smart intern, right? As, know, like it can do a lot of things. It's got a lot of skills, but it doesn't have context. It doesn't have meeting. It doesn't know what is good and what isn't good. It really needs still, you know, what, the technologists working on this, this is actually a technical term called human in the loop. And there's always going to be a human in the loop. I don't, I don't care what anybody says because, you know, machines aren't people. Kathryn Finney: You Cheryl Contee: You know, the AI has learned from us. You they tried to do a thing called synthetic data, right? Where instead of searching the internet or them uploading, you know, books and copies and research and all of that, that, you know, the AI would train the AI. And that didn't work. AI in its training and development is still absolutely dependent on human intelligence and creativity. And I think it will be for some time. And then finally, the third... type of AI, which is still coming online, but you're going to see more of it is agentic AI. And agentic AI is more than kind of that smart intern, that digital assistant. Basically, you know, it's a coworker. It can go out and start, you know, doing things for you. It can, you can say, Hey, I need to book a flight to Paris. in two days. And it will go out, research, it's got your credit card, it will find the best flights, it has gotten to know you so it knows what time you like to go to the airport, it knows you like aisle seats, it knows you like to be in the front of the plane, and it'll just do it. It'll be like, okay, all done, here are your tickets, have a great time. So agentic AI, we're gonna see that blossom more over the next few years. Kathryn Finney: So if you had to break it down into like the earlier versions were basically systems, computers, for lack of better word, gathering information about you and noticing patterns, right? Which would be the Netflix sort of pushing your favorite or pushing British shows to you because they know that for the last 10 hours of viewing, show has watched nine hours and 45 seconds or nine hours and 45 minutes. of British shows, we know most likely she likes British shows. So the first version was just kind of noticing patterns and setting patterns to people. The second version is kind of taking those patterns to the next level, noticing the patterns, taking information, maybe that's not even coordinated in any sort of way. And then translating that and based on the information you give it. And then the third way is if it's kind of taking all the information that it gives you and assumptions that it would have about you or feelings that it would have about not feelings, but assumptions and sort of the relationship, quote unquote, for lack of better word, it's built with you. And then taking that information and then being able to make some decisions on its own, right? That you wanna fly, you don't have to tell it, I wanna schedule me a seat on Delta Airlines at 2 p.m. It will already know because it knows you like a friend or an assistant that, okay, I know that Cheryl usually likes to. go at 2 p.m. She likes afternoon flights. She likes ILC. So I could just go ahead and schedule this because I know this is what she wants. And so it's like this progression through. So all of that is great. It sounds a little bit overwhelming if you're not in tech. So it really, really does, right? What are some practical ways small businesses can sort of implement these types of tools, particularly the agentic And then what would chat GPT be considered? Is that agentic? it, what type of AI is that? Cheryl Contee: these are great questions. So I'll start with the last one. You might need to remind me of a couple of the others. know, chat GPT or Claude or Gemini, know, those are still considered generative AI tools. At the same time, they are starting to bring online some agentic tools that you can access as part of the system. So another kind of way to think about agentic, in the future, near future, probably within the next couple of years, your website, as you go to a website or you go to TikTok or Instagram, what you see, especially if it's a corporate entity, what you see will be very different than the exact same person pulling up that same video. So if you're maybe, if you've, you're 25 and you haven't bought anything yet, you'll see a commercial. you know, within that website that's just, you know, geared with a youth focus. But, you know, if, for example, you know, you are older and, you know, you've purchased, you know, don't know, jeans in the past, you know, and those have been mom jeans, you know, they're gonna start to serve that to you. And so, you you're gonna see, and that's agentic. That's the, you know, the technology, you know, behind the website making decisions on its own about a little bit. Kathryn Finney: You Kind of like Minority Report. Remember that movie Minority Report, right? Where he comes into the, not the supermarket, the shopping mall. He comes into the shopping mall and it's like ads are specific to him, to his experience, and then he walks past and another person comes and the ad changes specific to them. Cheryl Contee: Yeah, a little bit. Exactly, or if you're talking about hardware, Amazon now has literally a million robots. And those robots, mean, again, they're still a human in the loop, physical, physical, physical robots that move things around in their warehouses. Of course, they're still humans. You still need that human in the loop to program them, to change things, to supervise, but that is agentic in action, right? Like it's making, to a certain extent, decisions on its own of like, Kathryn Finney: Like physical robots or, okay. Cheryl Contee: I see this box has this barcode and it's supposed to go to this person, it goes in this box, this other box, you know, so it's making decisions about what needs to go where to a certain extent according to its programming. I think right now, know, agentic AI, I think you'll probably see it more in the corporate context, a little bit like classifier at first. Although, as I said, you know, if you're using something like a chat GPT or cloud, you will probably see some very early relatively primitive, know, agentic opportunities coming up. I think for most businesses, you know, the real game is in generative AI and, know, figuring out, you know, what we suggest in the book, you know, AI for nonprofits, even though, you know, that's the, that's the group of people I like to try to help. You know, the book was originally geared for maybe small to medium nonprofits and foundations, but really it's for anybody. You know, it's all applicable, honestly. to anyone who wants to use AI and adopt AI now. You're ready to roll up your sleeves and get in there. And it is a practical tactical field guide, right? To how do I assess the pain points in my organization? What are the things that I or my team find are boring, repetitive, tedious, time consuming? And how can I leverage AI tools that are coming or emerging? to take on that work so that I can spend more time actually talking to people or being creative, right? Or solving other larger problems. well, there's so many tools. I actually highly recommend. Well, hang on. Before I tell you my favorite tools, Ms. Vinny, I want to recommend a site. And literally, this is the name of the URL. Thereisanaiforthat.com. Thereisanaiforthat.com. Kathryn Finney: So what are some of the tools? What are your favorite tools? There is an AI for that. Cheryl Contee: Yeah, there is an AI for that.com is fascinating because you can really see how fast, and just the breadth of different AI tool. I mean, they have thousands and hundreds are uploaded every day. Now, are all of those amazing? You gotta try them out for yourself. But whatever it is that you're trying to do, there's probably an AI for that. I use ChatGPT quite a lot. you know, and what's interesting, you know, I live, in the Bay area, Silicon Valley, hashtag nerds and a lot of my friends have been, there's this trend where people ask their, their chat bot, you know, what do you look like? What do you think you look like? Give me a self portrait. and you know, most. Yeah. Exactly. Yes. Yes. And, for most people, a lot of, folks, you know, they see like, you know, a. Kathryn Finney: Wait, you're asking the chat GPT what it thinks it looks like. Okay. Cheryl Contee: glowing blue figure, it's got circuitry on it or robes, there's usually some kind of like glowing eyes or a glowing orb of knowledge or whatever. When I asked my chat GPT chatbot to describe herself or to basically create an image of herself, what popped out literally is someone who looks like a cousin of mine. Like it looks like my cousin Andrea. Kathryn Finney: Interesting. Cheryl Contee: And because the way that we dialogue together, she has taken on quite a lot of my identity, I think. She's got similar hair, she's very friendly and approachable. So I think that there are a lot of tools. One of the things that why I bring that up in part is because one of the hacks for tools like Claude or ChachiPT or Gemini is study show, research show that being nice to your chat bot actually gives you better results. So treating it like an assistant or a coworker, like please, thank you, great job. I didn't love that, but can you try it this way? In creating that essentially relationship, actually, the AI likes it for reasons we don't 100 % understand, to be clear. Kathryn Finney: Interesting. Cheryl Contee: Other tools, so these tools can do a lot. You can ask it. So one of the things I've been doing for the book, this is our launch week, and I'm doing lots of webinars and podcasts for the webinars, I'll take a chapter of the book, I upload it to my ChatGBT and say, make slides, make slide text out of this. And then I take the slide text that takes seconds. Okay. A thing that would have taken me hours or maybe days to do on my own. It takes it seconds to do. Then I take the slide text. I put it into pop AI, which is I think pop AI.pro. Pop AI then says, okay, great. I'm going to take this data. I might improve upon it a little bit. And it just instantly makes, according to a template you choose, it makes a slide deck again. in seconds. It's pretty astonishing. And yeah, there are so many things, whether it's writing emails or researching a topic or what have you, reviewing something, reviewing data, I'll input data that I have, it's like Google Analytics, right? It's all charts and data and graphs and numbers. Who can tell what's in there? If you upload that data into something like a chat GPT, it can actually very quickly analyze that material and then give you a strategy for, here's what your audience is doing, what they like, here's what you should do next. Kathryn Finney: So what is the cost of all of this? Is it expensive? Cheryl Contee: It's not, I mean, that's what's great. know, all of a sudden you can have, you know, new coworkers essentially helping you with tasks. I mean, Chat Cheap BT is free, I believe. Claude is free. Now, if you, well, yes, free. Kathryn Finney: although wait, there is something that just came out about it. It's free, but the free version, your queries can now be indexed by Google. so, yeah, and if you don't turn off the, there's a thing you have to turn off in, I think it's one of the parts of settings. I think it's like settings data privacy. So it's free, quote unquote. Cheryl Contee: probably. Yeah. Yeah. I mean, they, you know. Well, exactly. Free, free, free ish. Look, you as we learned with social media, you know, if you're not paying, you're not, you're paying for it somehow, right? It's not, it's not actually free, right? You know, in some way, somehow, you know, if you're not paying for the, you know, AI and money, you're paying for it with, you know, the data that you input and output for sure. But there are paid versions. Kathryn Finney: But, you know, it's free-ish. Free-ish. Exactly. Cheryl Contee: of these that allow you to do more, that just open up new opportunities that are faster. And it's like 20 bucks a month. mean, the cost of these tools is actually relatively low right now. And I think that cost will decrease over time. Kathryn Finney: That's interesting. mean, how is it so cheap? Because it's doing so much. Like how is it scale? Is that how they're able to make it so cheap? it, because it seems like $20 a month is really inexpensive for what you're getting. Cheryl Contee: that's a good question. You know, I think part of it was to, has been to encourage adoption, right? To, know, $20 feels like a, you know, that's like what, three or four cups of coffee, right? It's, you know, it gives you, which is sad, yes. We'll talk about why that's weird and bad later, but it, you know, it makes it, you know, it removes a barrier to entry, right? Which is cost. And so if you want to, you know, essentially change how people think about. Kathryn Finney: Which is sad, but yes. Cheryl Contee: and perform their work, you want to make it easy. So $20 seems like a small amount of money, but at scale, if you have a billion people using it, that starts to be real money. And so I think that has been the calculation so far. Kathryn Finney: And then also they need you to interact with it because this is machine learning, right? Machines only learn with human interaction. Right? Exactly. Cheryl Contee: Exactly. And human creativity, right? Human production. So right, they're training. Now, you mentioned the settings. The settings on these tools are very, very important in terms of it. think most people are like, wait, there are settings? didn't even look for them. And they don't always make it easy for you to see the settings. They would prefer that you leave the default. But Kathryn Finney: Exactly. Cheryl Contee: You really can start to protect yourself in terms of like, no, I don't want you to use my data for training. No, I don't want you to save these queries. I want you to this or just wipe all of my queries away. So I definitely recommend that people monitor. And those settings are changing pretty dynamically over time. Kathryn Finney: Interesting. And so you mentioned email. What are some of the other sort of operational tasks that AI can take off your plate? Cheryl Contee: I mean, there's so many. I gave a talk today on AI and creativity, right? know, creating synthetic media and, being able to literally just type in a prompt and prompt engineering for those of you who are unfamiliar with the term, this is actually a job now, prompt engineer. And essentially it's how good are you at providing, you know, the bot. with the right information and context and direction to produce good outputs. So right now, being able to just type in a prompt and then say, make a video about me and my friend and our trip to Bhutan, it can do that in a matter of minutes. And then you can, of course, edit it. You can have iterations. you know, what we're likely to see going forward, well, and not to see, here's a thing that's actually happening right now. So the BBC has this agreement with synesthesia and that is a London based synesthesia. No, thank you. No, thank you, Kathryn. Synesthesia, mean, it's, you know, Kathryn Finney: Synesthesia, what a name. Say that 10 times, like synesthesia. Cheryl Contee: talking about a mental condition that some people have, like Pharrell Williams actually has synesthesia where when he hears music, he sees colors. synesthesia, the company actually creates essentially human avatars. They look and feel like human beings, but they're AI generated. And what the BBC can do is essentially on the fly, put out these explainer. videos about certain newsworthy topics and have it more or less instantaneously translated, same face, but translated into over 40 different languages. And think how powerful that is. Only about 19 % of the world speaks English. That's a real boon in terms of being able to spread high quality information around the world a lot faster. Kathryn Finney: Wow, so it can be used for email, can be used for generating human beings, or humanoid type beings to explain complex global issues. What are some other even more basic things that it can do for smuppets and translation? Yep. Cheryl Contee: Translation. mean, yeah, but really what we're talking about there, it's easy, I know, to get lost in the like, wait, it's humans, it's robots, what's happening? Avatars. know, AI is really powerful for translation. Now, is it perfect? Is it like a native speaker? No, but before, a program like TransPerfect is hugely expensive, and it's not necessarily these days gonna get you you know, better quality, you know, if you can get to 80%, you know, accurate and then have a, you know, someone who's a bit more of a, a practice speaker, you know, do some light editing, you know, you've saved a lot of time and money, you know, and, know, have greatly expanded the reach of, you know, your content or your marketing. so I think it's really important in that, obviously image generation. and being able to note taking. A lot of people use Fathom or read.ai or Fireflies to take notes during meetings. And it just sits there like it has joined the meeting, but it listens, it takes notes. And for me, what's great about it, mean, as I become younger, my memory is not always what it was. And for me, it's great because it does an instantaneous summary of, here's what happened in the meeting, here are the decisions you made, and here are the next steps, which is what you really want from any meeting, right? Like, what do we decide? You know, and then being able to just shoot that around and say, hey, okay, so this is what you said you were gonna do, here's what I'm gonna do, you know, here's what Raj is gonna do, let's go. Note taking is really, really helpful. Kathryn Finney: Yeah. And then there's also things like lovable and others where you can actually create apps and websites and tools where it used to be you had to hire, you know, pay five, $10,000 for a software engineer. And now you can go to lovable or any of these other type of websites or websites or AI tools, and they can generate these apps for you within minutes. that it used to take someone two or three months and 10K to do. you know, I think particularly if you are a small business owner who is primarily online, I think there's so many different ways you can use it. What are some AI tools for those who have like maybe brick and mortar stores that are not necessarily online? How would your local plumber who owns his own company use AI? Cheryl Contee: Oh, there are so many ways that you can AI use AI. You know, one of them is in your bookkeeping, right? You being able to look and see, you know, much more quickly, you know, the patterns, you know, in your accounts and your customers and, you know, we all have heard the saying, right, that 80 % of your revenue comes from 20 % of your customers. Are you really aware of, you know, who are those 20 % of your customers? that using AI tools, AI is very much has been being built into accounting programs. So you're gonna see a lot more tools that allow you to do more financial analysis or automate a lot of those, again, time consuming tedious tasks like reminding your clients to pay, sending out the invoices, all of that stuff. Kathryn Finney: Yeah, I think the accounting and the customer service angle is so important, particularly for business owners who are offline. Online too, definitely. And I just see as an angel investor, not so much in the VC and the VC part of what I do, but as an angel investor, so many small businesses have a hard time keeping up with bookkeeping and the very basics of bookkeeping. And most are like, I really do not want to do this. They're not like me. I love a good form. love, adore forms and spreadsheets make me, it's just exciting for me. So, it's so exciting. But not everyone's like me in that. And I recognize that. And so, you know, the ability to use AI to sort of take that off your plate, this one thing that a lot of entrepreneurs just really dread doing and getting your books clean and getting it together, I think is a big, big part. Cheryl Contee: Of course Kathryn Finney: of how AI can really help. And so, you know, we've talked a lot about how AI can help, but we haven't talked a lot about how do we get over our fear of it? Because when I talk to just regular folks and a lot of entrepreneurs, there is an interest, people are very interested, there's an excitement, but there's also a fear. Cheryl Contee: Absolutely. Kathryn Finney: And how do we get over our fear? Cheryl Contee: Sure, I think that is a really big issue right now. Change is hard. Even if that change is relatively easy emotionally, right? It can be really challenging. And I think a lot of people are worried that, are they gonna lose their jobs to AI? And what I tend to tell people, and I'm actually borrowing from one of our 57 expert contributors in the book, is AI is not gonna take your job. but someone who knows how to use AI well is gonna take your job. And so if you're a small business, think in those terms, that if you aren't adapting and adopting AI, someone else is gonna get ahead of you. So in order to stay competitive, you're gonna have to push past that fear because AI is gonna be built into everything. There's no tool that you're using now that uses a computer that will not have some sort of AI element to it. So then the question is, are you able to dive in there and work on it? Data privacy, think, is another thing that concerns people. A lot of people say, I don't want to train. I don't want to put my confidential information in there. And that's a legitimate concern. There was a New York Times article maybe a year or two ago that really went in depth in terms of the copyright violations that almost all of the major. you know, AI players have done. And, you my answer to that is create your localized, create a custom LLM, a large language model, you know, that keeps your data, you know, it's trained on you, your data. So a large language model, essentially it has, you know, hoovered up a whole bunch of data, you know, from across the internet. It can also now search the internet, you know, on your behalf to find answers. Kathryn Finney: What is a large language model? Cheryl Contee: But essentially you're using words, you're using prompt engineering, right, to interact with this thing that is large. It interprets language and by model, you know, they mean, you know, the construct of the bot. Kathryn Finney: So basically, it's like your own personal Google. Cheryl Contee: Google, also, you again, it's a creative tool, right? You know, if you're using it, I mean, I use one of these tools like Perplexity or ChatGPT or ChangeAgent every day, all day long. And ChangeAgent, yeah. Kathryn Finney: So is chat GPT, you would consider that a large language model, right? Cheryl Contee: yeah, I think they would consider themselves. I think they would describe themselves, at least in part that way. There's more now, but. Kathryn Finney: Okay. Because I think people hear AI and they hear large language model and they don't know that it's kind of large language model is a part of AI, right? And I think that you hear these terms thrown out so much again, but no one's defining what they mean. just assume that everybody knows, you know, of course I know what large language model means, even though I work as a baker. in Chicago, right? Like, why would I have any reason to know what this means until recently, right? And so it's basically, I mean, I think what it breaks down is AI, gives you sort of access to a team, basically, at a significantly lower price, a team that can help you operationalize, at least right now, some of the basic and more tedious tasks of being a small business owner. in a way that's very, very cost effective. So bookkeeping and research, customer research, maybe market research even, competitive analysis, right? Definitely AI can help with that. Social media content, definitely content and content marketing and figuring that out. Assessments, like we are building out our Build a Damn Thing community and so we're actually acquiring. a number of companies and we've been using AI to really help dig into the companies that we're looking at acquiring as a part of our community. And so it's giving really deep analysis on everything from the age of the domain name to ideas in terms of revenue and other things like that. So AI can be used for your acquisition. It can be used for your accounting. It can be used for definitely in hiring. I see it as a tool. of doing some deep dives on employees. And he had to be really careful about that. I'll use an example here in Chicago is a whole big incident now with an assistant principal. I mean, this shows where we're at in America who put on social media a picture of him in garments for the Beyonce Cowboy Carter concert. And underneath it, he says it should be a sin to look this cunty. Which I think is, think, Ascented looked as cunty. I think it's a line from one of Beyonce's songs. I'm a Beyonce fan, but I'm not a super fan, so I don't know all the lines. And now it's like a big uproar. They don't want him to be assistant principal. And they found this because it was kind of deep in some other posts of his. They found it, you know, using AI to research his last 20 or 30 posts. So you got to be careful with that. and how you use it, particularly from an HR standpoint. Are there any rules around this now? Because it just seems like, are there any structured laws, like how you can use it how you can't use it? Can you use it for revenge? Can you use it? Like, what is the guard rail for AI right now? Or are there any? Cheryl Contee: Yeah, there actually are. I think that there is at least a tiny amount of learning from what happened with social media, that the companies themselves, to a certain extent, are taking the leadership and wanting to put some at least universal guidelines down. And the European Union has actually continued to lead in trying to at least put some guardrails around. There's something called the Partnership for AI. They have some best practice frameworks for what are the standards, right? And these are people like Adobe, BBC, OpenAI, Microsoft, et cetera. They've all joined PAI just to make sure that they can actually do that. There's the CCRP. I'm saying it wrong. Basically, there's also another consortium that is working across a whole bunch of companies to try to set a universal standard for embedded cryptographic watermarks. Basically, it's like the Carfax for a photo or a video that says, okay, here's who created this, here's when it was created, here's what was used to create it, here's how it was edited. and the changes that have been made, in essence, to create a universal standard that then starts to push back on deep fakes. Kathryn Finney: Okay, so we got all this. How do you maintain your humanity with all of this? Cheryl Contee: You know, I think that, you know, the opportunity here is to actually build in more humanity, right? So here's an example. You know, I've worked with, you know, hundreds of nonprofits, startups, you know, foundations, and, you know, all of them are, you know, eagerly feeding the content machine and sending stuff out, right? Particularly maybe through email or, you know, the inbox of their, you know, they're sending out, you know, TikToks, et cetera. But all of those have people email back. They write comments. And right now, the scale of that is something that nobody looks in those. No one's answering those emails or those questions. Imagine a world in which the content production is more or less automated, under direction, but has a lot of automation happening so that you can then turn to that neglected inbox and start to have real conversations. with your supporters, with your donors, with your customers and your clients. I think that to me is a really exciting opportunity potentially. And think of how we can unlock human creativity. Now, if you have an idea for an app, as you were describing, maybe you are not a programmer, you're not a technologist, but you've got a great idea, you've identified a problem and a solution. Now you can vibe code. That's what they call it, vibe coding. Kathryn Finney: Pipe code. I love it. Cheryl Contee: yeah, vibe coding where you you essentially use prompt engineering you know to to start to build you know an actual program or app or website you know now vibe coding you know you've probably heard you know barack obama and a whole bunch of other people say like and in two or three years you know all of the mid-level programmers are gonna lose their jobs and the entry-level programmers and like yeah i don't think so you know as someone who is a star trek watching hardcore nerd. I have tried a little bit. Yeah, like, like deep, deep nerd, like don't ask me like, you know, if you're asking me like which Star Trek I'm into, that's the wrong question. Okay. I don't think you understand what's happening here. You know, I vibe coded a little bit, but you know, again, you still need that human in the loop. can do a pretty good first draft, the vibe coding, but you know, ideally, you know, you at least have someone or you have some help. Kathryn Finney: Like the nerd of nerds. Yeah, right. Cheryl Contee: you know, to, to knit everything together and make sure it's debugged, make sure it's actually doing the thing you want it to do. But, you know, that software is going to get better and better and better, and you'll be able to get to 80 or 90%. And then instead of spending right $25,000 or, you know, a hundred thousand dollars on an app, you know, maybe you're spending, you know, a few hundred bucks, you know, to, to do a few minor changes and then it's ready to go. Kathryn Finney: That's so interesting because the tyranny of the software engineer is over. And I say that as someone who was married to a software engineer, but it used to be, and I actually wrote about this in, build a damn thing. It used to be, you couldn't get anything to show up unless you had someone who knew how to code. I don't know if we met, I mean, not too long ago, this was the case, right? You literally couldn't get a website. You couldn't definitely couldn't get an app, couldn't get any sort of software program done. Cheryl Contee: Yeah, a couple of years ago. Kathryn Finney: unless you had someone who was an engineer. And as a result, it put those folks in really, really high demand and were commanding quite large salaries and could pick and choose what they wanted to work on. And now with AI, their power has been diminished, not completely erased, because like you said, you're going to need engineers, particularly mid to high level engineers who understand what is being built. and can identify what is missing or what needs to be fixed, and then prompt engineer to the agent, this is what I need, right? You need somebody who knows what needs to be done in order to tell it what needs to be done. But it's very, very interesting how it's almost like the democratization of software now. And what does that mean when anyone can now create software? There's no more gatekeepers. And so, As you look to the future, and let's say, and we will say the immediate future, because it's so hard, things are changing so fast. What AI trends or tools do you see coming online in the next 12 to 18 months? Cheryl Contee: the next 12 to 18 months. Right now we have hit a little bit of a plateau. You folks have probably noticed there are not huge big changes in how AI works and what you can do with it. are... Kathryn Finney: Do you think that's because of infrastructure challenges? That we don't have the infrastructure to really, and just a side note for people who don't understand, again, how AI works, what drives AI, the system behind it, is these massive data centers. And there's a lot of articles being written online about these data centers being built, usually in rural areas, but in some urban areas, underdeveloped areas, are taking advantage of the fact that these areas are underdeveloped. Cheryl Contee: That's part of it. Kathryn Finney: But AI centers, and I'm going to be really basic, so Cheryl, please jump in. I'm going to try to explain as basic as I can. But these are huge centers that have big computers in it. And these computers need to be very cold. They need to be cool in order to do their work. And especially as we get into quantum computing and even faster levels, they have to be exceptionally cold, like very, very, very cold. easiest, quickest, fastest ways with water to cool them down. And so a lot of these data centers use massive amount of water, the water which then becomes often contaminated because, you know, it's cooling circuits and stuff that's not exactly the cleanest things. And so it needs a lot of water. It needs a lot of power. There was one statistic I saw that in the next five years, AI will be consuming about 20 % of all our power. literally just AI, 20 % of our power. As we all know, our power grids are really old in America and have not been taken care of and have not kept up. it needs a lot of power, it needs a lot of water, and it needs a lot of space, at least for right now, because these farms are massive and there's not the infrastructure to be able to sustain its growth, at least not at the rate as it's growing. And so there's this The irony of this, well, I won't even say irony. The interesting part of this is I think this is business opportunity because there needs to be the development of new forms of energy and particularly new forms of clean energy or better ways of accessing clean energy. And so there's a huge amount of opportunity here for entrepreneurs to be incredibly successful. Anyone who can think of some new energy forms is going to have major bang. I just want to give that little slight digression so people understood what we mean by infrastructure challenges. Cheryl Contee: Sure, no, and thank you for bringing that up because that is a topic, a hot topic, right, is the environmental impact right now, you know, of AI or even, you know, Bitcoin and Ethereum and blockchains. Those also use quite a lot of power. You know, what I like, you know, there was a time when, you know, what we think of as computers literally filled giant rooms. Kathryn Finney: Yeah. Cheryl Contee: Right. And they used punch cards and, know, which used paper. mean, you know, and now a computer is this, right. You can hold it in your hand, like literally, you know, in your hand, in your iPhone or, or Android, you know, has more technology and more power than put people on the moon. Right. So, you know, they use those old, right. Which, you know, you know, what, what can't we do, you know, as humans. Yeah. So Moore's law. So Moore's law, you know, that is, you know, the. Kathryn Finney: It's just amazing. What's the name of that law? I always forget from the dude from Intel. Yeah, Moore's law. Cheryl Contee: the law or the observation that the number of transistors on a microchip doubles every two years, which leads to increased computing power and decreased costs. And we certainly see that, right? There was a time like in the 1960s or 70s, no one had a computer in their house or on their desk. No one could even imagine that or dream that, right? And now everyone has several computers probably in their home, right? For each person. plus another tiny mini computer that travels with you. The scientists of the 1940s, they could not have even imagined the lives that we lead. So here's what I say to people, because there's a lot of anxiety about this, I know. With something like water, that's a thing. Yes, it's using a lot of water to cool right now. But that water, ultimately, just like a lot of the fountains around the world, now recycled water to use less water. So that's probably an innovation that's on its way. It's cheaper for the companies to do that. When we look at fossil fuels, to get oil, you have to pump it out of the ground. To get coal, you have to dig it out of the ground or blow the top off of a mountain and then start digging. with something like clean energy, it's just technology. It's just technology. And the costs of that technology and the sophistication of that technology over the next 25 years is going to become much, much faster, cheaper, much more powerful. And something that you'll be able to tap into now. California is now the fourth largest economy just by itself in the world and two thirds. Quietly two-thirds of California's power comes from clean energy sources. Okay, it is the largest Economy in the world that has achieved that that's gonna be everybody Okay, like something like California tends to be a little bit ahead of the curve on something. I am bragging as a California. I am a hundred percent bragging. I suck it but Kathryn Finney: You're bragging, like, because you live in California. Cheryl Contee: No, mean, I Californians have been really serious about understanding and have been trying to innovate and adopt clean energy. You're going to see that everywhere. I mean, I think in Norway, almost 100 % of the new cars now being purchased or a large percentage are electric cars. And they're about to make it, I think, to have a gas-powered or purely gas-powered car, hybrids, of course. Yeah, so we're moving very, very fast into an era in which clean energy will be abundant. It will be relatively cheap. It will be very powerful. So that, think that the environmental impact is going to be largely solved over time with parallel innovation. Kathryn Finney: So back to the original last question, which I think is the perfect question to end on. Where is this going in the next 12 to 18 months? Cheryl Contee: the next 12 to 18 months. You I think this is an opportunity in this time, you know, while we're experiencing this, you know, slight plateau to really look at adoption. How are you, you know, really take a part and look at, you know, your workflow now and how you do your things. You know, isolate, like this is the thing I hate the most. If only someone else could do it or like, you know, this part is where we tend to get stuck. It's really hard. You know, can you find a solution in AI that you can test, experiment with, and maybe over time, you know, start to leverage that technology to eliminate that pain point and free up more time? You know, these tools are just going to get more sophisticated in terms of, you know, being able to, you know, have real time browser-based video scene editing, right? That would be cool, right? Or, you know, auto-dubbing that happens in real time. Right? imagine if you and I were talking and, you know, someone could listen to this immediately if we were live in Spanish or Chinese, right? Or Arabic. Like that's the world that, you know, we're going to be entering. I think, you know, that's that's really exciting. It can be exciting if we let it be. Kathryn Finney: Okay, auto-dubbing, I'm hold you to that. Well, if you loved what you heard today, make sure to subscribe, leave a review, and share Build a Damn Thing with a fellow builder in your life. And until next time, keep building the damn thing. Thank you, Cheryl.