CrushBank CTO, David Tan, connects the dots between early days of fantasy football and the ChatGPT hype of today. Along the way, he co-founds a company to help bring AI sanity to MSPs.
CrushBank CTO, David Tan, connects the dots between early days of fantasy football and the ChatGPT hype of today. Along the way, he co-founds a company to help bring AI sanity to MSPs.
Hello everyone and welcome to the Crush Bank AI for MSPs podcast, a podcast where we talk about everything related to artificial intelligence, machine learning, managed services and really anything else that I think about that I think you might find interesting that I feel like talking about. My name is David Tan, I'm your host and I'm really excited about today's episode because I think it's going to be a lot of fun. We're going to talk about what is arguably the hottest technology on the planet right now generative AI. But that's not it. We're going to talk about everything from financial reporting to high school football to fantasy football and a bunch of things in between, and if you have no idea how those things can be connected, strap in, because this is going to be a bit of a ride. We're going to have a little bit of story time. I'm going to talk about how I got interested in AI. I'm going to talk about the power of generative AI a bit. I'm going to talk about some of the dangers of generative AI. But, to get started, the idea behind this podcast is not to talk down to my audience. I fully expect that most of you will know what generative AI is, understand the technology, have a pretty good idea of what I'm talking about, so I'm really going to try and skip the basics where possible. That being said, I also think there's a really good possibility the only person that listens to these episodes is going to be my mom. She doesn't really know what generative AI is, so, for my mom's sake and anyone else, that's not technical.
Speaker 1:Let's start with a little bit of background. Generative AI is, as it sounds, a type of artificial intelligence where large language models, or large models, I should say, in general, are used to create original content. Now, that content can be anything from text to images, to computer code and a lot of things in between. Obviously, generative AI hit the mainstream back in about November of 2022, when OpenAI released their ChatGPT product, gave people the ability to log on for free and chat with a computer bot that could do anything from answer questions, create original content, write stories, write narratives, create outlines, build recipes you name it. You can do it with ChatGPT, but before ChatGPT, openai was getting a little bit of publicity around their Dolly product. Dolly is also a generative AI engine. The difference there is a graphic generative engine. What that means is you can describe a picture and Dolly will spit out what that picture would look like. So, in other words, you could tell Dolly to create you a picture of an Eskimo riding a motorcycle through Times Square on New Year's Eve and it would come back with something that probably doesn't accurately represent that, but at least in name represents that. In other words, it will have all of those elements. It will be recognizable as what you asked for. When OpenAI released ChatGPT, like I said, it got incredibly popular, incredibly mainstream, and everyone started using it and since then probably the year almost a year since not quite a lot of other companies have come out with similar technologies that people are starting to leverage in all aspects of their business. I mentioned, you can use it to create code. There's a lot of low code platforms built on generative AI. People are using it to create content. Like I said, people are using it to create images, slideshow, decks, things like that. You name it. There are really powerful models that can be used to create that.
Speaker 1:And again, I'm also not going to get too deeply technical on some of these episodes. There may be some where I am, but this one's a little bit more about having fun, a little bit more story time, so I thought it might be fun to take a step back and give you guys a little bit of a background of why I first ever got interested in artificial intelligence and where my fascination with some of this technology came from. So I'm going to go back, flashback to 1994. My business partner, evan, who you will hear on a bunch of these episodes, and I had just started our IT company. We were two snot-nosed kids living at our parents' homes, no expenses, really no idea what we wanted to do. But we got a great opportunity and we started an IT company. That's a long story for another day. Maybe it'll come up on another episode. That's not important. What's important was we really didn't have much to do. We had a couple of customers, we were trying to build a business, but again, this predated the internet in a lot of ways That'll become relevant in just a minute. It predated what we know of as managed services. It certainly predated the cloud. We got our start back then, just to give you an idea from a time frame standpoint gripping out Novel Network and installing Windows NT for businesses. I remember vividly having to sell our clients on email so not to pay to myself as someone that's a million years old, but you get the idea of when this was.
Speaker 1:So one of our friends had a brilliant idea of starting a fantasy football league. We thought it was great because we love football. Listen, I'm not going to say we were like the guys sitting in the back of a diner that created a rotisserie baseball back in the late 70s or early 80s I think it was but we were fairly early on in the fantasy football days and again, we're talking a time before mainstream internet as you think of it today, no worldwide web as you know it. So all of this stuff was done manually. So basically there was 12 of us in the league, a bunch of guys that all grew up together, went to high school together, great friends, 12 of us in the league and Evan and I were the commissioners. We ran the league. We were responsible essentially for managing the lineups and the scoring and basically the way it would work back then was on Sunday mornings, everyone in the league would call me, leave a message on my answering machine and leave their lineup, and then Sunday night I would download the box scores from copy server, prodigy, whatever dial up service I was using, and we would manually or with spreadsheets we'd compile the scores and then Tuesday morning after the Monday night football game we would fax the final scores out to all of our friends. And again, like I said, we didn't have much business to do at the time, so we certainly had a bunch of spare time. So I also took it upon myself to start writing newsletters Really, quite frankly, it was a way to make fun of each other. But it was also good fun. We called it the Pigskin Chronicles and it was a great time. I would write articles about the individual games I would write about. I'd have some sort of an interest piece, like talk about a made-up story about someone's life. It was fun. It sounds a little bit childish and probably nonsensical to you now, but trust me, at the time it was a good time. So that was 1994.
Speaker 1:Flash forward to about 1999, 2000. The internet as we know it started to come more into shape. Obviouslycom was all the rage back then and a bunch of websites to run fantasy football leagues started to sprout up. We chose CBS Sportsline for some reason. I don't remember if it was the only one or if we liked it, if it was free, if it was the cheapest, don't remember, but it's what we chose. We still use it to this day and you could do all of everything online. You could do it electronically. I didn't have to manually calculate scores, I didn't have to listen to my answering machine to get lineups. Everyone could handle it themselves. They could create their own lineups. And what would happen was, as time went on, cbs would send out an email every Tuesday morning with the results of the games, the final scores, but also a bit of a narrative.
Speaker 1:Now, at this point our business had started to grow so I didn't have time to write the pigs in chronicles anymore, and I can assure you CBS Sportsline, who had probably hundreds of thousands, if not millions, of leagues by this point in the early 2000s, certainly didn't have staff writers that were writing fantasy box score stories about the junkies that was my team playing the sledgehammers that was Evan's team. But every Tuesday we would get a newsletter from CBS that had a fairly well-written narrative about the outcome of the game and it would talk about coaching mistakes and sitting and starting the wrong player and basically just describe the games and the outcomes of the week, and then they would come out with another one which is a preview for the next week, and I was fascinated to find out how they did it. So I dug into it and it turns out they were using a very early form of generative AI. So basically they fed details about the scores of the game, the rosters, the possible starters and sit and players that could be set free agents, things like that. They'd feed all that into the machine I'm oversimplifying it, obviously and the machine would spit out a narrative that looked like an article about the particular game and, like I said, you couldn't necessarily tell that it was written by a computer. It was interesting. As time went on you started to sense a little bit that was written by a computer, but still very cool. And, like I said, I was super fascinated with the technology and I found out where it came from.
Speaker 1:So a few years prior to that, reuters, who had to cover financial reporting news, found themselves in a bit of a dilemma. So if you think about public companies, they report earnings quarterly, obviously, and big companies like Apple, google, microsoft, you know, caterpillar, you name it the large companies have teams of analysts that listen in on earnings calls and they dig through the filings and they look inside the P&Ls and the balance sheet and they write long narrative stories about the performance of these companies. That's all well and good for these large companies, but if you look at a small company that maybe trades over the counter, they're still responsible for reporting their financials every quarter and they're still interested to the stockholders to read about those reportings. So what Reuters did was they basically fed the same type of financial reporting data into a language model. I'm not going to call it a large language model because it very much predated what we have today. So they would feed these into a language model and they would write a financial news story about some over the counter stock. The company reported earnings. Again. The stock was probably trading at 12 cents.
Speaker 1:So I have no idea who was reading it probably the same people that are going to listen to my podcast but it was a great solution for Reuters and that kind of grew into, like I said, the technology that was used by fantasy football and that was really the precursor of what we see as generative AI now and again. It fascinated me. I've always been fascinated with technology, emerging technology specifically. I wouldn't say that I sat down and started to try and figure out business uses of it. It was the type of thing that was kind of in the back of my mind. I've told that story many times. It's like I said, it's lit the spark around AI specifically for me and when we started Crush Bank initially back in 2017, one of the reasons I was so fascinated with AI was kind of where that came from.
Speaker 1:So I'm going to fast forward now and this is going to get much more relevant in a minute. So thank you for indulging me and learning a little bit of my background. I'm going to fast forward now to a newspaper article from the Columbus Dispatch. It's a newspaper in Columbus, ohio, a date line August 18, 2023. It is a article about a high school football team game between two rival high schools and I'm going to read it. So please indulge me. I'm not going to read the entire thing, but so indulge me for a minute. I'm going to read a little bit of it to you.
Speaker 1:So the headline is Westerville North escapes Westerville Central in thin wind in Ohio, high School football action, and the byline of that is lead AI L-E-D-E-A-I. Obviously, lead AI. Well, not, obviously, but lead AI is a service that newspapers, news stations, newsrooms in general any news reporting service can subscribe to that will essentially quote, unquote, replace journalists and write AI articles for you Fairly well known product. I know a lot of newsrooms use it, so this is not just not meant to be critical of them. It's more of a conversation around generative AI and some of the things we should be thinking of when we embrace this technology. So that headline again West escapes Westerville Central and thin wind in Ohio not what your typical sports journalist would write, obviously right.
Speaker 1:I grew up reading Mike Lupica in the New York Post, right. I used to wake up early every morning to get the post to read Mike Lupica's article about the Ranger game from the night before. When I was in college I was in Michigan and Mitch Albom was the journalist of Note in Michigan and he wrote stories about the fab five or Michigan's football gate teams or things like that, and we couldn't wait to read them. Anything Mitch Albom wrote was gold. This is not the type of headline that someone like a Mitch Albom or Mike Lupica would write. Obviously, journalism in general has changed significantly in the last 30 to 35 years, but you can tell reading that that it is not, that it is written by AI and to the Columbus Dispatches credit they. The byline is the AI product that they use to write this article.
Speaker 1:I'm gonna read the first couple paragraphs, so please indulge me for another few seconds. The Westerville North Warriors defeated the Westerville Central War Hawks 2112 in an Ohio high school football game on Friday. Westerville North edged Westerville Central at 2112 in a close encounter of the athletic kind at Westerville North High on August 18th In Ohio football action. Those are the first two paragraphs. The first paragraph of a news article, if you don't know, is called the lead LED and the idea of it is to grasp your attention and make you interested in the article. Both of those paragraphs would qualify as a lead. So in other words they're both meant to be first paragraph Pieces or parts of a newspaper article. So that's the first strike there.
Speaker 1:The second strike is it just sounds very awkward, right? If you read it again. It's not what a ambitious Journalist, what journalists, would write, especially when and I'm extrapolating out now here but especially when you consider that someone covering high school football in Ohio Probably has dreams of covering the Browns or the Bengals or even Ohio State, not Westerville North versus Westerville Central. But you need to start somewhere and you would assume they would put their best foot forward. So we'll kind of leave all of that for a minute.
Speaker 1:I want to focus on one sentence there. I read it to you. Hopefully you caught it, hopefully you picked up on it when I was talking about it, but if not, just in case, I'm gonna read it again Westerville North edged Westerville Central 2112 in a close encounter of the athletic kind at Westerville North High. That sentence Jumps right out of the page at you. That is absolutely not something that a sportscaster would write. Our sports journalist would write. I should say again it's a I, it's meant to sound like a human, some ones and zeros got crossed someplace. So it thinks that that is a Compelling sentence in a sports article. It is not.
Speaker 1:But let's go a step further. So I find that sentence very interesting. So I want you to play along at home with me now and I want you to open up Google and I want you to type a close Encounter of the athletic kind into your search bar and as soon as you do that, I want you to switch to the news tab. So, in other words, you're getting news articles and you will notice dozens and dozens of articles that come up and if you click on any one of them, let's see. I'm gonna go back. I'm gonna go to the fourth page here I'm gonna go back to let's see how about January 10th of 2023? I'm gonna click on the article and I'm gonna read it for you. Raymond Lincoln would edge verdant North Mac 41 34 in a close encounter of the athletic kind on January 10th in Illinois boys high school basketball.
Speaker 1:And every one of them is the same. You can find that sentence no exaggeration in 50 articles without the blink of an eye, and probably even way more than that. And again, which is actually pretty funny, if you were to go to lead a eyes website, what they tout is original and creative content that sounds like it was written by a real person. I'm paraphrasing there. The original and creative part is what strikes out. That's funny to me. Now, again, I think that last article I picked a random was from like the Tennessee and or something, and the other one I read you, like I mentioned, was from the Columbus dispatch. So certainly neither of them are gonna lose any readers over relying on a generative AI engine that uses the same silly sounding sentence over and over again.
Speaker 1:But we need to take a step back and we need to think about this and how it affects us specifically. Again. Managed services, it companies in general, but really any enterprise, any business that's relying on this technology. So people often ask me my opinion on generative AI and the uses of it, and open AI specifically, and some other vendors and again, I'm very high on the technology. I think the promise of it is incredible. In a lot of ways, it's not ready for prime time.
Speaker 1:The biggest thing and we'll talk about this probably on future episodes, the biggest thing is that it's not required to be accurate. So, in other words, if you ask, if you go to open AI or chat, you can ask it to write something. It'll write something that sounds real and sounds realistic but doesn't necessarily mean it's true and I have a bunch of great stories that we'll share about that at some other time. So that's the biggest problem in it. So one of the things I often encourage people to use it for is to create content from something that you know is real. So, in other words, feeding in information about the final score of a high school football game and asking it to generate a notepad article about it. In that scenario, under those constraints, it will generally create an article that is accurate and truthful and, again, it will use the technology, will use the language that it finds appropriate, that thinks it's interesting or that it that is usable.
Speaker 1:In that case, the reason I bring this up and the reason I talk about it is because one of the most common uses that I recommend for things like this is around sales and marketing pieces. So, in other words, you're a small upstart MSP and you need to do a marketing blitz or an email campaign and you want to get a bunch of new clients to sign up for some new managed security plan that you're about to launch, but you just don't have the expertise on writing marketing pieces around managed services or managed security services in this case. Now you can go out and you can hire someone, or you can contract someone, or you can use one of the multitude of great partners there are on the channel to do stuff like that, but you might want to take a crack at it yourself. So you might want to log on to ChatGBT, you might want to describe the offering and you might want to ask it to write an email for you and, truth be told, it will probably do a pretty good job of that. It will write a compelling, interesting, accurate piece for an email for your website, for a marketing blast.
Speaker 1:Whatever you're doing with it based on those criteria, the problem is you don't want the guy down the street that selling and competing, offering to do the same thing for you and have the same language in there. Now again, I'm not saying that in every case this language is going to come out the same, but when you're using the same machine learning model or I should say the same large language model in this case, to generate similar content, chances are those results are going to be pretty close to each other. Now you can put your own slant on it. You can ask to write it like you're a 14-year-old skater or a 25-year-old gamer or a pirate living in the 1800s, like whatever you choose to slant to put on it. You can get it to write in sort of a different voice or a different method, but chances are you're just going to ask it to write as a business professional or not. Ask it all and it will imply that you sorry, it will infer, rather, that you think it should be written as a business professional.
Speaker 1:And that's the danger of a lot of this stuff. And leaving it what we call unshaperone, leaving it alone, is wrought for opportunity for things to go horribly wrong Again. You don't want your website to have the same content. I remember early on when we were starting to sell managed services, there were a bunch of marketing firms that were selling, offering kind of pre-canned articles that you could put on your website, or they were writing content for your website. And this is years and years ago. They've gotten way better. But if you had at the time, if you had asked one of these companies to generate content for your website and then you had Googled it, you would find other managed service providers that had the same exact content on their site. Or maybe you guys were using the same Microsoft articles or the same HP articles or whoever was putting out content for the SMB, for IT services, and we're kind of at risk of that happening again. So the moral of the story really is I think it's kind of fun and interesting and helps you understand sort of how this technology works, where it came from.
Speaker 1:But really the moral of the story is that, no matter what the use case, never let this AI work alone, never let it work in a vacuum, always have it chaperoned by some sort of professional right. So whether you're asking it to write code for you or write a bunch of scripts for you for some sort of automation great, because it will absolutely shortcut the work and it will get done much faster. But you're absolutely out of your mind if you roll that stuff out without having someone that understands it, checking it and rechecking it and running it in a sandbox and doing all the sort of due diligence you should do before that stuff gets deployed. By the same token, don't ask it to write an email, copy and paste that email into MailChimp and send it out. Read it, edit it, make it your own. And if you could start on third base, so to speak, with the amount of content you need written or the content you want, in this case, that's still saving you a tremendous amount of time. Don't assume it's going to do all your work for you, or, quite frankly, don't ask it to do all your work for you, and that, I think, is something to remember.
Speaker 1:Just about AI in general generative AI specifically, but AI in general. We like to say that AI will not replace people. People that use AI will replace people, and this is a perfect example of that at play. Here's this technology play around with it, see what's out there, use it to generate content, but never let it work alone in a vacuum. That would be my takeaway of the day, my big advice, my big lesson learned that, and don't ever draft a starting quarterback from the Dallas Cowboys for your fantasy football team. Two lessons of the day. Thanks again for joining me today. My name is David Tan. This has been the Crush Bank AI for MSPs podcast. Check me out on LinkedIn, check us out at CrushBankcom and enjoy the rest of your week.