Tales of Abundance
Tales of Abundance artwork

All episodes

How Everyday People Are Using AI: Real Stories from Work, Home, and Travel

Episode 47 · 29 October 2025 · 49 min

Open on Spotify

Show notes

How Everyday People Are Using AI: Real Stories from Work, Home and Travel

How do you use AI in everyday life? Randy Lorensen, Dr. John Oberg and Anthony Marino compare notes on how each of them uses it at work and at home, from planning a family trip to Japan to building AI agents inside their companies.

AI is no longer just for tech experts — it’s changing how we travel, work, and live every day.

Hosts Randy Lorensen, Dr. John Oberg, and Anthony Marino dive into how each of them is using artificial intelligence in daily life and business — from planning family trips with Gemini to building AI-driven systems inside companies.

They explore the rise of agentic automation, AI prompting, and what it means to “agentify” your workflows. Along the way, they debate the future of AI in medicine, the role of wisdom versus intellect, and even how 3D printing could revolutionize manufacturing the same way AI is transforming business.

This episode is perfect for entrepreneurs, leaders, and curious minds who want real-world examples of how AI is being adopted today — and what’s coming next.

Everyday AI Uses, How to Write a Good AI Prompt, and Agentic AI

What you'll discover:

  • How John uses agentic automation to "agentify" processes at work, and plans family trips with a Gemini gem
  • Anthony's shift from skeptic to power user: data analysis and legal documents in minutes
  • Randy's everyday AI uses, from marketing copy to calculating materials for a fence
  • How to write a good AI prompt: let the AI interview you first
  • Intellect vs. wisdom, the domain-expertise problem, and why Randy's AI is ready to take calls
  • Power demand, nuclear, and how AI and 3D printing could change manufacturing

Timestamps

  • [00:00] Welcome: why AI is leading toward abundance
  • [00:47] How each of us uses AI today: John on everyday questions and company documentation
  • [01:58] Agentic automation: "agentifying" a process
  • [02:51] Building the perfect prompt with a prompting agent
  • [03:28] Planning family trips with AI, and age-appropriate travel journals for the kids
  • [05:29] Setting up safe AI use, and "the vice president of the future"
  • [06:23] Anthony: from shopping questions to data analysis in minutes
  • [09:03] Reviewing legal documents, and how fast answers come now
  • [11:42] From AI skeptic to user, and the 1960s mechanic analogy
  • [14:32] Randy: AI as a muse, for marketing, and in place of Google
  • [16:14] Building a fence with AI-calculated materials
  • [17:07] How we learned to use it, and different AIs for different problems
  • [19:07] Anthony: everyone will use it differently
  • [20:41] Learning from each other, and getting better at prompting
  • [23:18] Have the AI interview you, then write the perfect prompt
  • [24:05] The big problems AI has to crack: health, security and "math problems"
  • [26:00] More sophisticated phishing attacks
  • [26:53] The domain-expertise problem
  • [27:53] Randy's AI that's ready to take calls: holding AI to a higher standard
  • [29:35] Intellect vs. wisdom: will AI develop wisdom?
  • [31:22] Doomsters, and the case for humanoid robots
  • [32:18] What's next in our companies: personalized AIs with tribal knowledge
  • [34:00] Alexa Plus disappoints
  • [34:56] Break: Premium Cabinets business opportunity
  • [35:33] Choosing models, Sam Altman's quote and compute limits
  • [37:39] Power demand vs. power production, and nuclear
  • [39:27] Data centers, electric bills and not-in-my-backyard
  • [42:05] Resource constraints and an inflection point
  • [43:49] AI today is where the internet was 25 years ago
  • [44:35] 3D printing: a $47,000 Black Hawk part, printed for far less
  • [46:31] Manual workarounds in manufacturing, and better designs
  • [48:19] A caution for highly regulated industries, and wrap-up

🌿 SPONSORED BY AG1 (Athletic Greens)

Fuel your health and performance from the inside out.

Try AG1: https://shop.drinkag1.com/tales

About the hosts

This is a hosts-only episode with no guest: Randy Lorensen, Dr. John Oberg and Anthony Marino. (Host bios are on the host pages linked below.)

Find us online:

Hit Subscribe so you never miss an episode of Tales of Abundance. 🔔

Related episodes

Meet the hosts: Randy Lorensen · Dr. John Oberg · Anthony Marino

Transcript

A full transcript is below. It comes from the episode recording and has speaker labels and time markers, which line up with the published audio. It hasn't been proofread, so expect some errors in names and terms.

Randy Lorensen [00:11]: Listeners welcome back to tales of abundance why settle for less I am out here with dr. John and Anthony as I guess is the norm these days welcome gentlemen

Anthony Marino [00:23]: How you doing, Randy?

Dr. John Oberg [00:23]: Pretty exciting day today.

Randy Lorensen [00:24]: Yeah, we're going to, we're going to talk about AI and we've been talking about AI almost weekly for a while. And I definitely think where AI is going is leading toward abundance in the world. as, and you know, I think we've talked about some of the pain points that are going to be there and worries about people getting jobs, but ultimately it's going to democratize and demonetize a lot of the things that are happening in the world. So we want to start out. We're going to, we're going to talk about how each of us use AI today. And I think all three of us use it fairly differently. John, do you want to start out with how you use it today?

Dr. John Oberg [01:01]: I mean, we're using it all over. I mean, personally at home, we use it for planning trips and for meal planning and for any day-to-day question that you could possibly think of. I'm helping it with a lot of the content that I write inside my company internally for standing operating procedures, any kind of documentation, including how to help with performance reviews. So anything we can do to use AI, we use AI. We're trying to identify kind of all of the small processes inside of our company. And then we use it extensively throughout our company, companies that I work with. And by extensively, mean every single place we can find a way to make somebody's load lighter. We're aggressively doing that. So we've identified across two companies, know, 400 different opportunities to use AI. And we're now working through each of those, like how do you identify each one of those little processes to make people's light lift.

Randy Lorensen [01:54]: Explain agen- I'm not sure everybody knows what agentify means.

Dr. John Oberg [01:57]: Yeah, it's like, how do you take a process and automate it so that an AI agent is doing that process for you? And so you prompt it, it does its thing. And then sometimes you manually move it to the next process. Sometimes you automate, automatically move it from one agentic process to the next.

Randy Lorensen [02:15]: Yeah, think I've heard that word being bantered around a lot lately. And the first time I heard it, I had to go Google it and find out what it meant. And so I don't know that everybody is on that.

Dr. John Oberg [02:23]: Yeah. It's essentially, it's a series of prompts for an AI. you have these AI prompts, right? You have like, you can give it documentation and a whole knowledge base. Then you give it a set of instructions and then you can prompt it. And so, and we do all those things. We have a knowledge base for AI's. We have some that are kind of standard knowledge base. We have instruction sets and how we create instruction sets for the agents. And then we have some kind of template and prompts that we use. And then we have kind of people, people have the ability to like unstructured. prompt it. And now one of the things we started working on yesterday was like how to make a perfect prompt. So how to take someone who's not great at prompting and how to use a prompting agent to make for better prompts so the AI delivers on its promise more effectively. So I'd say that we're, you know, we're working very hard to push AI into the org chart of our organizations, but at home, we're trying to use it for travel planning too. Like anything we can do to learn faster and get information and grab that information and process it faster. doing that.

Randy Lorensen [03:19]: Why don't you share the... Do feel comfortable sharing the travel planning story specifically?

Dr. John Oberg [03:25]: Yeah, I mean, we, so we used AI to plan a trip to Japan for two weeks this last year. We used it to plan a trip to Antarctica for next year.

Randy Lorensen [03:33]: You're underselling. Explain to the listeners to the extent that you're using it.

Dr. John Oberg [03:40]: Well, this is my wife is really the one who's done the hard work on this particular agent, but she's gone in and created a really extensive set of prompts and a Gemini gem in this particular case that goes through and gives it all of the background for our family, all of our allergy information, all of our travel preferences, all of our ages, all the socioeconomic background and all of our preferences around flights and hotels and all these things. and then all of the experiences we've had that are really good and then takes us to a part of the world and says, here's the part of the world we want to explore. And then the agent asks us a whole bunch of questions like a good travel agent would, and it gives us massive amounts of information. And then for our trip next year, I've gone and now created an educational agent and said, hey, we're going on this trip to Antarctica. I want it to be fun, educational for our kids who are, know, teenager to early twenties. And I want them to have a great educational experience and an age appropriate fun journal that would engage them so they could maximize the enjoyment, not just some kind of rote scientific education about the trip that we're taking. And so I'm creating this thing where I can kind of drip out to them like these cool anticipatory notes next year to kind of build the anticipation for that trip.

Randy Lorensen [04:59]: That's cool. So, and when you say you're using it in your companies, you have a person who works on like you do AI and a lot of people are probably doing little parts or but more prompting, but you've got somebody that's actually coding and setting things up.

Dr. John Oberg [05:15]: Yeah, I think in companies, there's a couple of things I think that people need to think through. I think there's a platform for AI inside of a company that should be set up by the technology department of the company. And that particular role needs to be hypersensitive to regulatory issues, legal issues, and to set up the right safe environment for others to use AI. I believe that the vice president of the future is going to have to be someone who knows how to create these agents and these GPTs and custom gems and these workflows on top of that platform. And then I think everyone else in the company is going to need to know how to prompt things as a part of their role. For anybody that touches AI, prompt engineering is going to be a part of what they do. And so I think there are going to be some entry-level roles where all you do is grab a templated prompt and run that prompt. And then think as you become more senior in an organization and more experienced, you're going to have to be better and better. prompting and engineering those prompts up to the point where when you're an officer of the company, I believe that it'll be part of your role to create those custom gems in the world that we work in. And we're gonna put all of that already is showing up on our org charts for our companies.

Randy Lorensen [06:23]: Anthony, how do you use it?

Anthony Marino [06:26]: Well, it's really evolving for me. You know, it started out, know, where do I buy a new suitcase at the domain to today? I look at it for me as kind of my Mr. and Mrs. Know-it-all in terms of information gathering and guidance and so forth. Of course, I've been doing all the things. Yeah, I think so.

Dr. John Oberg [06:51]: That's gonna be real problem for me in our relationship, by the way. I'm gonna

Anthony Marino [06:54]: Yeah,

Dr. John Oberg [06:54]: be replaced very quickly.

Anthony Marino [06:56]: very quickly. Exactly. So but it's it's all it's it's changing the way I think. And it's taken down a number of barriers in terms of my skill set. And specifically, I can now I am now a spreadsheet expert where before I would I would grind over an Excel spreadsheet to get data in and constantly making mistakes and so forth. Now you you literally just throw data into the chat bot and then tell it what output you're looking for and voila. You know, I do some consulting.

Randy Lorensen [07:39]: Hold on, you say you're a spreadsheet expert now, that means you're saying you give it the data and you have it create the XLS or create the spreadsheet file for you.

Anthony Marino [07:47]: Not just create the spreadsheet, tell it what I'm looking for. So if I'm analyzing data on purchase behavior of a population, I can ask it, give me the frequency of a particular zip code, what is the average ticket in that zip code, how many times was a purchase made in that zip code, compare it year over year, month over month, quarter over quarter, things that. in the past would take hours to put together and analyze. Not only did you have the backup of the full spreadsheets, but you have the answers. And yes, you have to go back sometimes and know what I meant was this as opposed to what I'm seeing here and so forth. I mean, it creates tabs, it creates new columns. It's a walk in the park. then in terms of document management, you know.

Randy Lorensen [08:48]: Hold on, before you go to document management, I want to go back before it, because I'm going to forget. You said something earlier. You said it's changed how you think. Can we unpack that?

Anthony Marino [08:51]: Okay. Okay. Mm hmm. Yeah, I'm not I'm not fearful of going down a rabbit hole. I, you know, I've got I've got a very powerful tool, you know, for OK, for example, the way I think. I got a legal document I have to review the old days, you read it and you interpret it or you hire you send it over to a lawyer and he tells you the ups and downs of it and so on and so forth. Now you plug it into and you should always use more than one, you plug it into an LLM and ask it, this is what I'm worried about, what would happen if this happened, if that happened? And the answer comes, depending on how big the document is, in no time at all in comparison to the, I'll get back to you tomorrow, or I'm in court tomorrow, so I can't review this document, I'll get back to you on Thursday. So it's changing. I'm not afraid to get after it. And my expectations have changed tremendously in terms of how fast I'm getting answers. now, because of that, I can think at 100 miles an hour as opposed to 25 miles an hour in that specific instance.

Randy Lorensen [10:20]: Okay, or unless you're not done with that, I didn't mean to interrupt you. Well, I did mean to interrupt you. just wanted to dig into that.

Anthony Marino [10:28]: Yeah. And, you know, other documents and, you know, what does this mean? Or, you know, you're reading a book. I'm reading a really complicated book right now on, you know, economics. And there's a lot of stuff in there that, you know, economics classes were 20 something years ago in grad school. And so you just plug in a quick question. What do they mean by this? and you get a quick answer and now you can, boom, you can keep going where in the past you have to look it up and maybe interpret it so on and so forth. So it just changed. It's changed the way I operate, much more confident. my next, my next phase is I, and I think this is going to be a problem for some people, particularly non-technical people is, you know, how do I integrate video? How do I, you know, can this really create a Facebook ad for me? Things like that that I don't know how to do just yet and I've been working on trying to figure that out and it's not as easy as it seems.

Dr. John Oberg [11:39]: Randy, didn't we have a guy on a couple weeks ago, Randy, that looked a lot like Anthony, who said, I was just a bunch of crap and it wasn't going to actually uphold any promise? that there's a guy just like you who

Randy Lorensen [11:49]: No, he had that

Anthony Marino [11:49]: No, no, no, no, no, hold on.

Randy Lorensen [11:50]: in a discussion, but by the time we got to the podcast, he'd done research and he'd changed the tune.

Dr. John Oberg [11:56]: That's right, I remember that. That's the way it went down.

Anthony Marino [11:57]: Well, hold on a second. I what I said was to the audience is I don't I still don't see the economic boom from a headcount perspective from, you know, John, talked about, you know, having a VP of AI inside companies. As you were saying that, I'm thinking, do they really know what problems they're going to need that they can solve with AI. Or they just got to be throwing stuff against the wall and seeing what sticks. So that's where I was a few weeks ago. I still am to some extent. where it goes, we'll see. And I think it's going to be constrained by, I'll use the automobile mechanic as an example. If you give me a 68 Mustang, There's a pretty good chance I can change the points, plugs and condensers, you know, maybe do a valve job, you know, wrench on that, wrench on that car. You know, you pull up in a car, any car created today, there's not a mechanic from the sixties or early seventies that's going to be getting, be able to get into that car without reeducating themselves and learning about new equipment and diagnostics and so on and so forth. So.

Dr. John Oberg [13:20]: Sure.

Anthony Marino [13:24]: Right now I feel like that 60s mechanic on AI and I want to be the 2025 mechanic, but I think it's a pretty damn big leap and I got to sort that out.

Dr. John Oberg [13:38]: Well, that's 65 years of technology in all fairness,

Anthony Marino [13:42]: Yeah, it's true. It's true of all innovation, right? You have to learn it. But how difficult this is going to be is a question mark for me. And how many people have, the intellectual capability to do it, or even have the need to do it, right? know, answer this question. This is hypothetical. What percent of people only need the fundamental LLM to really make their life easier. They don't need it for any other reasons other than the things that we've already talked about, spreadsheets and recipes and travel and who is this person and so on and so forth.

Randy Lorensen [14:24]: All right, let's. John took you down a rabbit hole, let's stay on how we're.

Dr. John Oberg [14:28]: Well, let's. Well, Randy, I want to hear about you too, because you're usually an early adopter more so than me and Anthony. And so let's turn the question to you. What are the things you're doing with AI? Where do you feel like you're ahead? Is there any place you feel like you're behind?

Randy Lorensen [14:46]: Well, yeah, with how we use it, I mean, I started out using it writing a book and doing, I wasn't an author, right? And I wrote, I ended up ultimately writing most of the book, like 90 % in AI wrote 10, but I will also tell you there's a 0 % chance I could have wrote the book. or not zero, I would have never done it without AI to help Blake break the writer's block and a lot of other things and give me ideas. And sometimes it would just write things and I'd go in it then. And I wouldn't know what I wanted to say until it said it. And I knew I wanted to say something different. So just as a, guess, almost like a muse, it was really helpful. I use it for marketing materials, the ad copy video pictures. I use it, it's replaced Google for me for the most part. I say I Googled something earlier, but in all likelihood, I looked it up on Grok or ChatGPT or something. I I use four of the main paid models. I'm using it in my company. We're developing services in the company to do everything from drip marketing, reputation management, AI call answering, et cetera. We had to make, needed a dog food recipe and I wanted a dog food recipe and I didn't want to do the math to multiply it by so many days. I put it in and I told it, it created a shopping list. bought it, did it. I literally I built a fence in my, on my property recently. I used AI to calculate the materials that I needed. mean, there's just, there's almost endless uses for it at this point, right? I mean, it's, there's almost more things I'm Using it for the not using it for in some weird ways. I mean I'm leaving out in the interest of time I don't know dozens of others by the way the clips for people that listen to us on YouTube or Follow us on social media like those little short clips are all done with AI I mean I go through and I look at them, but they're those are all AI

Dr. John Oberg [16:50]: It's kind of like asking somebody where they use the internet today, right? Like, I mean, there's a little bit what we're asking about. Like the internet gets used for like everything from, you know, how I track my bio rhythms when I sleep to how my lights get turned on and off to like, it's literally permeated the vast majority of technology in my life. And I think that's where AI is headed as well.

Randy Lorensen [17:13]: Well, and how we use it, I guess I'll go first on this one, because I already said it a little bit. Like for me, or I'm sorry, not how we use it, how we learned it is what I meant. How we learned it. I had played with it a little bit, but I wanted to know more. And I joined a how to write a book in five days seminar and less honestly about writing a book and more about learning AI. And I spent five days trying to break it. Now I had a book at the end of five days. It was pure crap. I would never have published it or put my name on it. But then I took that and I just kept learning and learning and learning and really learned a lot about that. My big thing there was prompting and my big lessons were speak in this tone, speak to this goal, speak to this audience. And I used that not only there, but in my business and... Other things. that's a lot of it for me. And I've watched videos and read articles and, you know, went to seminars and things, but really just for me digging into it and using it for something and trying to break it is how I learned it. So John, how did you, how'd you learn how to use it?

Dr. John Oberg [18:21]: It was really similar. just started using it for small things and started using it for bigger things and bigger things. that, for me, then one of my big learnings was like, how do I jump into different models? Like when, you know, a open AI created a new model, I learned that, oh, I have to prompt this a little bit differently. And then I jumped to Gemini, it was different again. And now with Claude, different again. so, and somebody asked me today, like, which AI are you using? Are you using Claude or using Gemini? Are you using open AI? my answer was, yes, I am. I use all three of them and I'm figuring out which one works best for which thing. In the same way that I have a pair of pliers for one problem and a wrench for a different problem and it's like a knife for a different problem. It's not probably ideal, but I think different AIs are well used for different problems.

Randy Lorensen [19:07]: Anthony, how about you? How'd you learn how to use it?

Anthony Marino [19:10]: just, you know, it's like riding a bike, get on it and start to figure it out. mean, I even asked you a question early on, you know, how do you not, how do you, how do you tab down? when you hit enter, it doesn't start asking the question. I mean, it was so, it was ridiculous what I did know when you first got on, there was no instruction manual, you know, I guess I could have asked it for an instruction manual, but I'm not sure that whatever yielded. I think it's just one of those things you got to play with and find what it's useful to you for. This is a utility. It's like having some people like the thermostat at 68, some people like it at 74. And how's your spouse going to use it versus how you're going to use it in business? If you're a sports nut, you're going to go down one way with it and maybe use it for your fantasy football choices and things like that. I just think the world is your oyster here. you can see the abundance this brings is the abundance of information across infinite subjects. it's just a inverted funnel. It's amazing.

Dr. John Oberg [20:34]: And, for all of us, hasn't been, but it hasn't just been, we try to ourselves. I'm thinking this through, like, did we just use it ourselves? But the truth is we talk about it, three of us talk about it pretty regularly. And we're trying to learn from each other all the time. Like we're constantly asking each other questions about this, that, the other thing around AI. And then I think the two of you probably also have other people you talk to about it. know that I do. And so I think the reality is this is interesting, at least to the three of us. And we're all trying to lean in and ask people what they're doing with it and what they've learned. I have my wife at home, my wife uses it all the time. And so I have the benefit of chatting with her and we show each other prompts and some of my prompts are just such utter crap. It's like, what did I do here?

Randy Lorensen [21:15]: But the thing that's, you know, at one point when you would have a bad prompt, it was easier to go back and re-prompt. And I don't know if that's true. It's probably still true. But now, like being able to just say, hey, all right, I should have given you this information or add that to the prompt and do it again seems to be very effective now. And it seems like the feedback. It's getting easier and easier to prompt. That's maybe it's just me getting more comfortable, but it seems like it adjusts and can be corrected easier than in the past. Would you say that's true?

Anthony Marino [21:55]: I can't wait on that. I don't know. I mean...

Randy Lorensen [21:57]: I just feel like in the past when I gave it a really crap prompt, I had to go back and just scrap that, edit that, start over, and now I can say, I gave you a crap prompt, here's why.

Anthony Marino [22:08]: That could just be that you're very nature because you're not sometimes a crystal clear vocalist, if you will. John, I'm sure, because he's John, really takes his time and really thinks through what the prompt is going to be. I'm probably somewhere in the middle, so I'm not having that same problem.

Dr. John Oberg [22:09]: Yeah. yeah, believe it or not, kind of, one of the things when I had one of the guys that I work with look over my shoulder, said that I kind of intuit my way through it. So opposite of the way I normally react, I just kind of, I prompt and then I sharpen, then I sharpen, then I sharpen, then I sharpen as a way of me learning. And what I find is that I have a better starting point each time, but there are times where I get far enough down the thread where I'm like, man, I really just need to start over. And those are the times when I usually ask the AI to summarize my intent for me. And then I take that kind of new intent to a new thread and sometimes to an entirely new AI.

Randy Lorensen [23:07]: Yeah. I've asked the AI so much lately, like even prompts when I make a prompt, I'll ask the AI like what other things should or what other types of information should I have in this prompt or what questions do you have for me? it'll ask me questions. I'll give it and I'll say, great. Now take my prompt, that information and write me the perfect prompt for, you know, with this outcome or whatever. Right. So, okay. Well, let's talk about the future.

Dr. John Oberg [23:27]: Yeah. Yeah.

Randy Lorensen [23:35]: Where do you think it's going in the future? And we talked about this a couple of weeks ago, but where do you want to use it? Where do you think it's going? How do you see implementing it moving forward? Because we're all using it, but what's next?

Dr. John Oberg [23:50]: Anthony Westar first.

Anthony Marino [23:52]: Well, you know, for me personally, I think the discussion's irrelevant and maybe for us in total. In terms of what we're hearing and the, you know, the noise on the street is that in order for this to be a good business for these AI companies, they've got to crack the code on some big problems, you know. in health, how does it evolve in the medical profession? It's there now, it's its foot in the door, but what really is gonna happen down the road, probably a lot faster than we think in terms of helping us become a more healthy society, both in prevention and in treatment. Problem solving. you know, and I won't do this justice, the 13 % of the most difficult math questions that have never been answered have been answered by AI. What does that mean in terms of the impact on things like security and so forth? And all things are a math problem. And what does that bring to us? Does it get us to the point where people where the machinery is smarter than the human, where does it take us to the point that it won't need to be prompted, it'll just do things. So I just think it's accelerating like crazy. Everyone's trying to make money on it, which generally speaking is a good thing. And then there's the other side of this, that man can be a problem. Can AI be used in a way to mess with our electrical grid, our transportation system, our financial system? Who knows?

Dr. John Oberg [26:00]: Well, so it's already being used for advanced phishing,

Anthony Marino [26:03]: Mm-hmm.

Dr. John Oberg [26:06]: I mean, I've never seen more sophisticated attacks than I'm seeing today.

Anthony Marino [26:11]: Your password is a math problem.

Dr. John Oberg [26:14]: Yeah, that's a thing that they'd be concerned about.

Anthony Marino [26:17]: Right. your bank account is at risk. How much risk?

Randy Lorensen [26:22]: Your bank account has, your debit card has a four digit pin on it.

Anthony Marino [26:27]: Right. Now I have to figure it out. Oh, I know.

Randy Lorensen [26:28]: It's not that complex.

Dr. John Oberg [26:32]: So step one, cancel all my debit cards today. That'll be the first thing that I do. I mean, I don't want to dive into the blockchain problem to see if blockchain solves some of this problem. I do have a couple of other things I would add that I see this slightly from a different angle than Anthony. I don't know that I disagree with anything you've said, but I think that for me, there's a big domain expertise problem here. Like I see people using AI in their own domain exceptionally well.

Randy Lorensen [26:35]: you

Anthony Marino [26:49]: Go ahead.

Dr. John Oberg [26:59]: And it's very hard to use AI outside of your domain exceptionally well, because you don't know what you don't know and you don't know how it's misleading you. And I see this most commonly like in medicine. When I'm sitting down next to Dustin, you know, I know behavioral health really well. He's an expert in medicine and I can't catch it where it gets the medicine wrong and he can't catch it where it gets the behavioral health wrong. And so, but we can each catch that inside of our domain. And that's a real problem because how do you take someone like my son who doesn't have domain expertise in anything yet? And how do you teach him domain expertise when you're relying on people with domain expertise to use AI, not giving young people jobs in those areas? We're going to have a real gap there that's problematic. And I don't know that AI can solve itself there. And I don't know what that means for the career of the 20 somethings and the teenagers of today. So that's, that's number one. I'll pause there so Randy can ask questions if he wants to.

Randy Lorensen [27:53]: Well, just, you know, I think I told you the other day, John, on a call or maybe you weren't on there, but one of the products I'm working on in my company is an AI voice answering system, initially meant to be used after hours, but if it's good enough, we can use it all the time. I got to tell you, right now, I'm not sure the average person would know. I don't think most people would know it was AI, and I think even people that suspected it wouldn't know for sure. But... As I kept, you know, it would come out and I'm like, okay, we got to fix, you know, start out with these five things. And then it was four and then it was two. And I kept being like, here's something else I test. And I kept trying to break it and kept changing it. And at some point, it was, was last Wednesday, I realized, I'm being really nitpicky. Like I have a much higher standard for this AI than I do a human. And if a human was this good, I would be letting them take calls. This AI is better than most of the humans in my organization. It's ready to, know, not that it still can't be improved, but it's ready to go. So I'm less concerned. Now you're dealing with medical things in people's lives, right? I mean, here I'm talking about booking an appointment. It's not, it's not life and death. And, but I think that a lot of times we have too high of standards for the AI than... what we have for humans. have higher expectations. We expect it to be perfect where we have some grace for humans not being perfect is my point.

Dr. John Oberg [29:24]: It depends on the problem. But my concern stands, which is if you stop giving people domain expertise, then you no longer have operators with domain expertise. That's a problem. The other thing I would say is that, well, Anthony, you said everything's a math problem. think, I mean, Every math problem is a math problem, but I talk all the time with people about the fact that AI solves intellect problems. It doesn't solve wisdom problems. And so, and there's a big part of our brain that doesn't get solved by an intellectual processor. And so I think that's a good thing for humans. I think that that makes AI a better tool for us so we can lean deeper into that. But I think the future is knowing how to lean into work. the other work front, at least where we have wisdom work that needs to be done by humans and intellectual work that we done by AI. So I think that's a real, a real future conversation people need to think about.

Randy Lorensen [30:21]: It'll be interesting who's right. I agree with Anthony that literally everything is a math problem, including wisdom. We've had this debate before, and I don't know, maybe you'll be right, maybe I'll be right. I think AI will develop wisdom. And in fact, sometimes when I talk to it, it seems wiser than some people in certain situations. I think it's already developing it, in my opinion. But, you know. If we can solve math, what Anthony's talking about, means we solve. I don't know, the implication from people smarter than me is that we basically become gods at that point. That we can create anything we want.

Dr. John Oberg [30:58]: Well, at that point, you get people that can maybe emulate wisdom. But I still think there's a difference. so we'll find out when we get there.

Randy Lorensen [31:08]: I hope you're right. There's a point where if AI gets so good, What does it do to humanity? I mean, that's a real question. Does it replace us? Do we, I think we kind of merge with it or we use it and I'm not the doomist there, but there's smart people that are. I think also in the future, I think a lot of the stuff that's going, think human, I'm very bullish on humanoid robots. And I think that a lot of what it's doing, whether it's driving cars, robots, drones, et cetera, all of that, I just, heard the other day that the U S military or not the S military Palmer Lucky, what are, don't know the name of his company, but they've got a, they've got fighter jets now that they're creating that are drone fighter jets that fly with the main pilot. So the main pilot is there. And he or she is flanked by like five or six of these drones that basically do all the fighting and take all the danger and you know, and they're pretty much AI just so they're like the best wingman ever, right? So, I mean, I think there's a lot of that stuff, but I mean, how else in your companies? Like, what's the next step for you and your companies,

Dr. John Oberg [32:02]: Okay. Yeah. I'm worried a little bit about like, think AI is making it tremendously easy for you, for small companies to innovate and go fast and catch up and do it cost effectively. I am nervous though, big companies have data already in place and that, so I think speed gives small companies a real advantage, but data scale gives large companies a big advantage too. And I don't think many people are thinking about that or talking about that. Like the amount of data that I don't have in a system already as a small company is a thing. And so. So I think that we're to have to find ways to lean into the speed where big companies are smartly going to lean into the scale. And that's going to be, it's going to be a different playing field than we've seen in the past, I think.

Anthony Marino [32:58]: Thanks.

Randy Lorensen [33:06]: Yeah, I'm hoping one of the next things that come in the future is more personalized AIs that are actually have, they got less knowledge, but more tribal knowledge. Like for a company having the personal AI for that company is going to be a lot of ways much more powerful than being able to just plug into all of chat GPT, because a lot of what it can do isn't going to help the average company.

Anthony Marino [33:20]: Okay.

Dr. John Oberg [33:33]: It's a small language model. It's already a thing.

Randy Lorensen [33:36]: I know, but this is where there's going to be choices, right? Like there's already reports that, you know, chat GPT is their new version. It throttles you down. Like sometimes even though you're on five, it's using three because it's trying to use less bandwidth and they're trying to move stuff. You know, they don't want everything going to deep research.

Dr. John Oberg [33:58]: Okay, so I'm gonna tell you, did either one of you get the advanced Alexa Plus just recently? I turned it on with my wife. We turned on Alexa Plus and we started using it. And it was like the new advanced version. It was slower. It was dumber. It made 10 times the amount of mistakes. It was horrible. And we turned it off. was terrible. So,

Anthony Marino [34:04]: No.

Randy Lorensen [34:05]: No. Is it because it has to be trained?

Anthony Marino [34:23]: Yeah.

Dr. John Oberg [34:25]: No, my guess was they were trying to use older models and save some cost. yeah, so I mean, that was my guess. I have no idea what really happened, but we turned it off immediately. We're pretty, but we're both pretty tech forward. Like we're pretty willing to like train a model again. But it was, I really think it was like, Hey, let's see if they notice. Like that's what it felt like to me.

Randy Lorensen [34:29]: Yeah. Right. All right. Well, listeners, we're going to go to a break. We'll be back shortly to continue with this and talk a little bit about 3D printing as well.

Anthony Marino [34:46]: interesting.

Randy Lorensen [35:31]: All right, listeners, welcome back to Tales of Abundance. Why settle for less? Still here with John and Anthony. Haven't scared either one of them off. And we talked a little bit before the break about making decisions between even, know, chat GPT is making decisions about which model and how much bandwidth. And Sam Altman had a quote and it says if a so Sam Altman, the head of chat GPT. Had a quote, it says, if AI stays on the trajectory that we think it will, then amazing things will be possible. Maybe with 10 gigawatts of compute, AI can figure out how to cure cancer. Or with 10 gigawatts of compute, AI can figure out how to provide customized tutoring to every student on earth. If we are limited by compute, we'll have to choose which one to prioritize. No one wants to make that choice. So let's go build. And

Dr. John Oberg [36:29]: Keep hold on. Keep reading. I want you to read the next paragraph too.

Randy Lorensen [36:32]: That's the rest. don't know what you're reading. Continue then, John.

Dr. John Oberg [36:36]: The whole blog, I can paraphrase the rest, can't quote it on it up, but the next paragraph was, our plan is to release one gigawatt of compute every week from now until whenever. And so this was a sales pitch inside of his blog about how they're gonna be able to solve both problems, not just one of them, so that you wouldn't have to solve one problem or the other. They'd be able to solve cancer and education within 20 weeks. That was the pitch.

Randy Lorensen [37:02]: Well, I hope that's right, right? But my point is with this, if we had to make that choice, what a wonderful choice to have, right? Like, either one is a huge step forward from where we are now. And solving both, if Anthony's right about solving math, then we solve both and we solve all the other problems too, right?

Dr. John Oberg [37:24]: When you wouldn't have to choose one or the other, you'd only have to, the only choice you'd have to make is which one comes first. Like that's, that's, you know.

Randy Lorensen [37:31]: Yeah. That's with the assumption though, that the power, like we have, know, John, you and I talked about this a few weeks ago. think Anthony was on the beach in Sicily, you know, ignoring us, the, you know, power production, like power demand is way up. Power production is not, and we had that chart and it's, there's people working on it and there's smart people working on it, but there's a lot of, like, I think it is going to be a choice unless we somehow. figure out how to, what, 10x, 20x, 100x the power generation we have in this country?

Dr. John Oberg [38:06]: It's a temporal choice. It's a choice for a period of time, but like anything else with technology, you build one thing, then you build the next and you have as much throughput as you have for technology. But I think that throughput is consistently getting wider. You know, we're going to find more ways to create more energy. There, there, I yeah, I just, there is an energy problem. The energy problem will need to get fixed. And although we may not have direct line of sight to it, we haven't put up a nuclear reactor. Maybe we've put up one or two in the last 50 years. Someone's got to fact check me. The government is looking into geothermal power right now. That's a whole other energy source that's on a relative basis untapped as I understand it, which is not very much at all.

Randy Lorensen [38:39]: I think one. I think one.

Anthony Marino [38:58]: But the constraint could also be, you know, a NIMBY constraint, not in my backyard. Already Amazon, and I won't get the cities right, both Amazon and Facebook cancelled permit requests for data centers. And one was in Wisconsin for sure. Maybe the other one was in Arizona, because people are concerned about

Randy Lorensen [39:21]: One was in Arizona.

Anthony Marino [39:28]: noise pollution, their electric bills going up, know, not down. Because these guys are, you know, supply and demand. Adam Smith, right? 1700s, the more demand with limited supply, price goes up. And I heard some guys on a podcast talk about talking at length about this. And, you know, they feel like the hyperscalers should be picking up the tab. You know, they have the cash on the balance sheet. They should be picking up these tabs. And if they want to...

Dr. John Oberg [40:04]: Instead of any tax breaks, you mean?

Anthony Marino [40:07]: instead, well, not it, they should just pay whatever they have to pay to keep the cost of electricity where it is today so that they don't get shut down by local population. So there isn't a not in my backyard phenomenon, which is part of what happened with the nuclear power at that end. The, you know, the Greenpeace type thinkers that thought it was an evil, evil source of power. nonetheless, not, yes, there are going to be other ways, but it takes years to put in even these small nuclear reactors. It takes years to put them in. Now China, someone told me had 30 of them that they're working on getting done along with their coal mines, along with everything else. So they don't have the constraints we have.

Dr. John Oberg [40:57]: You know what I haven't heard? I have not heard of anyone talking about a nuclear powered, like aircraft carrier that's being decommissioned, being turned into an offshore data center. I have no idea that's a good idea or not. Has someone talked about that? Is it really? That's so good. Okay. All right. And what do they do? They just put all the chips on the ship and then.

Randy Lorensen [41:09]: I've heard people that is a plan by the way. No, that is a Yes. Yeah, for sure. I don't know, but I mean, just for power generation, I mean, they could just plug it into the grid, right? So, but I do think there's gonna be interesting choices that we don't have to make today. But again, it's two good choices. John, I think back to, we were at a company where we'd have training classes and sometimes more people would show up than you had room for and you were like, well, that's a good problem, right? I got, you know, it's...

Dr. John Oberg [41:27]: I mean, yeah. Well, I think there are choices that are real and present though for like the everyday AI user today. And it's about like which platform to use. Like, you know, what do you choose to use Gemini or OpenAI or Anthropic? And so I live in a world where we're going to use all three. Like we're going to take all three of them and we're going to find a platform for us at Salesforce and we're going to plug all those things into Salesforce. But I think people are going to be, you know, resource constrained from money. And they're going to need to choose where they want to learn to use. think choosing the right model is going to be a thing that people bump into. So I think they're, kind of the big choices you're talking about energy and chips and rare earth minerals and all these things that are kind of national level issues that have to be addressed in some of them in global. And then there are the everyday issues that, you know, individuals and, companies are going to be faced with small businesses, like not big businesses, but small businesses. to survive. I just, you you think about what Amazon did to the small business and it changed the landscape for sure. It shut some down. It opened up a whole new marketplace for their small businesses, but we're at an inflection point and that's for sure. And so people are going to be faced with decisions and not making a choice as a choice too.

Anthony Marino [43:06]: Well, the last constraint or another constraint is going to be what business problems will people try to solve? Is curing cancer going to be worth it from a business perspective? From the pharmaceutical company's perspective, the answer is no, right? Because they want to treat disease. They don't want to cure disease. What will people work on? People are going to work on things that make money because they are They are living and breathing every day about their stock price. And if they don't produce quote unquote shareholder value is gonna go down and they're not gonna want that even though that might be the right answer.

Randy Lorensen [43:47]: The other thing I think we need to remind ourselves is AI right now is very much where the internet was 25 years ago or more, where the internet was there, but there wasn't a lot being built on it yet. mean, have just, like companies are just starting to company, layered companies on top of the internet and offer services. Like this is a, this is a brand new industry that is, is a long ways from maturing. And I think that'll help. you know, this growth in the scale as well.

Dr. John Oberg [44:18]: But it's happening at a faster rate than it did 25, 30 years ago too.

Anthony Marino [44:22]: it's unbelievable.

Randy Lorensen [44:22]: God, you know, I mean five times faster. I mean, I I don't know the answer, but it's a lot faster a lot faster. So, all

Dr. John Oberg [44:27]: I don't know.

Randy Lorensen [44:33]: right. Let's another thing. So we've talked about AI a lot and one of my things I kind of go into is 3d printing and I recently saw this article And I heard somebody talking about it. anyway, this is this will be real quick. But this is another exponential technology, right? Like so this was an army, I believe a general. I'm not sure. I'm just going to go through this real quick. But on Blackhawks, they have an external fuel tank that breaks a lot. And they've cost they've been paying one hundred forty four thousand dollars per. In just 43 days, they 3D scanned it, did the prototype, did everything, and for $3,000, they replaced. And now they can just print these whenever they want for, you know, I think it's under three grand each time, but from 144,000 to three grand. And the one that I thought was even down here further in the article, So there is basically a knob, but part of an assembly, I believe on a Blackhawk, that was a $47,000 to replace, they 3D printed it for 15 bucks. So I think that's another thing that we start getting AI applied to robotics, 3D printing. mean, just, the future.

Dr. John Oberg [45:55]: love that idea, but do you want to get onto a helicopter with a $15? I guess if it's a even a doorknob, like I can't think of a single knob in a helicopter that used to cost 50 grand, that now costs $15 that I'd be totally psyched about. There's other parts, my standup desk, I wouldn't mind that or like, but I don't know, man, like

Anthony Marino [46:17]: you ever been in a bowling plant or an auto auto plant?

Dr. John Oberg [46:24]: I've been in an auto plant, not in a plane. Actually, I've been in a plant for planes, but it was not Boeing. It was the Eclipse plant out in New Mexico.

Anthony Marino [46:34]: Anyway, we can do a deep dive on this. You would be stunned if you saw the manual workarounds people use on something you're driving 80 miles an hour down the highway. OK. Or, you know, do you not recall a Boeing plane having a door fly off the off the jet? And why was that? Because there were a couple of bolts.

Dr. John Oberg [46:50]: Yeah.

Anthony Marino [47:04]: tightly secured on a relatively new plane. I shouldn't say new plane, but so.

Randy Lorensen [47:11]: $15 charge, John, is because, and we can do a whole episode on this, but in a manufacturing process where they have to make molds and they got to have an assembly line and there's just a lot that goes into it, they can use better quality materials and a lot of times better designs and 3D print them and do it cheaper. The price is not what's representing the quality.

Dr. John Oberg [47:32]: Yeah, I think it's important to hear like, know, while a pencil may have a dollar's worth of materials, might have three cents of materials. It has a dollar price tag because of all of the other things it has to go through. And so that $15 part was probably more like a $2,000 part if it had to go through all the same regulatory scrutiny. the point's still the same. The point is that 3D printing has the same potential impact. on manufacturing that AI has on everything else we've talked about.

Anthony Marino [48:03]: Well, they'll be interlaced. They'll be interlaced. AI will generate one way. It will go both ways. Can we do this? Can we do this some other way? And AI will get involved, help develop an answer, help develop the process. I think that's where it's a practical application for AI.

Dr. John Oberg [48:05]: That's right. I love it. just working in a highly regulated industry, my caution to my team is we're in a highly regulated industry for a reason and those regulations, they matter. so like, and you know, Elon Musk is great about like, regulations, you know, unless you understand the regulator and he tries to blow through a lot of that stuff. I think there's high consequence and low consequence regulation. work in a high consequence regulation environment. And so I'm probably a little more, a little more conservative here.

Randy Lorensen [48:50]: We'll come, we have to wrap this up. have a hard out here and we'll come back to this, but I will say the primary places 3D printing is being used mostly is aerospace and medicine. And, you know, it doesn't seem to be problems, we can, this can be a, this can be a future topic. Listeners, as always, download, subscribe, tell your friends, neighbors, et cetera, but also send us comments. Go on social media. If you follow us on Facebook, Twitter, give us feedback, tell us. Tell us what you like, tell us what you don't like, if there's topics you'd like, we'd love to talk about that. And today we'd love to know how you're using AI and where you think it's going in the future. So you can email us at talesofabundance at gmail.com or find us on wherever you get your social media. everybody have a fantastic day. John, Anthony,