Dan Steele - AI, ChatGPT and Tech

Disclaimer: Unedited AI Transcript

Announcer (00:06):

You are connected and you are listening to specifically for seniors, the podcast for those in the Remember When Generation. Today's podcast is available everywhere you listen to podcasts and with video at specifically for senior's YouTube channel. Now, here's your host, Dr. Larry Barsh.

Larry (00:31):

We all have questions about artificial intelligence and chat, G p t. Our guest today on specifically for seniors is Dan Steele. Dan is here to answer those questions. Dan is the author of three books on the subject, the book on chat, g p t, the book on chat, g p t four, and AI unhinged. Dan is a highly accomplished entrepreneur, author, and business strategist with 25 plus years of experience in tech. He co-founded the largest influencer marketing company and is a sought after speaker on various topics, including artificial intelligence, social media, strategy, and entrepreneurship. Welcome to specifically fas Dan.

Dan Steele (01:27):

Thank you for having me, Larry. I'm glad to hop on

Larry (01:31):

Terrific. Let's let's get right to it. Can we start by defining the terms for our listeners who aren't aware of what each of these means? What is ai?

Dan Steele (01:46):

So, pretty simply, AI is when you take a lot of data, a robust set, and you combine it with computer science to enable problem solving at a really fast pace

Larry (02:01):

And chat. G p t

Dan Steele (02:04):

Chat, g p t, is a form of artificial intelligence that allow that generat, it was trained with this massive dataset, including most of the internet and a lot of language. So it does language best. But it, it's a, it's what we call a generative ai. And what a generative AI system does is, is it creates things so you feed it a massive dataset and then you train it on what it should and shouldn't do. And eventually it'll, it'll do creation on its own.

Larry (02:40):

And AI can also be used in image generation, music generation.

Dan Steele (02:48):

Yes. It so I, I guess to to give an example of what's significant about what's, what's happening now, we've, we've had AI for years. I don't know if you remember when IBM Watson won on Jeopardy, that that was all artificial intelligence a company I started in 2013, we did a partnership with ibm. We had to raise about 15 million in venture capital to build what we needed to build in our partnership with ibm. But what's happened now is it's almost democratized artificial intelligence chat, G P T, and it's made it accessible to where you don't need to raise what a company doesn't need to raise 15 million to build a product anymore. And with that comes a lot of people doing a lot of things, and it's moving really fast and it's kind of scary for people.

Larry (03:41):

So what does the future hold with this onslaught of ai? Are we all in danger of coming face to face with the Terminator?

Dan Steele (03:53):

That's the, there's a lot of theories and anybody who claims to know, I think would be a little disingenuous with that. So I, I went, when I was a kid I'd go to the hardware store and there's a big funnel and you'd put a quarter in it and it would roll around the funnel. I don't know if they do 'em anymore. So I, I am, I talk about it as if the internet's the very top of the funnel, the quarter going, and then around the middle of the funnel is when we invented the cell phone. And then at the very bottom, where it goes really, really, really fast is kinda where we're at now with with the way all of this is working. So to know exactly what it's gonna hold, I think people are making a lot of money claiming to know exactly what it's gonna hold, but I don't think they, I don't think anybody knows.

Larry (04:40):

The big news this week is that the so-called godfather of ai, Jeffrey Hinton, left Google to speak about the dangers of AI and it being an existential threat to humanity. What's your feeling on this?

Dan Steele (04:59):

That's, it's really interesting because so on one hand you have people coming from these big companies that have a massive advantage, especially if we slow down and there's a lot of regulation on it. The more regulation you have, the low, the more it'll slow down innovation. So you have folks stepping outta big companies and saying, Hey, we need a lot of regulation on this where you have other countries who probably have little to no regulation or a more uniform approach to taking, taking on ai. So it, it, it is interesting because if it is as dangerous as everybody, as these people are saying it is, we probably want to slow down, but also we wanna remain competitive with other countries. So thank goodness I'm not the president and I don't have to make these decisions.

Larry (05:48):

Okay, so we've gotten the threat version of AI out of the way. Let's talk about the benefits of ai. Where, where do we see it in use today?

Dan Steele (06:01):

So as I mentioned earlier, by making it more affordable, we're seeing a lot more uses. So an influential the product we built was it was an advertising product. So imagine you're Disney and you want to promote a new film, and we would go and we would match the audience that you're looking to reach. We would figure out who that audience was, and then we would find people with big social media followings that shared that same audience. And then we would use machine learning and natural language processing to figure out how those people talked about Disney. And we would create content that was very authentic to the Disney fans and the content. So we you know, kind of taught the influencers how to speak using artificial intelligence. And that content was a very effective ad. It was so effective that it was that it was getting more likes and tweets and shares than their normal content.

Dan Steele (07:01):

Like I said, that cost us about 15 million to do. Now I can go and I can prompt Chad G P T, and I can solve a lot of the same problems, and I can do that as an individual for about $10. Do all of that processing that we used to have to do and all that build and everything. We, we can go from 15 million down to $10. And what that allows is for small businesses that you, when you're a small business, you can't afford to hire a big ad agency, and big ad agencies would do things like search engine optimization. So, you know, when people would search your name, you, you, you would wanna be the top ranked for, you know, Larry, right? That was very expensive. Some prohibitively like 10 to $30,000 a month, and now individuals are able to use this service, just the ch core chat G B T service for $20 a month, and then another generative AI that would create images for you for like, maybe another $20 a month. And they're able to compete for search engine terms with these companies that are paying 10 to $20,000 a month. And, and it, it covers a lot of different fields, but that, that's a good example.

Larry (08:13):

So I could promote this podcast using ai.

Dan Steele (08:18):

Yeah, I a good, a good example is a friend went to work for an attorney and he got that attorney to rank for number one for dog bites in the Las Vegas metro area, and a couple other terms that were really, really expensive. And he was able to, you know, hire this kid, for lack of a better term, you know, for not that much money. And he was competing for these search terms where he wouldn't have been able to compete before.

Larry (08:47):

AI is in use in a lot of things that most people are unaware of autonomous vehicles and aircraft, Facebook, Twitter, most social media, digital assistance. Siri is a great example of that. Where else would you see this in use

Dan Steele (09:11):

In the grocery stores? A great example is when we, let's say when there was a, a shortage of toilet paper a few years ago, all of the it's called timely D delivery. So everything that sits on the shelf in a grocery store is there as short as it can. So we know that if that store, this store in this town sells 15 packages of toilet paper each day, they're gonna have about 16 packages of toilet paper, and they're not going to have a warehouse between the manufacturer and the store. It's just shipping straight out to stores now. So they're, they're cutting out steps and making things more efficient. But a negative side to that is they only get their 16 things to toilet paper every day. So if people go buy a bunch of them, the computer can't correct fast enough to keep the inventory stocked. It's more a, a stable environment. It makes it so that there's less waste and there's less carbon footprint because you're not filling another storage facility. You're not trucking it from one place to another to another. You're just going straight from the manufacturer. But it still has its limitations as well.

Larry (10:18):

What about when we're in our automobile and we use Google Maps or apple Maps ways?

Dan Steele (10:28):

Yeah. Waze is a great example of you know, one of the first places that people would've inter noticeably interacted with artificial intelligence. What ways does, it doesn't provide you just a route, it takes all of the cars that are using the service and it sees, okay, this car is supposed to be going 55, but it's going 20 something's going on. Oh, now we have five cars that are going 20 miles an hour. I'm going to start rerouting people around this congestion zone. And it would, in real time, it would update and learn and start rerouting you on maps. Google bought them very early on, but they allowed them to stay autonomous and build on their own for quite a while. And now I believe Apple's even buying data from them. So it's telling you when there's an accident ahead and it's rerouting you and that sort of thing.

Larry (11:17):

So it's more than just people reporting

Dan Steele (11:20):

Yeah. 

Larry (11:20):

What the traffic is.

Dan Steele (11:22):

Yeah. It it's even taking into consideration weather reports now,

Larry (11:28):

But there can be abuses as well as all these uses. Yes, plagiarism, deep fakes

Dan Steele (11:36):

It. We have a really it's really interesting to see because for instance, YouTube, which started in like 2003, 2004, somewhere in there, I'm not sure the exact year, we're just now starting to legislate around intellectual property on YouTube, you, you know, like 18, 20 years later. And we're, it, and it's, we're building a pretty good system too. But it's really, really slow with the speed at which AI is moving. I don't know that we're gonna be able to keep up with this and I think it's gonna be more of a, a, a bang for forgiveness than ask for permission thing. I think a lot of the laws and stuff will be we tend to do this online gambling is a really good example of you know, some companies like poker stars and party Poker, they all came out and they just launched in the US market. Eventually the government said it was illegal, they paid fines, and then within, you know, five, 10 years, they're able to operate their own online casinos. Here. I think we're gonna have the, the only way that we're gonna with the way that our government works is that's how the legislation's gonna go around. Copyright, that sort of thing here.

Larry (13:00):

A lot of people are worried about deep fakes and, and fake news bots. Is this gonna be a problem?

Dan Steele (13:09):

Yes. there's re there's, it's almost getting, it's almost getting to the point where you can't tell whe whether something's real or fake. And unfortunately with the way that the incentives are lined up for media you know, websites in particular, the, the more people come to their website, the more money they make. So right now most of the major news publications or all of them across the board will put out stories and then go back and retract things as opposed to the old way of journalism where you had to have confirmation three times. So I, I fear that, you know, we'll put some deep fakes will get put out there, altered videos, that sort of thing, and then two weeks later when it doesn't matter, they'll say, oh, we're sorry. We didn't mean to do that. But I, I don't know the solution for that either.

Larry (14:09):

Yeah, I was just going to ask you, is there any way to determine what's real and what's fake? I mean, just as the user without computer systems to help us along,

Dan Steele (14:20):

There's people building software but nothing super accurate yet. It's, you know, was it believe none of what you hear in half of what you see <laugh>, I'd probably take that down to 5%. Oh,

Larry (14:36):

Really? <Laugh>. Yeah. It's getting to be that much of a problem.

Dan Steele (14:40):

Yeah.

Larry (14:41):

Let's, let's talk about the future of AI and how it can benefit older adults. This is specifically for seniors. Yeah. So what, what can we look forward to?

Dan Steele (14:56):

So automated systems inside the home, you know, it's no longer just life alert. You, you have you know, you can have these little speakers put throughout your house. Examples like my wife and I have recently got into cooking a lot at home. So there's only two of us, and most recipes are for 48 people. So we use, we, we will put a copy and paste recipe into chat G P T and say, Hey, make this for two people. Or, Hey, here's the five recipes we wanna make this week. Cut 'em down to two people and then build us a a grocery list. A lot of simple tasks will become a lot easier now. And accessibility to just the world is their, you know, delivery. A Amazon is a great example. My mom is a senior citizen, lives in a retirement home in Michigan.

Dan Steele (15:56):

I'm able to order for her and all of the deliveries show up to her and all, she doesn't, she doesn't have to do much other than opening boxes now. When, you know, when she first moved in, she tried going to the store herself, and it was, she had to schedule a ride because she can't drive and all those things. Mm-Hmm. <affirmative>. And we were able to use Alexa through Amazon, and she's able to make shopping lists right through that. So there's, there's a lot of you know, when I was a kid, having bifocals was the sign that you were getting old. And I went in this year and I had to get what would've been bifocals, but now they call 'em progressive lenses, so now you can't even see that I'm getting old <laugh> as much. I, I think that it's gonna continue. The technology's going to continue to advance like that.

Larry (16:41):

Our white beards don't give anything away. Yeah.

Dan Steele (16:44):

<Laugh>,

Larry (16:46):

I, I wish the strap on the headphones was a little wider. <Laugh>

Dan Steele (16:50):

Cover up some

Larry (16:51):

Of the white and the bald spots. What about virtual companions interaction with robots? Is that coming for seniors?

Dan Steele (17:01):

I, you know, there's still something authentic. So the, the books that I wrote, we'll get to them, but I I sent one to my, my best friend's mom. She reads everything that I do, whether it's a business proposal or anything. And she was saying, you know, when when she'll read, you know, long form texts that I wrote, she would fly through the pages. She could feel my soul in it. You know, she's like, that's a person. But then when she read the book, she's like, it was just, I mean, it was structured perfectly. Grammar was perfect, but it, it didn't have that authentic Dan Steel behind it, you know? So I, I think we're a little ways away from that. Simple conversation, simple chat. Sure. But a deep, authentic conversation. You, you just can't, you can't fake that yet.

Larry (17:51):

So there won't be robots making meals for us.

Dan Steele (17:56):

We're, we're pretty far out from that.

Larry (17:58):

There's some talk about health trackers and waste to detect cognitive decline Yes. With AI as well.

Dan Steele (18:13):

Yeah. So there's there's different websites. I haven't spent too much time looking at it on, on the cognitive side. I did get apple Watch for the first time this year, and it does track. You know, I was I was coming down with a cold and I didn't realize, you know, you didn't realize it, but it was able to alert me, like, Hey, your breathing is off and your heart rate's a little off. You might want to you know, see a doctor get things checked out. And lo and behold, the next day I had a cold and I was able to stay home. But he got ahead of it. Just from looking at my vitals and sharing that information with healthcare will be will be invaluable.

Larry (18:54):

AI sounds like a miraculous tool, but there are things that cannot do.

Dan Steele (19:03):

Yes, yes.

Larry (19:05):

Like

Dan Steele (19:06):

Like give you a human experience. It can't make moral decisions. Thank goodness. It, it's programmed around that. But it, it's mostly around that authentic human experience. It's just not there. And I, I don't see it getting there anytime soon.

Larry (19:28):

I asked chat g p t at one time to tell me a few jokes. Okay. And that was another experience of things that AI doesn't do well, it has no sense of humor. <Laugh>, it, it listed about 12 dad jokes that were the worst ones I had ever heard or read, rather. So, yeah.

Dan Steele (19:54):

But if you fed it and archive cause we were talking about, you know, using computers with big data sets, if you fed it an archive of just comedians jokes, it would be able to read the jokes and it would be able to come up with a concept of funny. However, a comedian, which I think is one of the purest art forms, they spend a lifetime working on timing, reading a crowd, that sort of thing. Nuances, like you say the joke, and then there's a tagline after the joke. And sometimes you use the tagline, sometimes you don't, depending on how the crowd reacts, that that's ju we're so far from that with computers that I, I wouldn't be too worried about it.

Larry (20:31):

So, nuanced language is not gonna be a, a big feature of ai.

Dan Steele (20:39):

Not, I mean, that, that's where I could get proven wrong, but I see that as, you know, 10 years down the road, maybe, you know, the, the I, I guess the big picture of where I see it, I, and I guess this this impacts everybody. Imagine sitting down on Friday night and saying, I wanna watch Martin's film in the style of Martin's Scorsese sat in the 1950s with Leonardo DiCaprio, your star, and it's a romance. And then within 10, 15 minutes you could just watch that movie and you're just limited to how creative you are. That's gonna come a lot from more people using the AI and it learning more. But I I, I see that as like a, a long-term possibility.

Larry (21:27):

So you're alluding to the fact that maybe AI will actually help us improve our language skills?

Dan Steele (21:36):

Oh, absolutely. Absolutely. And one of the applications that, so I, I, I deal with a lot of other technologists and, you know, when this all came out, we kind of just had a phone call to see, you know, where what, what we were all thinking. And one of the applications that we thought would be the best right now is a way for people to chat with a small delay. So Cuban Spanish, and Argentinian Spanish is very different. It's almost like two different languages, but the AI models understand the difference in region. So I could say something in you know, Southern English, and it could go out to somebody who speaks from Argentina and somebody from Cuba, and it can come to them in a natural way, you know, with like a maybe one to two second delay. The, that's one that's the most interesting application right now.

Larry (22:31):

Yeah. I, and one of the things that I've found playing with chat G P T a lot depend, and, and the visual AI processes as well. You really have to be pretty skilled in your input to get good stuff out.

Dan Steele (22:55):

Yes. That so for writing the book, the, the way that I came up with the first book is I was talking with my wife and I said, wouldn't it be funny if I could get chat G B T to write a book about chat G B T and the outputs? So the normal paragraph is about 2000 words and the outputs, the longest output I could get out was 600 words. So I had to play with it. And so what I did is I broke each chapter down into sections, and then when I wrote, when it spit out a section, I'd be like, okay, now write section two. And I'd be reading it and it would just repeat this stuff from section one. So then I'd have to improve the prompt. It took me about a week to get all my prompts down in order to do that. But that's changing. It's the, that's where most of the consumer facing applications are, are, or, you know, the stuff that you and I would use every day would be, the different things are just making it easier and not having to use the be so good with the prompts. And that's where the, the short-term winners are going to be.

Larry (24:03):

So you mentioned your books. Did you actually use chat G p t to write these three books?

Dan Steele (24:11):

Yes. Yep. Really? Absolutely. Yeah. It took the process of writing a book di it took me, obviously I had to modify some language, but for the most part they were all written and structured by chat G P T, and it took the process of writing a book from, I've tried, I've maybe tried to start writing 50 or 60 books in my life, never got past the first chapter. And now I, I can get a book out in about eight hours.

Larry (24:42):

It, it, it gives me, I tried to write a novel at one time, and I'm gonna say the first chapter was pretty good. The rest of the 55,000 words <laugh> were awful. Would you suggest I put that first chapter into chat g p t as a prompt?

Dan Steele (25:08):

That could be interesting and see where it goes. Yeah, so I would you have all the chapter, you have the chapters laid out for your novel. No,

Larry (25:20):

That's one problems. I'm

Dan Steele (25:23):

Oh, you, you, you don't have like a, the story started here and then it just kind of gets lost in chapters two and three

Larry (25:29):

I and four and five, and it gets Yeah. Yeah.

Dan Steele (25:34):

That, that's you, that's definitely something that could help.

Larry (25:36):

I will. That's interesting. I'm gonna give it a try. And I wonder if people listening would try it as well. I had never heard the, the, the comments you made about recipes, and I think that might be of a lot of interest to some of the people listening here.

Dan Steele (26:00):

And then for the folks with grandkids and great grandkids, my, my friend, his wife, my friend and his wife made a application called Tall Tales. And what you do is you input the kids' name. So let's say Lucy and something that they're interested in, say, unicorns, and it spits out a bedtime story about Lucy playing with unicorns. And you can be the coolest grandmother or great-grandma if you can sit down with a kid. It's called Tall Tales. It's on iPhone. And you can do a unique bedtime story every night for him.

Larry (26:33):

It's a nap for,

Dan Steele (26:37):

For iPhone. Ah, and it's called Tall Tales.

Larry (26:41):

Tall Tales. I'm gonna note that in the description below, because I think that would be real fun for grandparents

Dan Steele (26:51):

Especially. I know I, I know I slip behind on technology sometimes, but to, to be the one, the first one to show it to a kid, I think that'd be pretty, pretty cool and pretty special.

Larry (27:03):

How did you come up with the idea of using chat g p t to write the books?

Dan Steele (27:08):

It just riding in the car with my wife. Like, wouldn't it be funny if I could do this? You know, it was kind of a, a cheeky thing. And then, then I, I like finishing tasks, so I like when I got started with it, there's a lot, it was a lot more complicated than I thought. You know, I, I had it write the book and then I kind of tried to lay it out the way that I thought a book would be laid out. And you come to find out that Amazon has their own templates and those templates that you get from Amazon are not the easiest to use, you know? So I spent quite a bit of time learning those templates. So it, it turned into just this learning process and the ability to go from, Hey, I have this idea to, hey, somebody paid for it is only, it only takes about two or three days before people are able to buy your book.

Larry (27:55):

Did we miss anything? Anything you'd like to add?

Dan Steele (28:02):

I would say the, I I really feel like this technology levels the playing field, and you don't have to be a genius sitting in a laboratory in Silicon Valley in order to do things anymore. And there's a lot of opportunity, especially for older folks, cuz it just makes things easier. And if you can embrace it and figure out how to use it, life will just be a little bit better.

Larry (28:26):

This has been great fun, Dan. I, I, I think people are gonna appreciate learning a little bit more and getting over the panic that the media seems to be instilling in everybody. It probably makes good copy.

Dan Steele (28:43):

Yeah, ma it, it gets clicks and people, it gets eyeballs. You know, if, if the headline was, everything's normal, nothing to worry about, would you click the article? Yeah. <laugh>,

Larry (28:53):

Thanks again, Dan. This has been super.

Dan Steele (28:57):

All right. Thanks for having me.

Announcer (28:59):

If you found this podcast interesting, fun or helpful, tell your friends and family and click on the follow or subscribe button. We'll let you know when new episodes are available. You've been listening to specifically for seniors. We'll talk more next time. Stay connected.