Generative AI Book Launch – Tom Taulli, Author

This is Text Strong tv. Well, I have a really great pleasure today. I'm joined by Tom Toley.

Tom is author, actually, of his second book on ai. And, uh, he's, uh, both an experienced writer, author. net, which has sold InfoSpace.

Welcome, Tom. Thanks very much for having me. Look forward to it.

Your timing couldn't be better. We were all talking about ai, generative ai, et cetera, more. But tell us a little bit about your book.

Um, the books styles generative ai, how Chat G P T and other AI tools Will Refle Revolutionize Business. And I think that's what's on everybody's mind, right? What does this mean for me?

Uh, maybe tell us a little about, about why you wrote this book, your, your timing just couldn't have been better on it. Mm-hmm. And then we can get into sort of where we are and understanding how generally AI applies, disrupts all of the above.

Mm-hmm. Oh, great. Yeah.

So, uh, as you mentioned, the, the first book I wrote was AI Basics, so it came out in 2019. The funny thing about that is it's selling more today than it was in 2019. So, uh, it gives you an indication of where we're at with the AI market, even though it has nothing to do with generative ai.

But I did have something in there about the early stages of generative ai, and that caught my, you know, caught my attention. And it's something I, I kept track of, especially when we saw these transformer models emerge, and G p t three came along, and I, you know, I just was following it and I thought, wow, this is a really interesting technology. And then last year, and I saw companies like Jasper just skyrocket, and then VCs were getting interested, like Sequoia, and I thought this could be an interesting topic.

And this was before chat G P T came out. Mm-hmm. Um, so I pitched my publisher on the idea, and they, they were kind of, yeah, it sounds interesting.

And, you know, they really weren't too excited about it. Uh, but they went ahead with it. And then Cheche PT came onto the scene, and then, uh, like I said, I had no idea this was gonna happen.

It's not like a, some Nostradamus or, you know, predictor here, but it, it happened and then all of a sudden, oh yeah, this might be a good idea for a book. And, um, the, the challenge though was as I was writing it, there was so much every day, there was so many things going on. And one of my concerns is that, you know, I knew ge uh, G PT four was gonna come out, but I didn't want the book to come out and have nothing about it because it hasn't been released yet.

So it was released while I wrote it, of course. And then there's just a lot of updating that I had to do. Um, so yeah, it was, it was a challenge, but that's kind of, you know, the way it is when it comes to writing about technology.

But, you know, I've been through different cycles before, but th this one is, uh, pretty fast paced. It does remind me a lot about the internet, uh, you know, early days of the internet, which is really exciting and, you know, um, so it's, it's, it was a lot of fun to, to write this book. And, uh, and I learned a lot along the way.

Well, you, it's demonstrating your, you talked about this in your book. I've, I've made it through a little bit of it so far. I definitely will get through all of it.

Um, just how quickly, like chat, pt, number of users adoption surpassed how quickly Facebook and TikTok and all of the, I mean, it was just massive. I think the, but maybe the best way is just having your GitHub account updated with the latest updates. It could.

Maybe that could be a full-time job for you. I hope it isn't, but Yeah. Yeah, Yeah.

Um, I'm curious, so talk a little bit about, you know, generative AI just seemed to kind of go mainstream to AI mainstream for everybody, cuz everybody could try it and see it. And it was impressive. And of course there's a lot of technology behind it.

Um, it, it seems to me we're, we're maybe people are still in that experimentation phase of understanding it, but also we're in this, we, does it mean for me, what does it mean for my job? What does it mean for my company? How should I be adapting to this?

Should I be rethinking my business? Um, you know, are we gonna replace search engines with this, uh, you and I were talking before about contributors on Stack Overflow and Reddit or up in arms, like, are you just gonna scrape my content and, you know, you just be gonna become a, uh, plagiarism engine engine on top of my stuff. And, uh, I go, well, code is he gonna write code for me?

We're all in this. Is this gonna happen? And what does it mean?

And the value that I was contributing, whether I'm right or whatever, is that still gonna be present? It seems like we're in that big question mark. And of course you hear doom and gloom on one side and it's gonna a be an a and helper to us.

And maybe the answer is some variation of all those things. Mm-hmm. Yeah.

You know, um, I've been covering this for a long time and even before I wrote the AI Basics book, and, you know, there's always been those concerns about losing jobs or some dystopian outcome, and it just never happened. So I just think people just, you know, when you cry wolf so many times, people just say, yeah, whatever. And then Chachi PT came along and I was like, wait a minute.

This, this is different. This could actually write my, if I were in high school, it could have written my, uh, my essay and probably would've gotten an a on it. Uh, and uh, so it was just such a, you know, such a huge, uh, leap in technology for so many people.

Although this had been going on for some time, for a lot of people not new to this technology, you know, have not really covered this technology and followed it. It was just spectacular. And now people have these, you know, uh, expectations that when you use ai, it should do what Chachi PT does.

So that has some companies scared, like, we really have to up our game. If we don't get on, you know, on the bandwagon here, we could be left behind. We, you know, even Google could be disrupted.

And if Google couldn't be disrupted, anybody can be disrupted. So on the enterprise side, on the company side, you know, generative AI has become, you know, one of the hot topics at the, you know, at the, the executive suite right up there with cloud migration and cybersecurity. Um, as for just people in general, your employees are wondering, wait a minute, does this mean I won't have a job?

Because, you know, you know, maybe a lot of my job is kind of tedious and, you know, perfunctory and, uh, you know, can Chet PT just do all what I'm doing? So they're starting to wonder, you know, maybe I should think about maybe changing my career. Maybe I should learn about prompt engineering and, and things of that sort.

Um, and then the other, like you mentioned is I think of really legitimate concern too is that the power of this technology is all in the data. And the data is what we people create. And we've created a lot of data over the years.

You know, we're mm-hmm. We're talking huge amounts of data on Wikipedia and Reddit and all these different sources, but a lot of times we've done it, you know, for free. Uh, and, um, you know, but if another company's gonna take that and make money off of it, shouldn't we get paid?

Uh, and it's not just for text, it's for images, uh, too, I think what, uh, Adobe has done is it's pretty interesting. Um, you know, they're trying to find ways to allow their creators to get compensated for this data. But at the same time, for companies who use this data to indemnify them against potential litigation, their belief is that by having this compensation program, there probably won't be litigation or not a lot of litigation or mitigation without merit.

So I think there's some companies doing some pretty, you know, thoughtful things about this. Uh, but you've already seen the lawsuits, like with Getty Images and mm-hmm. You know, it reminds me of what happened with Napster back in the, in the late nineties and, you know, that killed Napster.

So, you know, there, there is this struggle here. We got this great, great technology, but at the same time, we got creators and people who should be, uh, compensated for the work they've done and recognized for the work they've done. Um, and then you all add on top of that that, you know, this is content that is made by people.

We're, you know, we're fallible. We make mistakes, we have biases, we have certain things we're probably not too proud of. And that seeps into the data and that comes into the results.

And that may make it that, you know, this technology is far from perfect in terms of, you know, using it for your company or using it for your own content. So you got you, you know, the human is still very much in the loop, uh, with all this. So, uh, there's a lot going on.

Definitely. It, it's interesting, the, the whole focus around data. It's almost like it, we've elevated the value of data in some ways, but also because of this, how easy it is to access it.

You don't have to have permission, quote unquote, to technically get access to it, to scrape it off of whatever, pull it off of whatever. We both valued it, but also devalued the, maybe the contributor at the same time, except for maybe people will figure out, look, well, we know this is stuff we need to show some value for. We are gonna pay you for not just go rip off content off the now.

That's right. And worry about people chasing us, the trolls. Chase, what do you think, you know, are we headed one, one thought I had, are we headed to kind of the, the Tesla autopilot driving model of, well, yeah, we'll, we'll drive to, to an extent, but there has to be someone behind the wheel and, and yet need to be able to take over.

Is it kind of that model, or is it, do the work, have a human review it? What do you think the, the different mm-hmm. Kinda patterns that we're starting to see emerge?

Well, with chat chi, pt, you still need a human because you gotta have a prompt. Mm-hmm. Uh, we make prompts.

Humans are so far the ones that are making prompts. So, um, so there has to be a human in the loop when it comes to this technology. Now, ha the role of the human is definitely going to change.

Mm-hmm. So maybe the human is not, you know, writing the whole piece of content or creating all the images from scratch. They're using these tools to enable them to do that.

And for some of the processes that maybe they don't know how to do, I don't know how to create an image, you know, I just don't know. But I can describe an image that I want to create that I can do, uh, and that image will will show up. So, you know, I can write, but sometimes I don't, uh, I need an idea or, uh, some inspiration or some examples.

And, and, you know, Chachi pt, uh, can definitely do that. Help me with that, and maybe write that first draft or samples, different samples of first drafts, and then I'll look at 'em, and then I'll mold them to what, to what really fits my voice and what I'm looking for. Um, and also, again, this technology is far from perfect.

Um, you know, I've, I've done different experiments with chat bt, and, you know, some of the, you know, you know, the writer's strike, I'm, I'm in la uh, you know, the writer strike is, you know, AI is probably the first time a strike is involved AI to some extent. Um, but I, I actually used it to try to write some scenes and scripts and they weren't that great. Uh, even what GT for, um, now maybe for some programs it would've been fine.

They're not, not the great programs out there, but, you know, if I'm gonna, you know, some of these shows they'll spend 10, 15 million, you know, Yellowstone, they'll, they'll, they'll spend 30 million on an episode and probably are not gonna use chat G P t to, to do all the scripts. They'll probably have some pretty good script writers for that, but maybe these tools can help them mm-hmm. With that development.

But, um, but yeah, human definitely in the loop, definitely a major contributor. But again, you, you want to understand these technologies and how to make it so you can increase your own productivity and, and capabilities with these tools. It seems like as long as we have, I think one is a differentiation between the quality of writing for Yellowstone and the quality of writing of Mitch and Chachi PT is, you know, is, is universes away from each other, Um, but maybe in more skilled hands you can get closer to.

But it's still, you know, really good writing on shows like that just to pick on. Mm-hmm. Um, but, but also you, you, you mentioned, you mentioned Napster and music.

Mm-hmm. And in your book you talk about education, journalism, gaming, healthcare, finance, just as some of the industries and the impact on those industries. Mm-hmm.

To your point, um, you know, we have robotics and we still have people that work in factories, right? Mm-hmm. Maybe fewer of them, but we produce more goods, higher productivity, but we also have specialists in those technologies.

So it seems some of the uphill is shifts and changes in work, right? As much as replacing or eliminating work, maybe there will be elimination. I tend to be a little more optimistic of, you know, change creates disruptions, it creates new opportunity.

Mm-hmm. I, I agree too. Yeah.

And again, history has shown that whenever there's these disruptions, there's usually a lot of opportunity for people, you know, for employment, entrepreneurial op opportunities, freelance opportunities, job op, there's so much that usually comes out of that. And, uh, you know, let's just even look at the internet, all the types of job descriptions that did not exist before that that existed after. And, and for us, we can't really come up with these, you know, cheche PT came out six, seven months ago, so we're still the, you know, we gotta figure this out.

But yeah, I, I think there'll be a lot more opportunity going forward. Just frame this idea about, I was having a conversation with the company, and we're talking about sort of general uses of generative AI type technologies and, and training large language models and things like that. It seems to me, tell me if I'm on base, if you have, um, sort of a proprietary or access to data that is unique, very valuable because you, on an online SaaS system, you get on it, anonymized information about it, but most other people aren't gonna have.

Mm-hmm. You know, the general models are great, but where your value may be is applying what you do really well mm-hmm. Few other people do into AI models.

And that's where you're really gonna shine. You know, don't go after adding just chat g PT to the, to your product everywhere and make your product harder to use. Not that people will do that, but you know, it's gonna, I, I've seen that already.

Uh, yeah. Uh, yeah, I mean, they'll just, it, to use the open AI api, it's very easy. Uh, you go there, sign up, you actually get the, you know, $5 credit and mm-hmm.

You know, they, they pay based on tokens and it takes a while to actually use that, that $5. Mm-hmm. Um, so it's actually pretty easy.

You can do it in an afternoon and get, uh, you know, get your company up and running on, on the open AI api, but, you know, I could go to chat G P T and, you know, ask it to, you know, to show me, you know, to show a Python script for something I want to average number, whatever, it's, and they'll do it. So you really have to differentiate yourself in this market. And we're, and you, I think we're in that stage where, you know, you know, quickly going, going in that stage where you just can't, you know, throw something, spin it up, and people say, oh, that's great.

And, you know, but you know, they'll quickly realize that, well, I could just go to chat gpt for that. Right. Um, so, uh, it has to be differentiated.

Um, now, you know, for, for example, like customer service, you know, um, we were at the Atlassian conference, for example. Atlassian has, uh, this huge knowledge base of customer, uh, interactions, uh, of millions of customers. Um, not a lot of companies have that type of data, so they can create applications based on that, that for a company, I can get insights that I would not be able to get from my own, uh, you know, small set of data.

Mm-hmm. Uh, or even, you know, someone else who just, you know, some little startup that doesn't have a lot of data, but just, you know, bolted on some open ai. So yeah, if you're evaluating software, I think the first question to say, so, well, what, what's your source of data?

You know, um, you know, uh, if you're just saying, we have the open AI system, well that doesn't really differentiate you, because I could go there myself and sign up for it and start using it myself. So that's not a, that's not, that doesn't really set you apart. What does set you apart is the data, how you use that data, the kind of insights you can get out of that data, and then use cases of how you, you know, go through the, you know, uh, you know, work with a customer and get results.

And I think with Atlassian, they were using this internally and over 50% of their in, uh, uh, it service desk requests were pretty much closed loop automation. Mm-hmm. There, there was no need for human intervention.

So companies would look at that and say, Hmm, that, that sounds pretty good to me. You know, complete, you know, a self-service system that can handle a lot of this internal stuff that, you know, is really not core to our business can save us a lot of money. So I think that's where the, the, the true power of, uh, these systems lie.

And just be careful, you know, the latest startup that comes around claiming all these things, and really all they have is an api. Mm-hmm. It, it's a, it's a good use case of getting you two answers quickly, kind of solving in an another way the search problem, right.

In, in that case, in inside a company's products and things like that. Yeah. But we could, we could go on forever.

I, I definitely want folks to check out your book. Thanks. Uh, it's available on amazon, uh, dot com.

So, uh, make sure I get the title right again here. So it's, uh, generative AI how chat G P T and other AI tools will Revolutionize the Business. Comes right up if you just search on generative ai.

And Tom is the author and check out his other book, uh, artificial Intelligence Basics, A Nontechnical Introduction. Wasn't that long ago. You wrote it.

Maybe it seemed like a while for you, but it's still very relevant. So Tom, appreciate you coming on and we look forward to having you on, uh, some more things with Textron. You're a great, great resource for folks, so I hope they'll reach out to you.

Well, You're a great resource for everyone else too, so I really appreciate it. Awesome. Thank you Tom very much.

We'll see you again. Okay. See you.