Todd Gagne

Today's conversation is with Andy Jorgensen. Andy has a background in ecommerce and recently has been working kind of in the AI space. He has, a a conference that he does called AI Edge, which is for companies to get a better sense of AI and just how to leverage it in their existing business. Aidan talks a lot about just kind of of some of the use cases that are kinda unique that he does, and really about the future direction of AI, and has a ton of energy in of this discussion on it. So I've really enjoyed this conversation with Andy, and I hope you do too.

Andy, welcome to The Treehouse. I certainly appreciate you, filling in for me today, and I'm excited about this conversation.

Andy Jorgensen

Sounds good, Tom. Thanks for having me.

Todd Gagne

Okay. Well, why don't we start at the beginning for you Like, I I you and I haven't spent a ton of time, but maybe you could tell me a little bit about your journey and, and kind of what you're we're gonna talk a little bit about a AI today. And so, maybe talk to me a little bit about your background and and kinda what you're doing in the AI world.

Andy Jorgensen

Yeah. So my background is in ecommerce. I have a fireworks company that the goal was we're gonna bring ecommerce and fireworks together. So bringing the latest and greatest technology with industry is just gets me all jacked up. I love it.

So, claim to fame, we were the first ever United States company at least to do the, like, pickup and delivery of fireworks. So think Walmart Groceries. We're doing it with explosives. And, we got all the legal yeah. Our whole our whole dream was, like, we wanna streamline.

We wanna Walmart ties, if that's even a word. We wanna make that firework industry just as as efficient and stuff like that. So it's remained relatively small. It's a seasonal market. But now, my background is converting from, like, the ecommerce in fireworks, and then I went to ecommerce consulting for, you know, Shopify stores to what everyone is talking about right now, which is just artificial intelligence and automation.

I would say those are 2 in the same coin. That's when people think AI. They're actually thinking AI or automation. How do I automate some something either my brain, my hands, or whatever it else is happening in my business And that is fresh.

So I am, along with every other pioneer in the space, I am just beginning to understand what things are and how it applies to this, that, or the other thing, and it's been fun.

Todd Gagne

Okay. So maybe just to maybe put a finer point. So you've been kind of, doing these kind of events in different cities, and basically getting, people interested businesses interested in AI and starting to educate them and doing some consulting around that. Is that correct, or is it a little bit different

Andy Jorgensen

Yeah. So, essentially, we've got one event so far. We put it together in about 60 days last year. We had over 200 people show up. So if you want to throw an event in the name of AI, it's gonna happen, and it's gonna be fast and hard.

What we found is as I've recruited people to come to Sioux Falls, South Dakota, which is where we're at, they've had a little bit of a roadblock because it's hard to fly into here, and it's a, you know, half day event that we had. So we are thinking we're thinking big. We're probably gonna go national, to AI it's called AI Edge, and it would be AI Edge San Diego or something. And this would be the place that you would come once a year to meet and greet with all the other people who are either beginners or experts, and where our our vibe is very unique, I think. We're not PhDs.

We're not, like data experts. We're not large language model developers. We're not even coders. I mean, you could be a coder, but we are just everyday people who want to take advantage of this. Everyday business owners, everyday employees, everyday executives, that are maybe not super developer savvy.

So that's the vibe the vibe and the vision

Todd Gagne

of what we got. That's pretty cool. And I and I think, you just think back. I'm probably a little bit older than you are, but I think back of, like, all the different kind of technology shifts that we've seen. Right

And so, you know, I I started off on, like, an Apple 2 c. You know, so it was a monochrome, you know, Apple computer. And then basically, you think about, like, the Internet and all the changes that came with that and then all the mobile components of it and then AI. And it just it just amazes me over a 50 year period, that's a crazy amount of change. And this one seems like there's a lot of opportunity here.

And, you know, the question people are asking is, is that going to be more transformational than the Internet I'm curious if you have an opinion on the that progression and then the impact that AI could have over time.

Andy Jorgensen

Yeah. I do have an opinion on that, actually. So if we think about the Internet as the connectivity of all the information, all of our information we publish on the Internet, AI would be the brain of the Internet. So that's the 3rd grade version of what I like to think through is AI is like the so so, for example, the Internet is genius because it's all of our brains, and it's all of us conversing, all of our opinions, all of our emotions, and, you know, on Facebook posts and stuff, and it's just all of so now if you take all of that and conglomerate it into a brain that learns everybody's as much data as it soaks up, it's learning. So just think, futuristically, the brain of the Internet, the brain of information that starts to learn from data that it collects.

Whereas, the Internet is like human input, and AI is human input, but it actually starts to grow, which is cool. So in that, I think it will be as transformational, if not way more, but for sure as transformational, I'd say.

Todd Gagne

Yeah. And maybe to build on that, I think that, like, data in in the Internet is static. And, basically, what you're doing is putting some rational, logic and almost thought. Like, I mean, you're putting some processing to solving problems with all of that static data across the Internet. And so, to me, that's pretty interesting, and we'll see where it goes.

But, the breakneck pace that we've been going through over the last 2 years has been kind of interesting.

Andy Jorgensen

Yeah. I agree.

Todd Gagne

So let's maybe start with, some, different scenarios. Like, you know, I think when I talk to entrepreneurs and stuff, I think you got a wide range of spectrum. You got people that are doing some really interesting use cases that I wouldn't have ever thought about. You've got a lot of people that are doing these use cases that you have heard about, and then you've got people that are basically trying to use it like Google. And I I'm kinda curious on maybe your top 3 or 4, you know, best case scenarios or, like, use cases that you use chat or one of the other bots, for that maybe are unique.

Andy Jorgensen

Yeah. So, I mean, it it is a version of Google, but there's kind of a translation issue sometimes with, like, okay. Where did that data come from And so is it valid Is it, hallucinating is when an AI makes up something.

So, for example, if you use it as Google and it's false information, it doesn't say, I don't know. It gives you an answer, and it it will produce so it's called hallucinating. It basically just starts feeding you an answer because it's supposed to generate text, and you're doing that. So the limitations for a search engine are kinda like, well, why don't you can use Google to use your brain to decide which tools. I think that that usage will get primed, obviously.

It'll it'll become more authoritative just like search engines became authoritative and you could actually trust them. I I genuinely think that the best application right now is is that what you're talking about Like, currently today, what should we be using this for

Todd Gagne

Or just, like, your personal, use cases that are kinda unique. Right Like, I I'm just curious about how you use it. Right And do you have a use case that you think is unique that is interesting that maybe most people don't do

Andy Jorgensen

I do. I'm just gonna read them off here. So number 1, being my real estate investment expert. Okay As an entrepreneur, let's say I want to put some money towards real estate development here in town.

I get a prospectus with a lot of numbers. I'm not an expert. I can YouTube it and Google it for hours and hours and hours. I can send it to my real estate expert entrepreneurial friend, which I did. And he put, you know, an hour and a half in, and he sends me and we can sit down for lunch.

We can do this. I took the exact same PDF for real estate development, and I said to Chad GPT, I said, hey. I need you to be and you got a prompt that's smart. You gotta be smart. You'd be like, you are an expert.

You are from Harvard. You are doing this sort of thing. And I fed it the PDF, and I said, you should tell me why this is advantageous. You should criticize this and tell me what's a little bit fishy. You should and start to, like, get the brain.

Because if you think about even just chat GPT, let's talk about AI as chat GPT. It is way, way, way smarter than you or I. Way smarter. Yep. It is like a teenager who's not self aware yet, but is like Rain Man smart.

So if you think of it in those terms, if you actually just probe the right parts of its brain, you can unlock anything. So that's one usage. It's like real estate investment. I'm not gonna just go off of that yet, obviously. I'm still gonna talk to my real estate friend.

I'm still gonna do my due diligence. But what is it that it might discover with all the data that it has, like, an insane amount of data that my friend or me might not seen from this Number 2, spreadsheets. Hold on. Right

Oh, yeah. Go ahead.

Todd Gagne

Like, let's go backwards. Let's go backwards. Like, what did that tell you Like, so when you fed that in, how did that change or shape the conversation you had your with your real estate friend

Andy Jorgensen

Well, we haven't actually I haven't, rallied around with him quite yet. So I haven't I haven't to, like, figure I haven't I haven't actually, like, figured out with him, like, where we wanna go with this, but that was just, like, one thing. So, for example, legal and regulatory, are there any pending zoning or regulatory approvals that are required that are still not met Or are there any disputes right now for zoning or anything like that Because that can actually delay it.

Then the follow-up question for that was, if there are any disputes or legal things happening, how are those resolved so that the project isn't delayed Because you know, with real estate, there's always delays. But it thinks about, if there is this scenario, how is that resolved And have you thought about this Or, like, what are the critical milestones in place, and what are the, like, contingencies in play for those milestones

And I'm like, jeez, I would have just been like, what return do I get So Exactly. Does that answer your question

Todd Gagne

It does. That's good. Continue. Number 2. Yeah.

Number 2. So, let's just

Andy Jorgensen

say a very basic business uses a spreadsheet. Mine happens to be a 10 years over 10 years, my brain has built the spreadsheet to be exactly what I want it to do, And it takes maybe 4 hours for someone else to be able to use it. That's any typical good spreadsheet, right, that's being used. People run their businesses on spreadsheets every single day. And I have a virtual assistant that I can have run and add codes and stuff like that or maybe add this script to it.

Well, I just said, hey, chatgpt. This column needs to talk to this column. And they're like, they're being chatgpt. He said, attach the Google spreadsheet link, and it will produce the code for you. So instead of 1 year ago, I went to upwork.com.

I paid a guy $150 to help me with my spreadsheet. Now I just say, I would like column c to talk to tab 2, column d, and I want it to do this. And it'll give it its best shot. Now it's gonna mess up. But isn't isn't, you know, Vadim from Pakistan gonna mess up too

Right Or I hired a guy from Texas, and I paid him a $125 for an hour of his time, and I didn't really get much from it. I can do this. In 3 seconds, it produces the output, and then I say, no, that's wrong. It'll go, no problem.

Here, how about this And within 5 minutes, it can do the most advanced spreadsheet work that you've ever seen.

Todd Gagne

That's awesome. Do you have a 3

Andy Jorgensen

3. So, fun fact, the the the conference, this is the creative thing. Everybody's probably using Chassis PD for this. I started AI Edge and I said, hey, I'm a white canvas. I'm gonna start an AI conference.

What should I call it And it said, boo boo boo, here are the choices. And AI Edge was actually one of them. I was like, cool. So I chose that.

I said, okay. Now help me now help me think through structure. Now help me think through this. And it was some of it is like a conference is a conference. It's not like an it's not like a an innovative thing for a conference.

It's kind of the age old thing. But there's just idea generation. Right So I said, I would like to send a notification out to my user or my attendees or prospective attendees, and I don't want it to be cheesy. Give me some ideas.

Give me language, and large language models are good at that. So that was basically just running the running conference was the number 3.

Todd Gagne

Yep. Okay. Those are good. Mine are probably a little bit more benign. Although, like, this the first one's not mine.

The guy there's a guy that I, did do some work with, and he was starting to feed, a bunch of pictures into chat gbt and selling, hey, which one of these people are related and why And that was pretty interesting. I was like, holy cow. He's basically using it for facial recognition. And it was doing some, like, inform it was doing some measurement of, you know, the nose to the eyes and the ears.

And basically, it picked out all the people that were related. And I was just, like, kinda floored. I mean, you know, that's kinda CIA, NSA type stuff. And so I, you know, I didn't play around with it much, but I thought that was kinda interesting. The second one that I did was, we we have a camper, and we have 2 kids that go to college at the University of Utah.

And so we were talking about, dropping them off here, and basically doing a trip. And I would wanna do it in Southern Utah, and I had, like I don't know. Probably I probably had 17 or 18 locations that were kinda just in a random folder. And I said, I need a 30 day trip, in Southern Utah, and I want this in order. Here are the things that I like to do.

I like to bike. I like to hike. I like to kayak. And tell me how long I'm gonna travel and how long I should stay in each one of these spots, and it mapped the entire thing out for 30 days with the location starting in Salt Lake and ending in Salt Lake. And I was just like, okay.

Well, that would have taken me a lot longer to do, if I had to piece all that together and figure out where they are and then what the sequence events and then and it and I down to the point where what mountain biking trail should I do and what hiking trails are popular. Right And saying and so it was all in the prompt, but that's still pretty cool if you could actually jam that out. And I would say it was probably 90% really, really good. It was really positive.

Andy Jorgensen

That's great.

Todd Gagne

The last one I'll talk about that I think we use actually internally with Wildfire is, is basically with founders. So we get applications, and we just have a script that basically says, how big is this you how big is this market What would be a good beachhead What would be the average deal size What's the pain tree really look like

And what's the go to market strategy And it creates basically a PDF that we can then compare to what the entrepreneur gave us, and it's pretty good. Like, I mean, it's amazing, like, all the stuff, like, especially in the pain tree. Like, you're looking for you give a an explanation or the the pitch deck, and then you basically ask it to do all those prompts. And it does a pretty good job of of giving you a high level understanding of how big is this market, how hard is it going to be to penetrate, who are the demographics you're going after, what's the average deal size gonna be like, and what's the cost like, what's the lifetime value.

Right It's got some pretty good estimates on that. And so, again, not something that probably a ton of people do, but it's been pretty good tool for us.

Andy Jorgensen

Yeah. Yeah. I mean, you could take that as far as, like, all your legal contracts. So, there's NDAs and stuff, and I had to sign some NDAs to work with some clients. And I'll I'll spit it to ChatGPT.

I'll say, hey. Is this a fishy NDA What what am I not getting into Because you can pay a lawyer. What is it

Like, $50 an hour. Right And Yeah. They'll nickel and dime me for that for that. I'll just take the client NDA.

Easy. I'll feed it to chat ubt. I'll say, hey. Is this legitimate And, of course, you have to train that convo to be a legal expert.

You don't wanna just, like, leave it wide open, but it's like, hey. You're a Harvard lawyer, and I want you to make sure you're in my best interest. You're my lawyer. Here you go. So it's just it it it it's you could just figure out any brained thing is really great.

Todd Gagne

Do you think though that there's a shift from, you know, how we've used Google to being a little bit more creative about how to use these AI bots I mean, I feel like that's part of what you have to go do. Right You can't have the rigid I'm just gonna go ask Google a question, and I'm getting a bunch of blue links back with a bunch of sponsored stuff. I think you have to, like, start thinking a little bit more creative about how to use these tools.

Andy Jorgensen

Yeah. So you're wondering is, like, how how do I make it more than Google Is that kinda what you're thinking

Todd Gagne

Yeah. And just I think you have to have a level of creativity to actually really think through these and say, I've got all these problems. I might as well just give it a shot and see how it goes. But the prompt engineering is actually some of the stuff that is the special sauce to make it work well for you. Yep.

Andy Jorgensen

I would agree. I would agree. You have to train it. So you have to feed it information, and it's gonna learn and grow from that information. And the information, like you're saying, is coming from your brain.

It's like your brain is you gotta tell it what it wants, like, what you want it to do. So maybe talk

Todd Gagne

to me a little bit about your journey on kind of the training part of it. Right We call it prompt engineering. But basically, what we're saying is the better the direction you give it, the better the output's going to be. And I think you see some people, struggle with this, I'm kinda curious about your journey and, like, how how did you get better at this

Like, what what things did you do It sounds like you've done some things like act as an expert and basically in law or whatever field you're talking about, or real estate expert. But, like, give me some tips or tricks or ideas that you've done to, like, get better results by kind of, quote, unquote, training the prompts so that you get the results you're looking for.

Andy Jorgensen

I'll use an example that's maybe a little bit more, it's not business related, but I think it'll apply really, really well. And, Todd, everything coming through alright with audio and video and all that stuff Yep. It is. Okay.

So here's an example for you. I was thinking about if I have to write a speech, right, like, either business speech or a sermon for a church, right And I thought, let me just see. A lot of people give really good talks. Martin Luther King, Jr.

Gives really good talk. But in the world of religion, there's a speaker that a lot of people like. His name's, Tim Keller. And so regardless of religious affiliation, you would hear him talk and go, like, wow, that he's such a great speaker. And I thought, I wonder how long it would take for Chad GPT to really be able to write like him, and then if you could actually get it to start speaking like him too if you had, like, the verbal thing.

So what I did was I said I said, hey, why don't you write me a 30 minute sermon on this passage And I want you to come from this perspective of Tim Keller. And it was like, yes. 1st so I didn't start with that. I said, do you know who Tim Keller is

Now just insert whoever your favorite influencer Do you know who Gary v is Do you know who and insert the person. Like, I want you to build me an advisory board like this. So here's what I did. I said, do you know who Tim Keller is

Yes. Do you know this book Yes. Do you know the Bible Yes.

This is okay. I want you to write me a sermon based on this. Boom boom boom boom boom boom boom. And it sounds a little bit more like a little bit like Tim Keller, but I said, Okay. I need you to take it down to a 3rd grade level.

Boom, boom, boom, boom, boom, boom, boom, boom. Okay. It's a little long, and I think that you're using this much a little bit too much. So you basically have to read and review, read and review. You can front load all of that.

So for all of your users that have never used chatgpt, that have never really used any of this sort of stuff, you can front load it. And here's the best resource, is YouTube. Don't read books on AI. Don't read and you listen to this podcast, that's great. If you're in your car and you don't have time, but you can go to youtube.com and just say how to write a prompt.

Well and I'm serious. Like, a lot of people are gonna think you gotta buy some course for a $1,000. Literally, go to YouTube. People are giving away the best for free to do prompt prompt engineering for 5 minutes, and the best 5 minute training will give you. And so what I've learned to do is to front load as much as possible, but you're still gonna have to tweak it until you do that.

But I would say, like, you are this. Does that make sense Yes. And turn it into a conversation. That's the biggest tip that I could say in terms of, like, differentiating from Google, because Google would be like, give me information blah, blah, blah, 1 at a time.

This is like building upon your searches. So let's say, you are Gary Vaynerchuk. Do you know who that is Yes, I know who that is. You are, going to advise a small software startup, based on what they should be doing with their marketing strategy.

Does that make sense Yes. That makes sense. They don't have any money yet, but they might be getting some money. They have some partners.

Here are the partners. Does that make sense And you can basically say, does that make sense Just like if you were to hire somebody on upwork.com, fiverr.com, overseas virtual assistant, you would have to say, hey. Does this make sense

Or if I'm if you and I, Todd, are talking, just think of it like a new human who like I said, this is Rain Man. They're super, super, duper smart, but you don't really know what they are bad at, so you have to tell it what it's kinda good at. So I'd say, Todd, Todd, I need you to to help on real estate, but I don't you don't know me, and I don't know you. Do you know real estate You'd be like, yeah.

Actually, I've I've invested. I'm like, okay, sweet. You've you've established this. Do you know about interest rates Yes.

I do. Okay. So for some of your followers, maybe, if I'm gonna build a software, be like, you are a software expert. Do do you know what it takes to build a software Outline those steps for me.

Sweet. Awesome. Now you're Gary Vaynerchuk who's built the software. Does that make sense

Todd Gagne

Yeah. Oh, totally. And so I think the biggest difference is you're creating a conversation, with with chat versus just putting some information and expecting something static back. And so building on the assumptions that you may have for the ultimate goal that you have.

Andy Jorgensen

Yes. Absolutely. And here's one tip, one pro tip for every single person that uses it, it saves and retains what you said in that conversation. So, for example, if you have a new company, you can have a finance expert, the next conversation have a marketing expert, the next conversation have a marketing coordinator. I may be able to do marketing in 1, but it builds and builds and builds.

And when you refine it, you're refining it within that chat or within that conversation. So as you're talking about, like, it's not Google because it's like if Google remembered what you are trying to get it to do, and it was catering to what you're trying to get it to do, but it lasts forever. So in a month, I can be like, hey. I just invested. Here is the things that went wrong.

I'm gonna feed it that because I want, in 3 years' time, I may be making 1,000,000 of dollars in real estate deals because I trained, and I'm not kidding you. An AI with a an an IQ of a1000. Like, what is I don't know. Chat gpt has an IQ of, like, a1000, or maybe it's, like, 10,000 by now. But just think Albert Einstein has an IQ of, what, a 130.

And over 10 years, I took this brain, and I just fed it a bunch of real estate stuff, and it now is, like, oh, no. Don't do that. You're a human. You're silly. Do this.

And so you can contribute to this brain, and you're making it bigger and bigger and bigger in one specific area, in one specific chat. So I want my Gary Vaynerchuk advice to be one conversation. I want my spreadsheet expert in one conversation, and it gets really good the more that you prompt it.

Todd Gagne

Yeah. That's really good. That's really helpful. Maybe pivoting a little bit to, maybe the problems. So you you hosted this event, AI Edge, and you got a bunch of people to come.

Maybe talk to me a little bit about as you were talking to some of the participants, what were the problems that they're trying to solve How do they think about AI and their businesses, and, like, what's next

Andy Jorgensen

Yeah. So it it varies. Right We had, senator Mike Rounds, the colleges, and then a jewelry shop that attended our AI summit. So we get a broad spectrum.

We get every everybody. And we figured out, like, what is the number one problem that you're you're you're experiencing with this. And I'll just rattle off a few here, of actual meetings that I had afterwards. So there's a capital company. They're on the Inc 500, and they said, this is brilliant.

I said, let's buy I'm gonna buy you lunch, and I hear why it's brilliant. Because if a company like you cares about this summit, what was helpful

Todd Gagne

Yeah.

Andy Jorgensen

And they said well, for them, I think their particular issue is they're very, very automated, but they're like, how can we automate the little stuff Like, for them, it was like customer service, which, man, if AI can do anything, customer service is, like, the first the first thing. And that's been AI doing for a long time. That's the basic thing that they were like, how do we just do customer service Let's take it a step up.

I have a guy that does jewelry. He is a master jeweler. He made my wife's wedding ring, and it looks like a $1,000,000. Right I paid know, good money for it, but it looks like a $1,000,000.

But he deals with the cost of gold, the cost of the diamond, the cost of the metals, The market changes. And in the jewelry industry, I hope nobody takes this business idea because I think we might run with it. Actually, it's probably a software idea for some of your So here you go, guys. Make a $1,000,000. Just buy me a sandwich or something.

He gets the and I'm not kidding you, Todd. A physical book that a 70 year old man makes to update the cost of gold, to update the cost of the materials. So make so this is jewelry shops around the country. Give this guy $250 a year to update all of their, costs so that they can quote things appropriately. So when they're like, hey, I wanna buy a wedding ring for my wife, and they wanna custom build it.

They figure out how much to charge by flipping through a book. And this 65 year old jewelry shop guy is a, you know, he's a brilliant guy. He's like, Andy, couldn't we automate this I said, a 100%, we could automate. A 100%.

You could you could feed an AI and say, we wanna use this type of gold. Actually, we want to use rose gold, not silver gold. We want rose gold with this sort of diamond and then little, one, you know, milligram rubies on the sides. And the AI could just go, boom, boom, boom, boom, boom, boom. So we're we're, like, in talks about fixing that.

I just got done with a lunch, today with Mike, your partner, actually, and me sat down with manufacturing companies. And their their problems are a little bit more just business big problems right now, because we don't want to get into robotics talk yet. But when you talk manufacturing and AI, you talk about robotics. So that's, like, one thing that's, like, we're not quite there yet, but we'll talk about that maybe towards later in this conversation. And then Yep.

HR is huge. So I'm speaking at an HR summit in 3 months' time, and they're like, we just want the 1 on 1 basics. But let me just, like, blow your mind about HR, what's happening right now. You can feed 10,000 resumes to HR, like, AI bots right now. You can feed it 10,000 resumes, and it will give you the best ones.

And you can verifiably, there's HR, like, executives that are like, nope. This is the top 10. This is seriously the top 10. It'll do it in 2 minutes. 10,000 resumes, 2 minutes.

Next layer, this is not later. This is not futuristic. This is right now, 2024. We're sitting in June. You will interview with an AI that is not supposed to be a human being.

It says, hey. This is hi. Hi. I am the AI for Jet's company, and you are applying for this. And you're gonna start talking to it and say, tell me a little bit about yourself.

And then it's gonna start analyzing your microexpressions, your nonverbals. It's gonna an analyze your answers. It's gonna transcribe everything, and it's gonna actually know based on your nervousness, your confidence level. It starts filtering through the the people that had bad English, the people that had bad interview skills. Just like you said, you fed it a bunch of pictures.

You actually live talk to the AI, and it says, these are the top 10 candidates or the top 3 candidates from the top 10. And then the company is like, hey. Our executives are gonna talk to these 3 out of 10,000. And they just, they literally it it's automated, the whole entire thing.

Todd Gagne

So that's pretty incredible. Do you think that, like, society is ready for that Or do you think people just want a job And so, you know, a lot of lot of jobs will go that way, and people will totally be fine with it. I think from an efficiency standpoint, it makes a ton of sense.

And I think it takes a lot of variability out of interviews. Right Like, you know, you and I have probably done a ton of interviews, and it's like you have 5 different people. They all ask different questions trying to combine all that together. I think this could do a lot to solve that problem and give you a better sense.

But, like, there is kind of a social stigma, I think, right now to just talking to an AI bot for my next job.

Andy Jorgensen

Yeah. So here here's what I think of that. Do you remember early internet memes They were just were ugly. It just wasn't pretty.

Like, it it it it it was kind of the I don't wanna I don't wanna peg it. I'll just say it was kind of the, like, dingy Internet days. It was fun. It was super fun. But, like, executives weren't like, oh, greasy memes.

Great. That's great. Like, oh, that's so I can't wait to integrate that into my business because it looks silly. So I had an AI company that I was talking to about sponsoring our event. And they're like, look at our avatars.

And, man, they looked like ugly cartoon like ugly cartoons. And so I think, Todd, the answer to your question, it it turns into a quality question conversation about as soon as I can tell Photoshop let's actually equate it to Photoshop. Photoshop, early days, man, I used to go to Paint, and then I used to Photoshop. You're like, oh, that's super ugly. Now you can Photoshop Brad Pitt's head onto your wife's body.

My body. What I'm saying

Todd Gagne

Yep. Yeah.

Andy Jorgensen

Yeah. And and it but it looks compelling. And now you don't know what's photoshopped and what's not. And you just are like, whatever. Hey.

And you say, that's photoshopped. You go, I don't care. I don't care. It's like, you don't care in an ad if you photoshopped. So you're asking me, do you think there's a barrier to talking to an AI

As long as it's cheesy and looks terrible, then yes. But how far are we from it not being cheesy and terrible We're months from it not being cheesy and terrible. Months. And so when you're talking to someone, as long as it's, this is an AI, just so you know, and I'm Sally, but I'm gonna do your interview, is that okay

They'll be fine with it.

Todd Gagne

Yep. Yeah. It's true. Like, if you looked at, like, the chat gbto, demo that they did, like, that was super conversational. Right

Like, I mean, that was pretty impressive. And so if that was the bot you were talking to, you'd almost forget that there wasn't a human on the other side.

Andy Jorgensen

Yep. Yep. And so we're not even months away. We're, like, weeks away. It's just a matter of, like, is your company gonna be the one that utilizes the top leading technology, which I think is to be democratized very quickly.

So the answer is yes.

Todd Gagne

Yeah. Okay. Good. Well, why don't we pivot a little bit to talking about how entrepreneurs can, like, think about using AI in kind of their applications So I think a lot of the domain that we focus on is is mostly software.

We see, you know, phone apps. We see web apps. And so, you know, I think there's a wide range of stuff in that. I think there's been a lot of discussion and a lot of buzz on this. And I think one of the kind of ideas I'd love to beat up a little bit with you is, how is this continuum gonna happen over time

So we can debate on how fast this is gonna happen. But, you know, today, I feel like we're in an era where, you know, Microsoft just kinda coined the phrase, Copilot. And so, know, if I'm main word and I've got this turned on, then basically, it's trying to anticipate what I'm doing. If I'm doing, code and it's trying to and I'm writing a subroutine, it's gonna anticipate what I'm doing and give me an opportunity to do that. So I'm in control 80% of the time.

The AI is trying to help me out and and and provide some guidance the other 20% of the time. Then you think about it where maybe it's like a stand alone, agent. So something that basically is not other than an interface to tell it what I want to go do. It basically runs independently and returns me a result, and it's single threaded. So maybe that's the second kind of midpoint.

And then over time, do we get to a point where we have multiple agents that are kind of all meshed together And the example I use there is, like, travel. Like, I love to travel, but travel is a pain in the ass to, like, book. Right And so let's say I want to and I did a lot of business travel in my corporate world.

And so, you know, let's say, a client meeting runs long. I miss my flight. What I want the AI bot to do is to really rebook my flight from New York City to Seattle. I need a new, hotel reservation because the last one was a day late. And then, basically, I need an Uber to show up at the right time.

And whatever meeting that I had with a client for that evening, I need to let them know that I'm changing it. And, basically, I need that reservation in OpenTable. Right So there's a whole bunch of elements there that you need to pull together to solve that one problem. So my question to you, Andy, is do you see a progression like that, or do you see it differently

Andy Jorgensen

Yeah. Yeah. Absolutely. So, well, actually, I'm gonna rewind a little bit to you. You actually already fed chat gpt your travel plans because your was it about biking and stuff like that in the parts you wanted to visit

Was that was that you were talking about Yep. Okay. So you fed it, and it gave you the result. You had defeated the thing.

Right So you're saying is, can this evolve to a way where it anticipates what I need And you can say, hey. I'm running late, and it's gonna say, okay. Let me talk to the open table.

Let me do this. Let me do that. Is that what you're saying Like, is it like Yeah. Because there's

Todd Gagne

all these. Yeah. So in my example that I gave you, it was kinda single threaded. Right I gave you all the locations, and I told you how much time I wanna spend, and I told you what I'd like to do.

And so, basically, that was just a problem of, you know, optimizing and basically telling him, okay. What what can you do at this state park What can you do in this national park Where is that other place And they just put it together.

What I'm talking about is, like, you're gonna have to log in to my, Uber account. You have to log in to my Delta account. You're gonna have to log in to, you know, OpenTable. You have all these different tools that basically you're gonna need APIs to solve the problem that I just described.

Andy Jorgensen

Yeah. So maybe this just it doesn't challenge, but maybe, like, fast forwards what I think that already exists. Now it's not ChatGPT yet, but well, let's just think conceptually. Okay Self driving cars k

The self driving car is driving down the street, and Jerry Jerry so and so comes and he pulls in to the side. And the car actually anticipates that this happened, And so it stops and it slows down, but it doesn't stop fully and, like you know what I'm saying It already anticipates something that's unknown, and it adjusts, and you aren't touching the thing. And so there is no there's nobody on the joystick of that self driving car. There's nobody, like, sitting there, like, hey, time for you to stop.

Okay, I will stop. I'm a robot. It actually uses its training and sensory. Here's the second piece of proof that that already exists. It's just maybe not utilized in the all these tools yet.

I already said that when I establish context, it remembers the context and it communicates in one chat gpt conversation. So, for example, I fed it a bunch of stuff in the past about Gary Vaynerchik and Tim Keller, and I wanted to talk like that. And now every time I answer or give an input, it's within the context. Just that one chat, not not all of my chat gpt. It that one chat conversation utilizes that whole conversation's context to answer it within that framework.

So it is retaining that knowledge, and it's making decisions based on that. Now I think your question is, what is the name of the tool that I will be able to use Because it's not ChatGPT yet. Right It's like, how is it gonna be an independent travel agent where you're just talking to a travel agent constantly like a Chat GPT conversation, and you feed it your plans, and then it's like, sounds good.

I will be here for the next month. You've subscribed to a $1,000. You have my and then you just say, I'm late, plain delayed, and they'll say, sweet. No problem. I'm gonna do this for you.

Right That's kinda what you're thinking.

Todd Gagne

It is. And so I don't disagree with you that the technology there for, like, the anticipation part is there. I think what I'm more interested in is, the like, you have to there's a high degree of trust, what I just said. Right I give you my credit card if there's change orders.

Right And then I have to give you access to my, you know, Delta account, my Uber account, my, maybe my Marriott account, and my, OpenTable just to do that example. Right And so you're gonna need to pass information all the way through all of those login, authenticate, and then do an action that basically is tied. So, like, what I'm trying to understand is, doesn't there need to be kind of a a layer that allows you to do that

And I I mean, I think Apple in those latest, you know, in their development conference is talking about some of that. Right Where it's a platform where there's almost webhooks that you can basically take either these little models that are localized on there or chat gbt, and you can hook into them to do certain tasks within an app and do that. So I'm kind of I think that the the scenario is existing and maybe the technology is there, but is there an underlying, and infrastructure that we're gonna have to build out, on each one of these platforms Microsoft is doing this from an OS perspective.

We just talked about Apple. Is there something that's gonna happen in in the web that's like Zapier or something different that allows that connectivity and authentication to actually happen

Andy Jorgensen

Yeah. So I I think that actually has to do with politics of, is Apple gonna play nice with I mean, think about it. Is Apple gonna play nice with Microsoft gonna play nice with Google And right now, they're all fighting against each other, so the connectivity is gonna be kind of everyone's trying to be the connection point. So, I mean, I I could zoom this.

I'm gonna take the mouse and scroll out. Zoom out. Zoom out. Zoom out. And and if you look at, like, if AI is for you or against you is, like, a big fear.

Right And, like and I don't think it's gonna be, like, Terminator or anything like that. But it isn't going to be the AI that's threatening. It's going to be whoever is, like, in charge of that AI because the AI is gonna be, like, useful because of what it's used for. So, for example, if China and America just said, hey.

Let's put together a AI and find a way for us to just have peace in the world, and let's put all our energy and resources to never fighting again. You you bet that would happen. Like, that would happen. We we would be there. The reason we're not gonna do that is because resources on the planet are still scarce, and they still feel hurt by this, and we still feel control freaked about this.

And so here's what I'm gonna zoom back in. These companies have to play nice, Or there's one, like I want you to look up this if your users hear this. It's called Cheatlayer. It's a tool. Cheatlayer.

And it kinda hacks the Internet and does AI stuff with it, and it's pretty impressive. It pretty much it pretty much can do anything. But it you can you can look it up, and it kind of you can say, hey. I want you to tie this to this, and it's a kind of automated chat or, Zapier. It's kind of an automated thing, But I have a feeling that it's gonna breach some privacy and some protocols, and it's probably gonna get sued and that sort of stuff.

And so, I think to the extent at which you're talking about connectivity as the extent at which everyone can collaborate. Right Like, I don't know. Like, you just said that your reserve a table software has to talk to your flight software, has to talk to those. Currently, there's integrations that, like, you know, play nice with each other.

But it might be that there's this travel agency that integrates with all those things, and then it'll be pretty easy because they have a legal relationship. But maybe you're thinking is there just one that's gonna do all the things. Is that what you're thinking No. No.

Todd Gagne

I'm not. I think that, I do think that they're stand alone apps that you're basically connecting and passing information that's got conditional logic to go do something with. Right So Marriott still stands alone in in in their app, but what they're doing is there's a context layer that basically says, Todd's on this business trip, and, and something got delayed. And so everything on his calendar needs to change, and so I need to look at the dominoes that are gonna happen.

And there's an agent that sits there and says, what needs to change And then who do I need to communicate with Okay. I don't need that Uber anymore because he's not gonna be at the airport. I need to communicate with the client that says he was gonna have dinner at 6 o'clock.

We're not doing that anymore. It's gonna be 6 o'clock the next day, and and then and the hotel or the air, the, restaurant's the exact same. Right So it just goes down that list. And, you know, in some cases, this already exists with Zapier a little bit, right, where you can connect all of the plumbing together.

You can't write the conditional logic necessarily the way this is happening, but the chatbot could do that. And so I I I I do agree with you that saying, everybody wants a cut of the action and they want the money, and so they they're gonna control. But I do think that there's a world where all these apps are now being controlled by a user scenario that you dictate, and they connect to it to, like, give you a much better experience where I'm not tracking all that crap down because that's what I was doing when I was traveling. As soon as that happened, like, that whole domino, you're in the lounge trying to figure that stuff out and calling people, and you miss things. Right

You just miss it.

Andy Jorgensen

Yeah. Yeah. So you're saying, like, can my frustrating experience be evolved eventually by AI We're reading crystal balls here. And Ray Ray Dalio says, he who reads the future he who reads crystal balls eats tempered glass, meaning it just, like, it's terrible idea to try to read the future.

So, like, you're asking me to read the crystal ball I honestly think, yes, that's that's my prediction of what's that's worth. I think that's yes because if you just think of AI as a child, the first AIs ever just learned how to pick up a block. That's all they did. And they missed it, and they missed it, and they missed it.

And until they finally and this is a Google executive talk about this. The first AI that they built finally picked up the block, the red block, not the blue block. The red block, not the blue block. And then it got it right a 100 percent of the time because it had learned the red block from the blue block. My 5 month old son, Charlie, is doing the exact same thing.

He is at home figuring out which blocks go into which thing, and it's like this proud dad moment. It's like, okay. That's great. He put the spoon of food in his mouth, not in his ear. Oh my gosh.

And now he's finally done that. If if we, as as humans, start to understand that this is a young mind who is developing its understanding, and it already has the capabilities. You're just wondering if we can just play nice with everything. And I'm saying, well, yeah, there's a Mailchimp integration in every dang software in the planet. Of course, there's gonna be an AI software, and I think that's pretty probable.

Todd Gagne

Yeah. That's cool. Well, I'm excited about the the opportunity there. You bring up an interesting point. You talked a little bit about just robotics and and Google and and and and then Charlie.

So that that's all good. But like, there was a quote that I read the other day that said something about, this woman was saying, I want AI to do my laundry and my dishes, not my art and my writing. And I I I I feel like a lot of people that I talk to about AI, I think they hear the buzz, but I think the things that are important to them are not necessarily the things that are actually we're innovating today. And so the ability over time for us to take AI and put it into robotics, and, yes, we're gonna have figure out batteries, and we're gonna have to figure out power. And there's a whole bunch of things there, but you can see this over time kind of evolving to that, that that intelligence that we're kind of talking about embedded in things.

You know, could it do the dishes for me Could it do laundry Can I take away, some of the menial tasks that we don't like to do so that we can focus on higher value types of things

Andy Jorgensen

Yeah. Yeah. So, I mean, I mean, this guy's not a secret, but Elon Musk is trying to do that right now with the Tesla bots. Like, he's trying his darnedest. I mean, it's kind of I, Robot in real time.

Right It's like you all of us have seen the Will Smith pro. If you haven't seen I, Robot, you need to go watch it right now, because Stop. Just do it. Yeah.

Just pause this episode. Go watch iRobot because I don't think it's gonna turn out like that. But I saw a lot on talking about their the Tesla bots. And I thought, no kidding. This is legit, like the exact function.

And you know, it's funny. I want one. I'm like, yeah. I don't want one. Like, it's gonna it's gonna vacuum your house, and it's gonna, like, it's gonna do all those things.

And it's it's Alon, so there's a chance it's actually gonna happen. Right You know, love him or hate him. He's productive. He's a productive human being, and he makes good in his word.

So I don't know. Is is your question, is that gonna happen Is your question, like, when is that gonna happen Or what's the what's your thought on that

Todd Gagne

Well, I like, maybe talk to me about some of the hurdles to get there. Maybe that's the better part to do it. Right Like, is, power and battery a big component of that Right

I mean, there's batteries in a lot of things. So, you know, if you've got something that's, robotic running around, does it have to be plugged into the wall or, you know, like, there's just a lot of applications that I think are moving in the right direction. Does AI really accelerate this You know, like, I think robotic things in Amazon and and robotics in general has come a long way. As we add AI to it, really, what does that accelerate

Is it accelerate the types of things it can do Is it the the motor function to make better decisions I mean, I think it's a lot of those things. But, like, what do you think some of the predetermined issues are that we have to resolve before we get there to having, the Jensen's, you know, house cleaner for us

Andy Jorgensen

Yeah. Yeah. I mean, off the top of my head, if you if you think about, like, a Roomba, the Roomba is an AI vacuum because when it bumps, it turns and it goes, and it, like, it it's it's code. So it's, like, not necessarily very complex AI, but that is a sort of, like, learning thing where it, like, okay, it's not gonna go in this direction in the house anymore. So it's existed for quite a long time.

So we know that a Tesla bot is simply a very, very advanced Roomba. It's very easy, very easy for it to go low on battery, go back to home base because I've seen my Roomba do that a 1000 times. Right I've you've you've everybody has one of those things. It's like, how does it find its charge station

It just does. Right And, like, it's pretty dang smart for a little vacuum that doesn't talk or have any interface. I think the mean, if you want to talk about a bigger bigger picture, you're gonna talk about computing and processing. So what does it look like to have chat GPT in your Roomba

Right Well, it costs a lot of electricity. And I'm not I'm not smart enough to talk about, like, well, okay. Here's how we're gonna solve all the energy problems. You're right.

Energy is an issue. But the from the actual software perspective to the intelligence perspective, absolutely. You're just gonna beam in chatgpt4 0, and it's gonna beam into this robot here. Let's let's talk about the the the mom who wants their dishes washed. I'll tell you what's going to happen.

The person that can find a sensitive machine that actually doesn't crush my fine plates. Right Bend my spoons. You know, I have a glass, and I have very soft hands because I'm a human being. I have sensitive because I have pain receptors.

And this glass, I sometimes squeeze it, and I'm like, oh, snap. Like, it's gonna slip out of my hands. So I think you have to have an AI that is and maybe we should all research what Elon's really doing, but it has to have haptic sensory touch feedback so that my glass, which is a Walmart like, take a look at this glass right here. For those that are looking at this on video, I'm holding a glass. I can crush it with my hand and I bleed.

Robot will just pick it up, and if it breaks, it'll it'll say, loud noise. Oopsies. Sorry. And then there's glass all over. You know

I don't know. That's probably what I would think of. It's like the sensitivity of a robot doing my dishes in that use case.

Todd Gagne

Yeah. Well, that's good. I mean, it's fun to think about, and it'll be interesting to see. I think our our, next 10 years will be pretty interesting, to do it. So maybe in closing, I'm curious about as you think as you think about, like, entrepreneurs that are listening to this, what do you what's your advice or what guidance would you give them about, about AI in in their designs

I mean, like, you you've talked a lot about tinkering. You've talked about a lot about experimenting. But are there any sort of kind of, maybe guidance that you give them that's saying, you know, like, if you really wanna leverage this technology, do x. Yeah.

Andy Jorgensen

Well, start and just get your phone out, download ChatGPT, and start talking to it. There's literally a verbal mode where you just talk and have a conversation with it. You just say, hey. This is my business and start giving it. I know this is scary to, like, start feeding your data.

What happens to my data Who cares Everybody has your data anyway. Like, we've we all have each other's data. Right

Like, it's it's nobody's dying yet on that. Maybe I'm gonna get sued for saying something like that because I'm insensitive. But but as an entrepreneur, you're risk takers. And if there's anything you can learn from anyone who's done entrepreneurship is you're the guys that are bold enough to just say, oh, heck. Let's do this.

And you take on the risk, and you say, oh, heck. Let's do this. So discover for yourself what ChatGPT can give you insights that you weren't thinking, ideas you weren't thinking. There's a guy named Peter Diamandis. He's an he's an innovator.

He's, like, a, you know, $1,000,000,000 thinker, that sort of guy. He's connected to all the, like, world's best technologies. He said, we are using chatgpt to increase our email open rates. 30%. We're making 1,000,000 of dollars because we feed an email.

When we write a customer email, we feed it, and we say, optimize this for open rates, and it works. And it works better than any e email marketer, you know, that you'd pay a $100,000 for. So what one is in their marketing, they should do that. In your product design, I don't know the answer on the AI application for this quite yet, besides just start asking chat gpt questions and gleaning insights. But before you build a software, and this is you guys at Wildfire, you're gonna, like, make sure that there's a demand for it, that there's lifetime value of customer, that this is actually gonna be bankrolled.

You guys are gonna be good at that. You're already using that. So take whatever gaps are filled before you launch and say, how can I posture and position this And then once you start to launch, literally just start going and trying stuff and failing with chat gpt. Like, just start feeding it.

And, and I'm saying chat gpt, use Claude. You know, use and whatever your flavor of the day is, man, we might in 6 months, chat gpt might be laughable. I don't know. I doubt it. But, they're pretty well backed.

So, I don't know. Just entrepreneurs. Start getting going. Don't go read a book. Don't spend money on a course.

Just go on YouTube. How do I do this And then just start figuring it out. And that's, the guys who take, action, the ones that are gonna be the most fortunate, I think.

Todd Gagne

I think the last part is is dead on, man. Take action. Right You know, get off the sidelines, start playing around with it, make some mistakes, but take some action. So I think that's a great way to place the lead it.

Andy, this has been a fun conversation. I I appreciate your energy, your ideas, and your passion around this. And so thanks for taking the time.

Andy Jorgensen

Yeah. Absolutely. I appreciate it.

Todd Gagne

Okay. Take care. If you enjoyed the podcast today, please just take a moment to, like, rate it or comment on it for us. This feedback really helps us, and it helps us get the word out to, like, helping other entrepreneurs and founders. Thank you.