Artificial Intelligence Podcast: ChatGPT, Claude, Midjourney and all other AI Tools

AI Versus CTOs with Mark Wormgoor

Jonathan Green : Artificial Intelligence Expert and Author of ChatGPT Profits Episode 324

Welcome to the Artificial Intelligence Podcast with Jonathan Green! In this episode, we explore the evolving role of Chief Technology Officers (CTOs) in the age of AI with our special guest, Mark Wormgoor, a seasoned CTO and expert in technology leadership.

Mark discusses the rapid advancements in technology and how CTOs are at the forefront of integrating these innovations into business strategies. He emphasizes the importance of CTOs balancing their deep technical understanding with the ability to communicate and implement technology solutions that align with business objectives.

Notable Quotes:

  • “Technology is not just evolving in a straight line; it’s truly exponential.” - [Mark Wormgoor]
  • “CTOs need to bridge the gap between deep technical knowledge and business needs.” - [Mark Wormgoor]
  • “Start with the business problem. Don’t just adopt AI for the sake of it.” - [Mark Wormgoor]
  • “Awareness and education are key in managing the security risks that come with AI and technology.” - [Mark Wormgoor]

Mark shares practical advice for CTOs on how to effectively integrate AI into their organizations. He recommends starting with small pilot projects to test the potential impact of AI tools before scaling up. Mark also highlights the importance of having a collaborative approach, involving various department heads in discussions about AI adoption to ensure it addresses real business needs.

Connect with Mark Wormgoor:

Mark offers valuable resources for CTOs looking to enhance their skills and better manage stakeholder expectations. His “CTO Compass” project provides insights and guidance for technology leaders navigating the complex landscape of modern business technology.

Connect with Jonathan Green

Jonathan Green 2024: [00:00:00] The battle between artificial intelligence experts and chief technology officers on today's amazing episode with special guest Mark Warner.

Today's episode is brought to you by the bestseller Chat, GPT Profits. This book is the Missing Instruction Manual to get you up and running with chat g bt in a matter of minutes as a special gift. You can get it absolutely free@artificialintelligencepod.com slash gift, or at the link right below this episode.

Make sure to grab your copy before it goes back up to full price.

Are you tired of dealing with your boss? Do you feel underpaid and underappreciated? If you wanna make it online, fire your boss and start living your retirement dreams now. Then you can come to the right place. Welcome to the Artificial Intelligence Podcast. You will learn how to use artificial intelligence to open new revenue streams and make money while you sleep.

Presented live from a tropical island in the South Pacific by bestselling author Jonathan Green. Now here's your host.

Mark, it's really exciting to have you here. It's great to have you back on the show, and I really wanna get into this core issue because it's becoming a bigger and bigger issue. But what I really wanna get to first is like really just get people excited and explain exactly. What they're gonna learn over this episode, like what's the most important thing and what should they be excited about to learn?

Because you're an absolute expert on Technology Officer Inc. 

Mark Wormgoor: So here's what I believe. Technology has been developing at an incredible rate, right? And it's exponential. So technology isn't going in a straight line. It's truly exponential. And every year technology just doubles and it gets more exciting every year.

We've seen it with AI and it's gonna continue, right? I don't think that's gonna stop. I think CTOs are the people that are really gonna lead the pack and that change. On the one hand, there are people that really have. A very deep understanding of technology. And that's really important. But what a CTO also needs to do is bring that technology to their co-founders, their counterparts and their business.

And I think joining those two worlds, the business side on one side, with that really deep technology understanding, I think that's the crucial part of the CTO world. And I think that's what we're gonna talk about today is bringing those two worlds together, not just the technology side, but really merging that with the business side.

Jonathan Green 2024: I think that's the most important thing is we're seeing a real shift with how people talk about technology and there's this huge gap between usefulness and hype spec, like AI is the big word. And I saw something today where someone's there was a really amazing, someone said, you always hear people saying about now, oh, now AI can write all the code for me, but it's never a code or saying that.

It's always someone who and then, and it's really true. It's like the people who love AI art are not already artists. There are some, 

Mark Wormgoor: but 

Jonathan Green 2024: not everyone, like the majority, it's not. And that's something really interesting, and I think that there's this huge gap, right? There's three areas. There's the area of, I understand how LMS work.

I program lms, I'm a data scientist, machine learning. I build the ai. That's one category. The second category is the. Super hype enthusiast. I love everything. ai, I'm trying every new tool. Every tool I try is great. Everything's amazing. I love this movie, even though I love this video with ai, even though it has nightmare teeth like the right, and to me it's real.

And then there's the category of people who I think is the majority, which is well, which I don't wanna learn. A tool that's no longer useful in three weeks. That's, I think that's the, especially for A CTO, I think this is a core fear, which is I adopt, I say, Hey, everyone, we're switching to ChatGPT a month later.

Chachi BT iss out. It's all about Claude, right? Because if you watch the news cycles, it's really stressful because it feels like you're playing double it. You're making the wrong decision. What if we adopt this technology and then it shifts? Because this is, right now, this is happening a lot with video.

There'll be a new video tool. They'll go, oh, this is the killer of last week's tool. And then. Another tool comes out and then the other one has an update and it's, they're constantly, if you aren't an expert, I had a great conversation with a film expert the other day. He goes, I use all three video tools each as good at different parts of the creation process.

I was like, literally, no one else has ever said that to me. Yeah. It takes an actual filmmaker to understand that they're storyboarding in a story. It's not a short three sentence prompt. For a technology officer or someone who wants to consult with a technology officer, 'cause I think that might be the future, which if someone comes into technology and says, lemme explain all the tools, here's what they do, then you know what your team needs.

I feel like maybe that's the ideal AI consultant. I'll tell you all the tools. I'll say, you don't, this is what this can do. It can't really do that. I'll just tell you the truth instead of the hype. I feel like there's such a gap there. But for a CTO officer who's, oh my gosh, I wanna make the right decision.

But it's so hard with how fast you cannot keep up with every tool. How could you possibly do that? What's the right approach or what's the starting point? I. 

Mark Wormgoor: So I like what you just said. I think there, there's another angle to it. Yes. And it's not just the CTO, right? I think you need to have those discussions with head of sales, ahead of marketing, ahead of supply chain manufacturing, finance, hr, like all the other people that are sitting on that board.

Probably need to play a role in that discussion. So on, on one hand it's the CTO Chief AI officer that needs to have that discussion, right? They need to sit down to their, with their sales guy, and they need to say, these are all the things that are really possible in the world today. And it's an incredible amount of what's possible in sales today in AI or technology.

And it's them really that the head of sales that needs to say, this is probably something we can use. This is really gonna make our sales process better. This is how we improve, and this is a couple of ideas that we can go and try out. In my world, I would hope that A CTO is actually capable of bringing that.

I understand what you're saying, that maybe they will need an external consultant. 'cause if you're doing a 40 hour a week job as in CTO, keeping up with all the ai. AI next. Impossible. But maybe the three of them together, I think it does need that business person. So I think that's the first step.

Just thinking, looking at business processes, looking at the sales process, what is possible, what can we do? And from there on, just come up with a couple of ideas. So these are the top three ideas that we can think of that are really gonna make our sales process different. And start a small pilot. So top three, just start a small pilot, see if it can actually work.

If you can do something and only then decides is it really gonna make more money? Is it gonna help us sell more? Is it gonna be more efficient? And from there, scale up. So I think there's a process to it. Coming up with ideas, testing them out at a small scale and only then scaling up. But it needs involvement from the business person, the business head of sales or marketing or supply chain or someone else.

Jonathan Green 2024: I feel like the nightmare scenario for A CTO is the CEO comes in and says, Hey, I saw an amazing new AI tool and I bought it. Yeah. Figure out how we're gonna use it. And this is the, I think of this as the hype first mentality. I bought the tool, let's figure out how to use it. When I consult with someone or work with someone, I say, what's your biggest problem?

Rather than, here's the tool, let's. Find a way to hammer it into the hole. And I think that this is the problem with AI enthusiasts and a lot of people who've changed their profiles, now they're AI consultant and it's you're really excited about the mechanics, but understanding what most people just want is I want it to just work.

I want it to save me time. I want it to make my employees complain about me less. Like I want less negative reviews on Glassdoor. And I want things to be faster and I don't care how it works. There's this, and I would think about this a lot 'cause people, a lot of people in my space are like teaching prompt engineering and I said, I don't think anyone.

It's first of all, the worst name you've, that's ever been invented for something. It's sounds really horrible and hard. It's like the hardest, they've combined two really hard words. 

Mark Wormgoor: Yeah. And 

Jonathan Green 2024: it's like horrible. No one wants to actually be that. No one wants to learn 500 prompts. It's what I really want is when I wanna send an email, I click send, I write the email, and it sends right the machine inside that functionality.

So if I have an AI tool that, let's say it's gonna help me fill out a form faster. I just want it to work. I don't wanna have to ask the perfect question. Yeah. That's where I feel a lot of people get stuck in the weeds as the mechanics. And it's I just want something that, let's say every time I have a meeting, it generates notes.

I don't wanna have to write the perfect prompt. And if I say it wrong, like I got a fight with chat today. I was told you I was redoing a contract. It gave me the contract as a Word document and then we had to make some addendums. And I said, and then he goes, I'm not allowed to give you Word documents.

I was like, you just did in this conversation. So for someone who's not me, right? Like I was obviously frustrated, right? And I was like, oh, just copy and paste gave you before. And I was like, wait. Then wait, what? And that's a legitimate frustration. 'cause the beauty of AI is that it can do different things, but the negative is that it's inconsistent.

You ask the same question 10 times, you'll get 10 slightly different answers. And sometimes it'll say things like, I can't give you Word documents. I'm not allowed. And it's you definitely did. So I feel like. That's one of the core problems. So how can we bridge that gap where we become like, what, instead of looking, instead of finding a tool hammer and looking for a nail, we say what's a problem we need to solve?

And 'cause I feel like the things that were most exciting are the least useful. Like everyone wants to do AI video. And I would say how many videos did your company publish last year? 

Oh zero. The why? That means it's not solving a problem. 'cause the two biggest time wasters, 'cause I research all the time, the two biggest number one is email.

Yeah. And number two is looking for stuff, finding files on your computer. And this is everyone, right? We all, every one of us in the last week has looked for something and we go, what was that video called? Imagine if you just never, you never lost anything, ever again. Like that's a world I would love to live in.

And. Those are two areas and most people, when they think of email, they think, oh, I want something to write the emails for me. I was like what if you only saw the emails that you needed to reply to? And that way instead of a thousand emails, you saw five emails a day, and then you could actually write the email.

So your person to person wouldn't that's my approach to it. I'm like, so it's like right problem, wrong solution. And then the file finding, everyone's that's boring. I'm like, yeah, but. [00:10:00] That's what Spotlight was supposed to do, and Finder was supposed to do and search my file as was supposed to do.

And they can't do it. 

If you say to your computer find the picture from my anniversary two years ago. No. 'cause the file name is some horrible camera name, right? And if your camera has the wrong time and date on it, which mine often does, like in all VCR, never gonna, you're never gonna find it.

So you're scrolling through, looking at past picture. I do this so often, right? No. So I just. That's my core philosophy. So how can people who, from the two sides who wanna bring AI into other companies and 'cause a lot of, I feel like what's happening a lot is board directors will say, Hey, we need ai.

And the CEO goes what do you mean? They go, we don't know, but we definitely need it. It's it's like uhoh. How do you implement that when someone doesn't know what they want? But you just have to guess. 

Mark Wormgoor: Yeah. So first, I think you have, I think you said shiny object syndrome. I think it's, unfortunately it's not just the CEOs that deal with that, it's CTOs as well.

So the amount of time that I've seen CTOs come by and say, oh, we've got ai. Let's go and implement. Chat, GPT, cloud, copilot, whatever, just because we have to be on ai, not because there's a business benefit, but because it's a shiny tool. I don't think that works. So I still think the focus should be on the business problem.

And I think what you just said where A CEO comes in and says, we have to do ai. I think AI can do a lot of things, and I think the inconsistency you said is true. If it's actually paired with an expert, so I think it can write incredible emails sometimes. Sometimes they're less good, sometimes they're spot on.

You never know. As long as there's an expert that looks at it, I would never trust it to, to send my email for me because there's gonna be some idiot sentences in there. Some are gonna be spot on. They're gonna be amazing. I use it to write article sometimes, and then I redact half of it because.

With all my prompting, I can't get the tone right still. Because it comes up with the strangest words and the strangest sentences sometimes, and sometimes it's fine. So I think AI paired with an expert, we talked about coding. I think it can really help speed up the work of people. I think if A CEO comes in and says, we've gotta do ai, I think the first question is what?

Where in our business is that really gonna make an impact? First top line, so this our revenue. Where are we really gonna make an impact and grow our top line using ai? And those are the things to look at. I think that's the first one. I think we often talk about AI in terms of efficiency. So making things more effective, like increasing the bottom line.

I'm a lesser fan of that because yes, we can be a bit more efficient, but even if we save five or 10% on the bottom line, I mean in a really big company that's gonna make some impact, but often the investment is more that you have to do to actually get there with AI than it's worth. The savings at the bottom line.

So big fan of top line AI impact actually growing our revenue and growing revenue. It can be new business models, right? So say you want to go into a new business, it could be as simple as how do we reduce churn, right? How do we actually make our existing customers happier so they stick around for longer?

Or how do we get more customers? Those are the simple questions to ask. If you look at some of the examples out there one of the companies that, I didn't expect it from Walmart, they actually had their QT two earnings a couple of weeks ago, I think, and they actually said they use AI to update their product catalog to rewrite some of the descriptions, make it more searchable, to make it easier to use.

That really helps them on the front line where they're interfacing with customers and it makes their customers lives easier. So for me, that's really where to focus on, not just say we should do ai. Even A CEO, even if he doesn't understand what AI can do. A CEO should understand their business processes and where it can actually go and start to make an impact.

So I think just taking that question without having a follow up conversation, I think is not the way to go. 

Jonathan Green 2024: I think that's really good, is to. Do that approach because a lot of people, they get excited by elements or, and it is exciting. That's the thing. It is really exciting and it's interesting and one of the things that's sometimes, it's hard to explain.

I've been in business for myself for almost 14 years now, and I see the biggest jumps in my business when I do the most boring of tasks. So the most viable tasks for me are project management. Organization optimization. Looking for mistakes in my flow in my technology. Does a series of emails have a dead end?

Or is there a mistake like that? Those things make the absolute biggest difference. Looking at automations for improvements, the really boring stuff is the most significant, not the exciting. It's not, if I could spend an hour shooting a video with you, it's great, but it'll move the needle more in my business, if I find, oh, this thing is blo and broken and a thousand of my customers haven't been emailed in a month that's gonna make a bigger difference.

And there's. This challenge of, because you're hearing so much information, it starts to become, oh, I think this is what's important, and we, the things that most people are talking about is content creation, image generation. Those are fun. 

I do those with my son to make coloring book pages and to put pictures into Minecraft for him.

Those, to me, are great uses. And I love image generation. But to me, image generation is a replacement for stock photos. So if you're already using stock photos, then it makes sense because image generation is cheaper than my stop photo subscription. So that to me is how I explain it. But. There's this really big temptation to think that every problem can be solved by AI or that you need to add AI into every process.

And there's something you brought up that I think is very interesting, which is that you talk about how much efficiency or how much time it's gonna save, you also have to factor in the learning curve. How long does it take someone to learn this tool and for it to become a process and to go through the pushback of, I, I've always done it this way, now I've gotta learn a new way.

And I always look at, the things that AI are really good, is good at consistently is like data analysis. One of the cool things you can do, you can take a picture of your social media profile and say, what should I post more of? And it can tell you that. And I am really bad at data analysis. Like I'm really bad at looking at spreadsheets.

I can't measure crowds. This is one of my weaknesses. I could be in a room, people like how many people in the room? And I'll go, I have no idea. I was once in a room with 25,000 people at the DC Armory. It's one of the biggest buildings in America. I was like, how many people in the room? Someone was like, I was like, I dunno, 500.

They were like, there's 25,000 people in here. And I was like, I can't do, I can't tell. I just cannot count crowds. It's one of the things I can't do, but you can give a picture to AI and you can count the number of people, but people. Are so busy looking at, oh, it can write an email for me, or it can do this.

I'm like, yeah, but it can also just the data analysis, which is what most people are bad at. Most people, unless that's your specialty, right? They're really bad at, 'cause you go, how much time? That's why I have time trackers, like how much time do you spend this week watching TV or talking each client.

And like one of my clients that I work with in sales, he records all of his sales calls, uploads them to the AI and says, this one made a sale, this one didn't. What? What was different? As opposed to having the AI write a script and then he reads the script and it's this is the magic. And I think that especially more, I've been thinking about this as like technical officers and information officers who have this technical approach.

When you think of it this way, it really makes sense for them. Like instead of thinking of it as the silly stuff, it's wait, this can help people too. Analyze their data, see how they're using their time better. See, the 80 20 rule, like everyone says the 80 20 rule, but you have no I, no one really tracks themselves or really knows what they're wasting their time on, but it's oh, this can actually do that.

So those are the things that me the most excited, even though they're most boring, right? It's it's like the stuff that everyone considers boring. Like it's, so that's what has me excited. So for someone who is a chief technology officer and they're thinking, okay, my. A company wants to implement ai, can they approach it problem first and just say, okay what's a problem we need to solve before we grab a tool?

Does it solve a problem that we have? Is it already a problem? If you already have a problem everyone's showing up late, so that's why they made time cards where you punch in so you, it stops the late showing up. So is it solving a problem? 'cause a lot of it, that's the first level to me.

Okay. If you have a problem, then solve with a tool. Then I look at what's the difference between the tool we're using now and this new version, like how much of an improvement, because often it's like comparing like one tool to doing nothing. Of course it's a huge difference, but there's not a huge, sometimes it's an incremental difference, like when people pop, when they release a new AI model, they go, this is so much better than the competition.

I looked it up one time. I looked at the statistics and there was a new release of Claude that was 0.06% better than chat GBT. I was like, that's immeasurable. That's an, they were like, it's so much better. I was like, that's not what so much means. Like that's invisible to me. It's can you imagine if my wife was like, I think I gained 0.0 0 4, 6 pounds.

I was like, I have no idea what that is. It's like a gram. It's like such a small number when you look, think of it that way. But they get caught up in the hype. When you look at actual numbers, they're like, oh, it's really not that different. All of these things people talk about are such incremental changes that once you have one tool, there's not that much of a difference.

What do you do when you're talking to some CTOs and you're advising CTOs and you're helping them to figure out what to do to navigate these waters? When they're like my company wants to do it. What's the easiest? Like, how do I find the right place to start? That would actually be useful. 

Mark Wormgoor: Yeah. So I think what I said at the beginning, talk to people, right?

I know that's not maybe the most natural go-to for CTOs, but I do advise them. Always go and talk to people. Go talk to everyone else on the board, speak to your peers, speak to their team leads. Figure out what their problems are. So I think before you can even start to think of a solution, you need to understand the business, their core processes, and what their problems are.

So I think that's one angle and it's the angle that I always tell everyone to take. I think the other one you just said it is, it's really good adds data analysis. So the other thing you can go and look at is what are the actual data sets that we have in house as the CTO? You probably know the data landscape.

You know everything that you have in house, what gets recorded what customer data you have, what sales data you have. What can you do with AI on that data to actually come up with the recommendations? If you have all the sales data, what kind of customer is worth more, which brings in more revenue?

Which sales ICPs go better, right? [00:20:00] So what are the people to pre approach? What are the best leads? There's a lot of data analysis you can do that maybe you couldn't do before when you were using standard tools and AI is probably a lot better at. So I think those are the two. Talk to people. And really go and look at the data sets you have and can you extract more value from those.

Jonathan Green 2024: One thing I've noticed is the way people approach technology in the last 20 years has changed. When I was first entering the workforce, I graduated college in 2002, most companies had an intranet, which I had to learn was not the internet. They would have a fixed. Closed ecosystem. So even my friend who worked for very large telecom in the sales department, every day the telecom would download their own website.

So you could visit the website on a computer, but you weren't visiting the real website. So you could see what the prices were, where the promises were. 'cause people would say, oh, I saw this on the website. You have to be able to check that. But they didn't have access to the actual internet. They couldn't exit the building.

And that was a security thing. But now. Every single company that seems to have gone away, or at least I never hear about that anymore. Maybe some companies still do it. And like I was like, oh, that makes sense. You're locking, but now everyone's oh, I need accident, I need to be able to send emails, all of that stuff.

And it was like, there's gotta be a way to do it. But now the issue you have is that when people, for example, add Microsoft 365 to every computer in the building, what you've done is taken every piece of data and now it's going to the internet and coming back. 

Which is a massive, like a massive security issue.

Because every time, even if there's nothing, like you don't have a problem with Microsoft, don't have a problem at your end. You now just by transmitting there's just a possibility. You've now entered the possibility of a man in the middle attack or an interception. Just there's a huge difference between me handing you a note and me emailing it to you, even for in the same building.

Because of going through the internet and coming back, and very few people, and I realize this is more chief security officer area, but. For a chief technology officer who everyone's like, Hey, I got an idea. What if we com? What if we take every single document we have and tr and start transmitting it, right?

What if we take all of our customer data and we've seen it happen with doctors, which is a huge, you have a HIPAA violation. We've seen it happen with lawyers. They get in trouble because there's for different segments, right? There's top secret data. There's like data that's restricted and things.

So how can a CTO approach this with the data security mindset? And this is something I think, 'cause people aren't talking about this very much. 

Maybe 'cause not a lot of things haven't happened yet. And it's I don't wanna be the test case. You don't wanna be the one who would, you don't wanna be the lawyer who gets in trouble because you're the one who submitted fake cases from chat, GBT.

Mark Wormgoor: Yeah, no, I agree. So where do I've been in security for over 20 years, right? So this is really my my path. And I've seen quite a few of the breaches and all the things that you just talked about happen in real life. I think that's the other challenge you have as a CTO. If you're a developer, life is simple.

You develop code, you test it, and that's it. If you grow up and become a CTO suddenly you don't just deal with codes, you deal with databases, you deal with infrastructure, you deal with cybersecurity. And cybersecurity is just a fuel of its own. It's, and it's grown in the last 25 years.

Incredibly complex. It used to be, like you said, air gaps. No access to the internet. Life is simple. Everything's on the internet today. My phone, there's emails on there, there's everything's on my phone. My computer, everything's on there as well. I sent everything across the internet. So I think a couple of things.

First, make sure you have someone in the team that really understands security. I think every CTO should have either be close to a security officer themselves or have someone on their team that really understands security. I think that's first. Second, I think awareness, making sure that you train people what to do and what not to do.

And even then I see technology people just upload stuff into chat, GPT without a contract where I'm like, you know that's gonna get trained right? And you don't know where that's gonna show up next. They may use it on the model for the next release, and your data just gets sent to chat GPT. So even people in technology sometimes forget that the data that they upload.

It gets used by the models unless you are contractually covered, right? Because with most lms if you're having paying contracts, it's okay, and you can ask them to over reuse the data if you're using the free version of them to GPT or the others as well. Your data just goes into the model and gets used for training, and it actually says that in terms of conditions, but most people forget or don't realize, and.

People in technology don't know, people in sales and marketing have no idea. So I think step number one is education and awareness. Making sure people know what to do, what not to do. Second, for those that want to use an LLM like chat, GPT, copilot, cloud, whatever. Make sure you have a contract, it's in place, and you actually explain people how to use it so it doesn't get reused.

But people are insecurity and data in general, people are the weak link. I think the technology, it's made it easier. But I think people really need to be educated and aware of what's possible and what's not possible and what they should do and shouldn't do because they will. A lot of difficult things are too easy to do.

A lot of bad things are easy to do. 

Jonathan Green 2024: I recently was talking to someone who said, oh, I was analyzing my customer data and I accidentally included all of their phone numbers when I uploaded the data. And I was like, that's a really big deal, right? It's 'cause I think about this. I download a spreadsheet of all my customer database if it includes let's say phone numbers, credit card numbers, and I uploaded that, that's a huge security breach for those people, right?

I've just transmitted their. And there was a, it turned out recently that the chat GBT Mac app, right? You can have the desktop app had no security on it. It was storing everything you ever transmitted as simple text files, which is like what? Like the lowest of all securities, like 1980 security. So yeah, you can't without, because we make this assumption that, oh, Microsoft owns part of this company, so it must be super secure.

And it's that's such an assumption and I think that there's these big element that you have to add in a training to explain to people what's secure and what's not. Which data we never wanna upload and which data we is okay to use. It's fine if you remove the last names and you remove a bunch of other data, then you can analyze, oh, which country are my customers from?

But you have to think about what to remove and what you upload. 'cause you can't, there's no undo. It's out there. And then. The other thing is just thinking about, yeah, people wanna be like completely connected. Like a lot of, I don't have any of those. I don't have any voice activated devices. I'm like, why would I wire myself?

I'm not worried about committing a crime. I'm worried about saying something stupid, like saying something dumb. I don't need a record of it. And people are like, oh, they don't listen all the time. And then they found out, oh they do. And it transcribes everything. And you find out, oh it is listening.

Of course it is. Right? And then now people are finding out that your car, like your AI car, will tell on you for speeding. Huh. And all of these things are happening where we're buying devices that are turning against us. So I think there's, this is the other thing that really has me worried is that people are very excited by tools.

And they're so excited. They're not thinking about the implications. The other thing I'm interested in is you mentioned social engineering. I remember when if you wanted to break into a company, you would just leave a bunch of USB sticks in the parking lot and someone would pick one up and bring it inside and plug it in.

And it's that's all you have to do. It's so easy. And now. I think we've seen people put all of their data into a chat bot, like everyone wants the customer service bot, which I don't know why I think it's, I think that's a really poor use case. I think it's the worst AI use case. It's I'm already mad if I'm contacting customer support, right?

Like, why make me mad? Or now they're like, oh, we're gonna have an AI voice on customer support. I'm like, oh, that will definitely solve their problem is now I'm talking to a fake person that doesn't have the power to issue refunds or send me a missing part. Perfect. It's just like, why do you wanna make your customers mad or mad at you?

I don't know. It's like the worst place. But that's what everyone's excited about because we could fire all of our customer support people. It's I feel like those people. There's other areas that I'm a big fan of. Like pre-sales chatbots are fine 'cause people are not mad yet. But if someone calls customer support, they have a problem and they want it fixed.

I've never met anyone who says, oh yeah, and AI chat bott solved my problem. No, never met anyone ever. And I always ask that question, but the danger is you upload all of your data. You can socially engineer chat, GBT, like you can trick it into information. Like one of the things a friend of mine and I do is we take Fortune 500 companies, GPTs, that they put on ChatGPT in the dashboard.

We've broken the code on almost all of 'em. They're all unsecured code, like whatever you think. And it's my friend then who was like breaking into 'em. I was like, oh my gosh, what if he can break into mine? We did. We're doing a live call, but my security measures, no one even had, most of 'em had none. They hadn't even attempted to lock it.

They hadn't put anything in there to stop social engineering. Don't let people trick you. You have to tell 'em like, just so there are these, it's a new vector. The idea of someone through a computer doing social engineering is a new concept. So how do we navigate that world? I. 

Mark Wormgoor: I think it's getting harder and harder.

Social engineering used to be like a phishing email or a fake call from your bank, right? That's used to be advanced. Even six, seven years ago. I've already seen quite. Realistic WhatsApp messages come through which seemed to come from board members. Actually had the proper mobile number, had the proper picture.

It was their mobile number. It was someone else's, it was a random prepaid card, but apart from that, it was like the actual board member, his picture, his name, everything on the WhatsApp was correct. And this is already six, seven years ago. And it's gonna get more advanced. We've seen, we've heard the one story from Hong Kong about this company that actually ended up in a Zoom call with two or three people from their company that were both fake that were actually fake AI videos.

I. I think that's gonna go very quickly. So I think that at some point very quickly we contrast anything anymore that we see. If the videos are getting much and much better. Soon we're gonna be talking to somebody on the phone, on a WhatsApp call, on a Zoom call. We won't know if they are who they are who they say they are.

I think the only thing we can do is awareness. Train people just to be suspicious of everything they see. They read, and it's quite bad if you have to be suspicious. But I am actually now every email I get I double check, [00:30:00] which is quite bad. If there's something important, if it's a friendly email, I don't care, right?

If there's something important for me to do, like an invoice, I check it two, three times before I do anything. And that's quite bad. And I think that's gonna get worse and worse over time. Yeah, and I think the one thing that we should always teach people to do is call back. Something happens.

Yeah. Drop call back, reach out to the person by their email, what, their phone number, and do things like that. But it's a horrible world. It really is. 

Jonathan Green 2024: I think that I heard something from someone who, like another person that I know, and he said, I told my bank if I wanna move money, I come in person.

I don't call because your voice now, anyone could duplicate. I was like, this guy's a really, that's a really clever thing to think about is that. Someone can call and sound like you person at the bank is you get the minimum wage person who's been working there two days. Oh. And it's I, yeah, I get so many fake invoices now and I'm actually, what's interesting is that my a my email service says, oh, we've implemented ai.

And I go, that's interesting 'cause I'm getting a higher rate of like sketchy emails than before. And so I'm like, is not very good. What's going on here? Mic Microsoft. So that's one thing that goes about, but I think that. This is something that I've been thinking about is that I think there's gonna be a shift to face where it's like now, oh you have to bring the invoice in person again.

Like we're, it's like the only way to know it's really you is you have to come in person and, because how? So like with, and the good thing is that I think we're gonna reach that tip with social media where people just wanna spend time together in person because it's like. You don't know what's going on.

So I was like, oh, then maybe there's a good effect. And there's good and bad about, we only do deals in person. We shake your hand. I need to look you in the eye and to touch your face to make sure you don't have a mission possible face on and right. Like it's like it is good and bad because it's like you have to fly to another country sometimes.

And I think there are gonna be some changes, but it's, that's one of the things, I wonder how far it's gonna go if it's gonna lead to that calibration where the only way we trust you is if you're in the same room as you. It has to be a face-to-face meeting. 

Mark Wormgoor: Yeah. I'd love that, but I don't know about banks over there, but banks over here, all the offices are closed.

There's no way you can get to a bank anymore, into a bank office. I think you need to make an appointment. You have to explain to them why before you can even get, and they've closed 80% of offices, right? So everything's online. There's no way to get to bank office anymore. So I think businesses are making it harder and harder to do anything in person.

They're adding more and more checks with every step of their new app and their new process. And I think it's a continuous. War almost between the fishers and the scammers on one side and the official institutions. And it's not just banks. It's all the other business as well on the other side.

How can you prove that an invoice is real? It's really hard. Can you trust an email sender? A lot of them don't even come from the company anymore, just from some random accounting software or somewhere else. It's yeah, it's I'd say it is a war out there and that's not gonna change anytime soon.

I think it's it's probably gonna get worse before it gets better. 

Jonathan Green 2024: Yeah, I think it's really good to mix in the excitement with the challenge because it's so easy to get caught up in the excitement that you leave yourself vulnerable or make a mistake and yeah, when someone says to me like, oh I just don't know.

I say, this is a company. I say, if you want a secure ai, here's what you do. You lock your computers into an intranet. With whatever data you need, you have data can come in, but it can't go out. That's the first thing, right? Yeah. The second thing is you can have a local LM, you just need an expensive computer.

And not that expensive. I'm talking three to $5,000. You can have an AI that your whole team can use, like it's not crazy. Yeah. That's what I have over there. It's a gaming pc 'cause there's only one computer builder in the whole country. So it has a lot of LEDs, but I can run really big models on it.

And if I wanna run the biggest models in the world, I have to up the ram up in the ram to the highest level, $600 to be able to run the equivalent of ChatGPT 4.0. So it's not, 

Mark Wormgoor: yeah, 

Jonathan Green 2024: that When you think about our comfort scare and you say here's how you can do it. The other way is if you want to be completely secure, use a laptop at the end.

Every day you destroy the laptop. You have a, it's not connected. No one wants to do that, right? Because it's to be totally secure, it means you have you're spending $3,000 a day on laptops destroying it every day, or racing the data. But it is possible to be secure. But we're measuring security versus convenience.

And I think that's the challenge we've become, I want it convenient, I want it immediate, and I wanna use this ai it's the AI component in your computer is 0.0067% worse, you'll never know. You won't be able to tell that it's worse. And I think that, I just am I don't know. I think there's po opportunities there because really most your workers at work don't really need access to the internet most of the time.

Most of the time they're in like monster.com or hot jobs.com. That's what I was doing when I was working somewhere. They're mostly right. They're mostly looking at job searching, right? Yeah. And you can have email, have access to the internet. That's it. Like you run it through a port. Even I know how to do that.

It's not that it's a little complicated but it's the idea. I understand. So do you think it's possible to pull back internet access from your employees or to say, this is a work computer, this is a personal computer. These two things are separate. You have two computers on your desk. This is for this task, this is for internet tasks, and you don't connect the two computers ever.

Do you think that's possible? Or because people have become so used to it and they go, you can't take away my right to send WhatsApp messages during work. Do you think we've reached that point? Or is it possible to create a new type of security? 

Mark Wormgoor: I think you're already seeing that more and more.

Just last week I had a discussion with with the CIL about WhatsApp, more computers because they actually had people sending sensitive. Company information through WhatsApp to their coworkers where they shouldn't, but they still do. So the discussion we had is, should we just re remove WhatsApp, the tool, but also WhatsApp web clients should just block it on their PCs.

They can still use it on their phone, but that's like completely segregated on their phone from the company data. I think those are steps that companies are taking to, are taking already today. And I think that's a good first step. Also they're blocking a lot of different websites. So the site, I don't know if that's gonna be the mon support.com, but or the Indeed or the LinkedIn.

But we can already see that there's a lot of websites that are starting to get blocked of work workplaces, and I think that's really fair. A lot of these security companies have lists of, these are all the websites that you shouldn't get to. And of course, they're. The more adult sites, but there's the gambling sites, but there's also all the scam and the protection site that already gets you some security.

It's by no means perfect. Companies that I've worked for actually blocked all file transfer websites. So you couldn't get to a Google Drive, you couldn't get to a, we transfer, you couldn't get to a Dropbox, you couldn't get anywhere because that was out of policy, right? You could only use the, whatever the company wanted to upload files to.

So I think companies are really starting to block a lot of. Those kind of websites are ready to protect their data. And I think the same is gonna happen with lms. You will see that companies start to block chat GPT. If they have a Google contract, they will only allow Gemini. If they have a Microsoft contract and they've contracted copilot, they will only allow that and start to block all the others, at least from the work PCs.

And I think that's a really good first start because you can't stop people from just. Sending something to chat, GPT, uploading, uploading sensitive data. Like you said it's a one second task and you're never gonna be able to delete the data again. It's really hard. Yes. Maybe not the entire internet, but big parts of it.

Yes. 

Jonathan Green 2024: Oh. When I was a kid, my dad could remote into the office in order to do this. The IT department came to our house, did a special setup, and they gave him a physical two fa so it was rotating numbers all the time. So he had a physical thing for him to log in and remote into the office took about 15 minutes.

We had a high speed internet connection, but the process, the security. 

And it was what, and it was like the only way to do it, it's like this is. You have to get the physical device right. If it's on your phone, like two, five is okay, but it's still doable, right? Your phone can be clowned.

So there was like. Really these layers. So we have to, I think you do at least one complicated password, and they're making him change his password all the time. So I was writing it on, he was always writing it on a piece of paper. Of course, but it's like it's not, like at least they would have to, but he could only do it from our house. But now you have a lot of people that wanna work from home. So what, how can you. Now you have this interesting vector, which is, okay, I work remotely. I can do company work. Not allowed to do that on the company laptop, but at home I can go to whatever website I want.

Can you actually secure a device that's leaving your building all the time? Is that possible? 'cause people are always, I. Finding a way around it. 

Mark Wormgoor: So I think with the latest technology, you can go a long way into, actually this is Microsoft technology, but you can secure the laptop that they take, right?

It's fully encrypted. It needs either face recognition or a fingerprint to unlock. But still you need to physical access to the laptop. Which is still protected and it can, yes, its connected to the internet, but still all the websites are still blocked, so you can still block, even if someone is working from home, you can still block all the same websites.

Accessing on another PC is through web browser only. You can't get into anything else. You can't get into any company data. I have that with my it's some of the work that I do. The only way to get to the email is through web browser. I can't use Outlook or any other clients to access the email.

So I think a lot is possible. I think bigger corporates are doing quite well. If you're a small a hundred people company. I think that's hard because it does require serious experts to set that all up and make sure that it works and to keep 

Jonathan Green 2024: it running. I think that there's this mistake people make, which is, we're too small.

So I used to have this belief on my website. I was like, who's gonna, my blog doesn't get that much traffic. And then I put on, so I put on layers of security and I. Made the mistake of leaving on the alerts. So I'm getting an email like every 10 minutes. Attack. I recently a few months ago, I was getting, I kept my login, kept getting locked out on one of my websites because I have really good security on it, and I was like, I had to end up and I have the user for a usernames.

I use a randomly generated. For the username and for the password. 

So the username, a lot of people's username is like admin or whatever. That's not what I do. So it's completely randomly generated two. So now you have to crack two things. And you only get a [00:40:00] certain number of tries and it locks you out for a day.

So I had to actually go back in, in addition to the two FA, then it had to go in and add I to just change the username. 'cause they kept attacking this username over and over again, which somehow. The website had leaked, right? The put the, it accidentally put the long, the code username somewhere, and I got changed it again, and that problem went away.

But you make this mistake of thinking, I'm too small, that we don't have to worry about security yet. And that's the, and I think you're exactly right. It's the. Somewhere between two and a hundred employees is the really scary area. Because we even see like something McDonald just got scammed for a huge amount of money recently through their Instagram account.

We see large companies and small companies and the as the assumption is that, like the two assumptions I think that are dangerous, the first is that, oh, I'm sure the company that made the software secured it really well. 

Mark Wormgoor: And the 

Jonathan Green 2024: second one is that I'm too small. That it's, and it can be, very cost prohibitive or figured out, or should I do this?

And it's those are the two places. Like I get hit all the time. Here's the one I get all the time, which is the, we videotaped you naked from your computer. Send us Bitcoin. And I was like, host it. I told my wife, she goes, yeah, I don't care. Post it. Like whatever. But you have to 

Like you have to, you have that kind of place where you're like. Because also like my webcam, it's a physical camera, right? I turn it off after this call, I'll turn it off. I was like, 

They definitely did. But they're sending really aggressive messages. Like someone once sent me a picture, like a blackmail picture of my wife, okay?

Sitting on the beach next to someone. And I was like that's her cousin. 

Mark Wormgoor: Yeah. 

Jonathan Green 2024: I was like, you're not very good at this. Yeah. And also I showed it, I was like very annoyed because they send it to like my business Facebook page. Instead of my personal account. And I was just like, yeah, I just can't, there's no scenario where I just would care about personal stuff, and maybe that's just me. It's do you wanna po Because I just can't imagine anyone being interested. Naked pictures of me. I'm like, I'm not exci. I'm not a celebrity. I'm not exciting. Very dad. Oh look, I'm a dad. But it's that's the new, these, all these vectors, they're constantly trying them. And it can be if, if you're not security conscious, like when I, when my first jobs was in it and I, my first job was to audit every computer in the building to see what software people had on it.

And I was shocked. And this is Uhhuh. This was back in the day of tape drives. So they had a fixed amount of data that could be on all the computers in the building. 'cause every day at the end of the day, they would make a tape drive and once a month they would send the tape drive off to the data storage.

Right? They were trying to be really, and he was like, how come we're going over the limit? And it was all people were installing dumps like games. It was mostly games. It was like, back then it was mostly games and mostly it was like card games and one person had a really big game. It wasn't anything too crazy.

But we also, this was during the I Love You virus and one guy. He did it twice. He clicked it. Computer fried. You have to reinstall everything, which is what I did. And then he fell for it again. Which is and you don't wanna hurt someone's feelings, but it's come on man. No one loves you.

I got the news. So it is really interesting to see that like the same things have been happening for a long time. It's just now with internet people can access you more and more. But I do hearing, 'cause this is new to me that. Companies are having a new approach to security saying, this is your work laptop.

We'll issue it to you. Here's the things you can do with it. Here's you can't, you have your personal thing. These two things are separate. 'cause what I've seen is a lot of people who are self entrepreneurs work from home. They mix work and pledge on their computer and then they're on a live stream and they click the wrong link and suddenly they're fired or whatever.

They have a problem. And it's like the answer is just to separate the personal and the business. And that's, we've. Turned laptops. It's oh, my laptop is my work and my source of entertainment. So for smaller entrepreneurs, I've saw a lot of people, this has just killed their businesses and they, on a smaller scale, you have to I have a firewall.

Like I would never play a video game on this computer. I have a PlayStation five, three feet away. If that computer breaks, I wouldn't connect the PlayStation five, even one of these monitors. I, in my head I'm like, that's insane. I can't imagine doing that. Because I've created that strong firewall in my head, or like watching a mo, like I'll watch a movie while I'm working, but the thought of just sitting with my wife, let's watch a movie.

Now we have to copy it to a USB key and then plug it into the tv even if the TV's smaller. But a lot of people, and I think this is for smaller people that are listening, you have to start off with this mindset of separating work and play for both being able to do your job. I'm at work right now. I dressed up for you.

I wear a polo shirt. I don't wear a polo shirt except for work. I separate those two things, but having these things in place also means because we're, there's the question of people working from home and lost their productivity. If you. Are at work or not at work. The more you, even if it's you're not going to the office, then it's also gonna help you with that.

But also it's like you work with yourself. A lot of people, they go, I have so much free time and I don't know what to do, and I, oh, I have eight hours now instead of two. I can get, and you don't have that focus thing. 

Mark Wormgoor: Those 

Jonathan Green 2024: are the areas for smaller companies where they get into danger.

And I think these, it's, the cool thing is the solution is the same thing. Work and play. Separate those two things from your life, because I. It's really because it's the same thing as it adds security as well. And it's like not a, and again, it's not a fun topic to talk about until your website gets hacked or until someone's in your LinkedIn and or whatever, and then it becomes a really big problem and think that it's such an important topic that I'm glad we talked about.

It's really nice 'cause you have a great deep expertise. This has been an amazing episode. We went way longer than I was planning because I'm having a good time and I just think this is really great. So guys, if you're thinking about. I wanna be a CTO, I wanna be a consultant. 'cause there's so many wannabe AI consultants that you first have to understand what's the problem?

Businesses need solved. What, how? Like, how does CTOs approach problems? How do they, what type of problems do they have? Because there's certainly things that could be helpful to them. But what the CTO does not want is someone who's threatening their job, which is I'm the AI expert and now I, because I think that's where a lot of people are approaching it from. It's different. If you approach it from, I can supplement you, you don't have to watch the news every day. I'll keep you updated in tech. If there's a security breach in the AI space, I'll tell you that is helpful. I think that can be a really cool space.

I think just with how fast tools are changing. That makes sense. 'cause I spend three hours a day just reading AI news 'cause it's my job because, and it's, there's so many updates, so many things happening. So that I think is. Very useful. So when you're thinking about who you wanna work with, think about that approach, right?

If you wanna work with the company, think, I don't wanna come in and threaten one of the employees and say, no, I'm here to solve one of your problems, because that's what we all want. Here's the problem you have, I will solve it. Your boss says they want ai, Hey, I'm gonna ask the questions and go, okay, what type, what's the problem?

And I'll tell you what those solutions are to choose from. Make it easy. That's really useful. So I think this has been very helpful for people that wanna learn more from you, connect with you more, where's the best place for 'em to find you online and see the amazing things that you're doing?

Mark Wormgoor: Yep. Two things. I think first of all. I've been in the industry for 25 plus years now. I've started to document a lot of that. Nowhere near done, but I think I've made a good start. I call it CTO, compass. If people are interested, reach out to me on LinkedIn or Instagram. Just send me dm compass, call it CTO Compass.

I think the second thing, we have our website, the Maverick CTO the maverick cto.com, where I actually work with CTOs to help them grow their skills and manage their stakeholders better. So 

Jonathan Green 2024: those two. It's amazing guys. Thank you so much for listening everyone. This has been another amazing episode of the Artificial Intelligence Podcast. 

Thanks for listening to today's episode starting with ai. It can be Scary. ChatGPT Profits is not only a bestseller, but also the Missing Instruction Manual to make Mastering Chat, GBTA Breeze bypass the hard stuff and get straight to success with chat g profits. As always, I would love for you to support the show by paying full price on Amazon, but you can get it absolutely free for a limited time@artificialintelligencepod.com slash gift.

Thank you for listening to this week's episode of the Artificial Intelligence Podcast. Make sure to subscribe so you never miss another episode. We'll be back next Monday with more tips and tactics on how to leverage AI to escape that rat race. Head over to artificial intelligence pod.com now to see past episodes.

Leave, review and check out all of our socials.