MSP Dumpster Fire Season 1 Episode 1: Leadership and AI: Starting at the Top

Welcome to the MSP Dumpster Fire series. This is season 1, the impact AI will have on your MSP. Welcome to the show. I'm Robert Chiaffi, your co-host alongside my friend Alex. Wes Spencer.

Hey Robert, how we doing today?

I'm wonderful, thanks. How are you?

I'm super excited to get this series going. This is one you and I have been talking about for quite a while, and, uh, it's good to finally, uh, put digital pen to paper and get this thing moving.

I totally agree. That's all we've been doing is, uh, talking about this when we see each other at conferences. Uh, so it's a definitely a topic that I'm both passionate about and have become I've taken the academic route and tried to really educate myself on this topic. So looking forward to sharing as well as to listening to the audience to see what they've got to say over time.

Yeah, that's awesome. This is a topic that as soon as you said, I think a lot of MSPs are in this place where they don't know what they don't know. And some of them are maybe afraid to raise their hand. I went, oh crap, he's talking about me. And that was when I said, we got to do this show. Like, and for those of you to put this in a little perspective, Rob and I were on some sort of a group working session in Empath or something. I think Wes Spencer was hosting something and you and I hopped on and Rob said, I think there's a lot of folks out there who realize we're, you know, we're 2 years into this AI revolution and maybe we don't know what we don't know. And but now we've been on this ride so long, we're afraid to raise our hand and go, hey, I'm kind of a Luddite and I'm not sure. I'm not sure it's okay to ask questions at this point because we're supposed to be the technology experts. And as somebody who really feels like that Luddite, I went, we should do this. We should talk about this. I'll come out in public and ask all the stupid questions because I'm just a dumb sales guy anyway. I'm totally, totally good with that MO. And, and Rob, you've kind of done some homework on this, right? But so once you introduce yourself and then you can kind of tell us where you come from, your background, and then a little bit about the homework you've been doing on this.

So again, Robert Chiaffi, I'm the CTO and co-founder of NMS. Of an MSP based in Yonkers, New York. I've been running this company for 31 years now. I don't know whether I'm due some congratulations or maybe a slap upside the side of my head to say, what the hell are you doing it for so long? I'm passionate about this industry, love what we do, love what we do for our clients. But like many other MSPs, part of the dark side of this business is this wicked roller coaster ride that we constantly find ourselves on. That sometimes I wish the ride would just stop so I could get off and just, you know, relax and enjoy the view for a few moments. But here we go. Here's, here's one of these wicked roller coaster upside down turns called AI. I think before this was cybersecurity. And for us old dogs that have been around a while, we probably remember 5 or 6 different iterations of these, you know, sharp turns and just like kind of stomach upsetting, nauseating kind of feeling that you get of just not knowing where, you know, just when I figured it all out, suddenly the world has changed again. So that's been my impetus for this conversation, Alex.

Yeah, well, I've been around this, you know, I've been in the MSP space about 20 years. I'm so old, I remember when a Cisco ASA firewall and a Symantec Enterprise antivirus license was enough. And all we had to worry about was whether or not somebody changed the backup tape in the morning. Right. You know, and so cybersecurity, certainly a huge shift for our industry. But I feel like the thing that's common in this industry is that there's always the next huge shift. And so this is the next iteration of what is an MSP. And so I think it's a super interesting conversation. When you said this, I went, man, Robert's going out and studying this. He's going to go learn a little bit about this and get up to speed. And we can just kind of play through the— let's ask some of the questions that maybe MSPs are afraid to ask. So, you know, for those of you who don't know me, Alex Farling, 20-year MSP, 17-year MSP, and 3 years at Lifecycle Insights. Been in this industry quite a while, and I've got the pleasure right now to be working along Wes Spencer and Kyle Christiansen to build out this thing called Empath. So super exciting journey, and the best part of it is I get to do cool stuff like this. So Let's dive in. Let's, let's kind of start and dig into the history of AI and where this all started and where it comes from. Because for most of us, we started talking about it with ChatGPT, what, 18 months ago or something like that?

Just about.

But AI is way older than that, right?

Way, way, way older. And you know what? Just, you know, before we get there, which I'm really— I'm a huge history nut. Like, you could probably tell what television shows that I watch. Certainly not like Housewives of whatever. I'm always on the History Channel or something like that. But my point is that I don't want to mischaracterize, at least for myself, that I'm some sort of AI expert. I'm a student just like everybody else in this space, learning. And I— and the way I felt that I needed to learn was to go back in time. And I started you know, doing what many people do, get on the internet and just start looking things up. Actually, I think I used the traditional Google searching before I got into like ChatGPT. But certainly at some point I pivoted to that. But I wanted to understand where AI came from. And it made me start to think about me being in college in the late '80s, computer science student. I have a Bachelor of Science in Computer Information Science from now I own a university, back then I own a college, and I pride myself on my educational background. Like a lot of the things I learned in college were very useful tools for me throughout my career. And I remember then talking a little bit about AI. I mean, it was like just a really quick kind of conversation here and there. And then when I peeled back the layers of the onion, I realized that AI actually was born in 1956. Arguably maybe even sooner, but 1956, there was a small group of scientists gathered for something called the Dartmouth Summer Research Project on artificial intelligence. And this is what's considered to be the birth of the field of research, right? And there was a, you know, a few famous guys, John McCarthy from Dartmouth College, Marvin Minsky from Harvard, Claude Shannon from Bell Telephone, right? So some people from the the Bell Telephone Laboratories from industry, right? And Nathaniel Rochester from IBM Corporation. And I got a little bit of a letter that I found online about this proposal. And the letter is dated August 31st, 19— I'm sorry, August 31st of 1955. And it says, I'm just going to do a quick, quick excerpt here. We propose that a 2-month, 10-man study of artificial intelligence be carried out during the summer of 1956 at Dartmouth College in Hanover, New Hampshire. The study is to proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it. An attempt will be made to find how to make machines use language, form abstractions and concepts, solve kinds of problems now reserved for humans, and improve themselves. I was blown away when I read this, right? Like, wow.

It's like predicting the future, right? It only took 70 years, but here we are.

Well, I mean, not only did they predict the future, I think they helped shape the future because these words, I mean, as you see, there's tons of tons of technology now starting to formulate at our fingertips that were all right around this history. And so right there, I was like, I got to get deep in on this, right? Like, I mean, I don't— I got a full-time job and I got a million other things going on in my life, but I got to dig into this.

But as an MSP today, like, you have customers asking about this, right? They want to know what is this? What's the impact on their business? I was just at a trade show where I saw a company pop up just around managing the policy and procedure and security nightmare that ChatGPT turns into. Yes, this is a real thing. It's taken over our industry. We've gone a long way between the time when we thought Bell Telephone was a high-tech, you know, institution, right? A technology leader. And today. And so a lot's changed and we should kind of dive into that journey and talk about what's happened in the last 70 years.

Well, and so that is very interesting. So I started to build a timeline, I started to, you know, look around and listen, you know, nothing that I've found required hours or weeks worth of, you know, library time looking in old encyclopedia, and it's none of that stuff, right? A lot of the stuff that I'm talking about were things that I could find readily available, like, you know, in Wikipedia and some other websites. But so I started to build a little bit of a timeline of things that I thought were kind of interesting. There's one phenomenon that happened mainly what's considered between 1970 and the year 2000, where this is considered the AI winter. So as much as we're talking about history, right, I think it's important to understand that a lot of the concepts that were developed at Dartmouth College in that summer of 1956 remained as such because the processing power the computing power required, the storage capabilities required for AI to come to life just weren't there, right? And we needed to wait a long time before Moore's Law, so to speak, caught up and enabled us to have cheap, fast, you know, technology that could run at the processing speeds we needed and to be able to store the amount of data that we needed.

So we almost needed cryptocurrency to happen, right? I mean, that was the thing that pushed processing to GPUs that really quickly showed us, hey, this is how you can do a ton of processing really fast, maybe with less electrical usage than it would be if you did it on traditional processors and things like that. So that we could scale something as big as AI.

So Alex, I have a question for you. Little trivia question. I'm preparing you for GPT-4, right?

Uh-oh, I'm gonna be in trouble.

So no, and it's okay. I just like, you're gonna probably get it wrong. I wouldn't expect you to know this, but the answer kind of surprised me. When was the first driverless car? You know, when did the first driverless car get invented and kind of hit the road?

Can I ask ChatGPT if it knows?

You can.

Let's say it was, I'm gonna guess 1998.

Well, you're not terribly far off. And if you're a Met fan, you'll remember 1986. So I was dumbfounded by this, right? There was the first driverless car and I forget forget where it was, but I know it was at some university. And the thing looked ugly as sin, right? Had all kinds of contraptions all over it, big bulky cameras, you know, not like what we consider today. But in 1986, right, there's probably a lot of people listening to this that weren't even born in that year.

My first car was an '87 Accord station wagon.

So, and it wasn't all red with a blue back door.

No, I bought it for $800.

So, well, you were, you were the driver. It wasn't driverless.

It might as well have been, right?

So let's fast forward now a little bit because then, you know, and I looked through like all these different timeline things and I plucked out the things at least that were interesting to me. And then if everybody remembers IBM Deep Blue versus Garry Kasparov, right? The, the, the, and it beat him, right? What year was that? Any idea?

Was that In the 2000s?

No, it was earlier.

Was it really? Okay.

And that's the thing that blows my mind. It was earlier.

Yeah, time flies so fast.

It was 1997. But again, I mean, Deep Blue was running on an IBM mainframe, and you know, the, the size of those, uh, things were, you know, they were behemoths, right? They were—

it was the size of a server room.

Yeah, well, it was the server room. It was a server building, right? You know, it was a whole data center, uh, these things occupied. Maybe it wasn't quite that large, but, you know, tractor-trailer-sized or larger, right? Now, you know, I can't beat— I can't play chess on expert mode. I gotta play it at best, I can play it maybe intermediate mode. But yeah, I get destroyed. And that thing is on my phone, right? On my smartphone.

Yeah, right.

Ridiculous. Now let's get a little fun and get into the home. What about a technology like Roomba, right? If you think about it, Roomba's got artificial intelligence because it's got some programming in it and it learns based upon the room, right? By bouncing off things and making mistakes. But what year was Roomba released?

2005.

You're pretty darn close. It was 2000.

I owned one in like, it was what?

It was 2002, but you said you owned one.

Okay, well, I owned one in my bachelor pad in like 2008. So, you know, I knew they had been around for a little while by then, but as a bachelor, I didn't want to vacuum. I was happy to give that to automation.

Today it's 2024. So it's 22 years ago. Like that thing's been around a long time. Right. And then IBM Watson versus Jeopardy Champs was 2011. I'm just kind of speeding up now here. And then AlphaGo was this program versus Larry Sedol. Larry Sedol, or I don't know if I got his last name pronounced, not Larry, I'm saying, sorry, Lee Sedol was like the world champion. And when AlphaGo beat him in 2016, the observers were dumbfounded about how creative the solution was when he lost to AlphaGo. And just in its very creative style of play, it wasn't just a programmatic, like all out out, calculate you. It did some very, very creative things, right? And now we get into ChatGPT, which actually landed— ChatGPT-3 released in 2020. But where were you and I on AI and the rest of the world in that year? Sleeping, right? Because we didn't know what AI was. Granted, we had Google Assistant and Alexa and Siri. Those are all AI technologies, but like, was AI really a household name then? No, it wasn't.

I mean, in 2020, I was still shutting all those off because I considered them to be marketing data collection machines, not useful AI.

Right.

You know, there was a way different thought around those and spin on those in 2020 than there absolutely is today, where we look at them and say it can work alongside you and make you better. And I actually sat on my back porch about 10 days ago talking to a kid who is a, he's a software developer for a medical company. And he says, I have a, a well-trained GPT assistant that helps me write code. And he says, I still have to know what I'm doing. It writes 90% of my code, right? And I go back and clean it up and clarify it and tell it what I really meant. But it writes 90% of my code. He says, that's not going to replace me. It can't work without me.

Alex, right there is a whole episode for us to talk about, about how it's enhancing our lives, not, not taking things away. You know, I heard somebody say the other day that everybody thought the computer was going to take jobs away. Right. And what did it create? Right. It created our industry. So for those of us that feel in this industry, feel like AI is going to take our jobs away. No, it's not. Right. It's going to create more jobs. Now, will it take some jobs away? Certainly. But those are probably—

Typically it's going to replace jobs, right? It's going to shift what people do. We're going to upskill people and train them. I think about all the new jobs that exist because of it. Prompt engineer. Yes. That's— if you had said in 1995 or even in 2020, a prompt engineer is going to be a job, we'd have laughed you out of the room for saying that.

Totally. Totally.

Today, it's a real thing.

So I was— So, you know, thank you for allowing me to share the little bit of research and the very quick, brief history of the last 70 years of AI technology. I felt foundationally, for me, the way my brain works is I've got to understand where we've come from in order to understand where we are and how we may be moving forward, right? And it helped me root or get grounded in the fundamentals of what AI is both capable of, proven, as well as what trajectory we may be going on. So I think it's really important that we discuss this. And so thank you for allowing me to do that.

Yeah, I think, no, it's amazing. I think where we come from is as valuable as where we're going. At least where we come from, we know for sure. Where we're going, we're still trying to figure that out. And we'll discuss a lot of that in future episodes. As we kind of wrap up this episode, do we want to give everybody kind of a sneak peek of maybe where we're headed and some of the things we're going to talk about in the rest of the series so they know why they should come back for the next one?

I totally concur, Alex. I think there's, there's, first of all, there is so much to talk about this topic of AI. There are, it's like a spider. It's got legs going in a lot of different directions. But sure, I mean, so one of the things I think we're going to talk about is what AI is and what it is not. Right? That's certainly a thing. And how do we get education and training on AI is another topic.

Yeah.

How about you, Alex? What are you thinking?

Well, I think, you know, there's a lot of different kinds of AI and that, you know, LLM is kind of the thing that has everybody's attention today because it's the thing that we can interact with and we touch the most. But there's a lot of other kinds of AI, you know, deep learning and NLP and all that. We'll get into all of those and kind of just set a rough framework. We're not going to dive into any of them, make anybody an expert on this podcast, but we'll give you an idea of the different types of AI that are out there.

Right.

We'll help kind of talk through, and this will get a little bit humorous, but we'll talk about some of the confusion around AI, the confusion of AI, how we get it to hallucinate and give us false data. We'll talk about some of the downsides of trusting AI with a little too much and maybe a couple of example blunders that have happened with AI. As we go. We'll talk, you know, we were talking about my buddy that's doing some programming and stuff with AI, but we'll talk about jobs in AI and kind of what those opportunities might look like for a young engineer who's just getting started at an MSP. You know, there may be a day 5, 7 years from now where an MSP employs an AI expert who that's all they do. They spend their day in the AI side of an MSP. I would totally agree with that.

Yes. Yep.

Yeah.

You know, Alex, also when you were talking earlier, just, you know, a minute or two ago, you made me think that, you know, some of the other things that we'll insert in some of those conversations is even just establishing a vocabulary, right? We need to understand when these terms are thrown around, not just we understand what the acronyms mean, but what do they really mean, right?

We'll have to publish the AI vocabulary. I think that would be a good takeaway document for everybody to have.

So terms to know, I think, or learning to speak the language, right? It's like walking into a foreign country and not knowing how to speak the language. If you know how to speak it, you can navigate things a lot easier. So definitely, I think we're going to be spending some good time on that, especially as we're talking to our customers, right?

Let's figure out how to get this vocabulary out of the nerd words and into our customers' everyday language and make sure that when we explain it to them, we use the words right so that they understand what the words mean.

Yep. And listen, you know, you made me also think about how do we, how do we apply our change management skill sets here, right? There's a lot of fear about AI, a lot of phobias, right? So how do we overcome them for ourselves within our industry? But also, as you mentioned, as we help navigate clients through this, how do we help them get over the many fears and phobias about the negative side of what AI might potentially represent.

Well, and that kind of plays into the fact that, right, every tool now has some kind of AI in it. We can't just flip a switch and block it and turn it off. Our customers are going to be exposed to AI, so we need to join them on that journey. They're probably going to hand some pretty sensitive information into AI in one shape or another as they're just doing their daily job and interacting with AI in it. Um, we want to make sure they understand and that we're setting MSPs up to understand What are the risks there? What are the privacy and confidentiality problems that might come up? You know, how do we write policies and deal with those kind of things? And we got a great guest we can bring on to talk through some of that. You know, we'll go see if we can strong-arm Brian Blakely into coming in and talking to us about some of that kind of stuff as he's really focused on it today. So I'm just name dropping there just a little as we get some smart people in here to join us.

But I don't think we need to convince him. I think he'll be willing, very willing and happy participant.

So. Yeah, he's so good, so good.

You know, and one other thought that I had that we'll likely be covering is around AI leadership, not just, you know, managing change or being in that reactive mode, right? Like, let's think like what the way a normal MSP thinks about technology in general. We, although part of our jobs is reactive, we want to be strategic about things. And so therefore, how do we be strategic about AI? How do we integrate that at our leadership levels? And then how do we as leaders of our organizations help drive that conversation along with clients? Now, I know we kind of touched on that already, but I think, you know, this is something that has to happen at a very, very, very high level first, right? Just like most things, it's got to start at the top before it can permeate down.

Yeah, yeah, I think we have to educate leadership teams. We have to educate boards of directors. You know, this is really where this has got to be a top-down approach because if it's a bottom-up approach, it'll be a real leaky sieve and the data will be out there in every AI environment you can possibly imagine.

Totally.

So, and then data breaches will happen and the typical, the typical data problems will happen, right? If we, if we don't address this.

So I hope we've whet enough appetites for people to keep tuning in as we unpack each one of these topics and go deeper and wider and bring in some other subject matter experts. I mean, or I should say we should bring in subject matter experts Alex, you know, I know you and I are still learning here, but I think the value that we're bringing to the table is at least we can help facilitate the conversation. So I hope that's what the audience is taking out of this first episode as we prepare to take this plunge into the cold water together.

Absolutely. I love it. Hopefully we've managed to sort out a little bit of that AI dumpster fire for folks and at least let them know why they should be be back for more. So I think that wraps it up. That's an amazing episode. Um, come back next week and we'll, we'll have another one drop for you. Thanks much, everybody.

Thank you.

This podcast was created by me, Alex Farling, and my good friend Robert Jaffe. This podcast is produced by Francis Bellin. The music and artwork for this series were produced by artificial intelligence and edited by Francis Bellin.