Speaking on AI

Podcast transcript

Learning from Legends: Peter Switzer Discusses Artificial Intelligence and its impact on business, stocks, and our lives

Complexica's CEO, Matt Michalewicz, joins Peter Switzer on the Learning from Legends podcast to discuss the huge impact of Artificial Intelligence on business, stocks, and our lives.

[00:00:05.580] - PS

Hello and welcome to The Learning from Legends Show with me, Peter Switzer. Thanks for joining us. And today I get to learn about something that is going to be huge with a capital H in business, in our life and the stock market. And it's called Artificial Intelligence. So, hey, I don't get bored. This is going to be a really interesting and critically important interview. I've become more interested in Artificial Intelligence because I've bought into the local stock Appen. Now, it was $43 at its peak. It's fallen down to about $10, I think. Well, I bought into it and I just thought this is a company of the future. And I realised when I started investigating, it's a company working in the Artificial Intelligence space. And so I thought I'd talk to a guy I know very well who is an AI expert when it comes to business and artificial intelligence generally. His father was a professor of Artificial Intelligence in the US and his name is Matthew Michalewicz. And he makes AI very, very interesting. So if you want to be rich, I think it's a pretty good thing to desire in the absence of other better things. So if you decide to be rich in the future, listen to my chat with Matthew Michalewicz. Matt, thanks for coming on the program.

[00:01:32.340] - MM

My absolute pleasure, Peter. Thanks for having me.

[00:01:36.340] - PS

Now, we talked some months ago and we talked about, you know, we positioned you with your book Life in Half a Second for those people who never come across Matthew Michalewicz, unlucky as they are. Tell us about the book, just in a nutshell.

[00:01:50.520] - MM

I've been an entrepreneur all my life and I've collected scientific studies around success and Life in half a Second is a summary of the five primary drivers of success that are proven through research that you can point to and read. And it's a summation of things that people can do, whether in business, sports or their personal life, to improve the probability of success in their endeavors.

[00:02:14.700] - PS

Yeah, and those five things in a nutshell, again, are?

[00:02:18.060] - MM

The first is to have a clear goal, something that you can work towards. The second is to align your goal, to desire to your intrinsic motivations and ambitions. The third is to have belief in the goals. So there's plenty of research to suggest we don't achieve things we don't believe that we can achieve. The fourth is knowledge. Knowledge is a deterrent for many people. They don't know how to achieve their goals and they stumble along the way. And the last is action to take systematic daily action towards the goals that you've defined.

[00:02:50.610] - PS

One of my favourite quotes comes from Muhammad Ali, and I'm sure you come across it, it goes: The repetition of affirmations leads to self-belief, when that belief becomes a deep conviction, things begin to happen. That's exactly what I thought about when I thought about those five factors.

[00:03:08.160] - MM

Yeah, that's that's a great quote. Very, very good.

[00:03:11.730] - PS

All right. So apart from the fact that you came to Australia, listed a tech company on the Nasdaq, kind of catapult you into fame on the speaking circuit in Australia. And some people, when they hear people trying to help people get successful think: Oh, yeah, yeah. But as you point out, you took a scientific approach to trying to find those factors, not only to help yourself, but help other people as well.

[00:03:43.920] - PS

And people loved the book and people loved the stuff you talk about. But deep down there is something very - I don't know - should I use the word scientific, you might catch me out on that, but you certainly are a technical kind of guy, and your latest book is called The Rise of Artificial Intelligence. All right. Tell us why you wrote the book, first up.

[00:04:11.850] - MM

Yeah, so I've been in artificial intelligence all my life because my father has been a scientist all his working life. So since I was six years old, I spent my childhood in universities listening to lectures on neural networks and machine learning and Turing test and so on. So having been around that all my life, I guess was inevitable or probable that I would go into business in that direction. And for the last twenty two years, I've been an entrepreneur running software companies that use artificial intelligence to improve decision making.

[00:04:44.820] - MM

And this book is really how the field looks today, what AI is and isn't, where it's going, and most importantly, what it means for business managers. The "so what?" question. There's a lot of hype and talk around Artificial Intelligence about really what does it mean for business people and what's the relevance and why they should be paying attention?

[00:05:06.240] - PS

Yeah, and I think when people hear words like artificial intelligence, as I said, that's stuff for scientific type people, but not for normal people. It's for big business, not medium or small business. And I guess I want to flush all this out with you because, you know, you're not a scientist really are you? You had more like a business background at university. But course, as you say, probably your parents even raised you on artificial intelligence, you know, and that's why you turn out to be such a perfect human being.

[00:05:47.340] - MM

I'd have to record this and put this on my LinkedIn profile as that is the ultimate endorsement, Peter. But I think that the upbringing that I had led me to think in an evidence-based way. So you in that last book, you know, Life and Half a Second around success, it's an approach from an evidence-based scientific point of view. So even though I'm not a scientist, I have a degree in corporate finance. I think about things from a data perspective, from an evidence perspective, rather than just the hype or misconception.

[00:06:17.760] - PS

Hmm. Okay. I want to talk about. The business that you've created, which in the not too distant future will be floating, correct?

[00:06:27.810] - MM

Correct.

[00:06:28.330] - PS

OK, so we'll end up on that, so for all those people out there who love to make money or keep them dramatically on a string waiting for the name of the company and was going to do and little list and all that stuff. But it does actually. Well, I wonder beforehand is, you know, establishing your credentials around artificial intelligence and anything I know about artificial intelligence has been pretty well inspired by our conversations over the years.

[00:06:58.620] - PS

How long is artificial intelligence actually been around and being used by governments and big business?

[00:07:07.420] - MM

It's actually a very old field, so more than 70 years old, so it predates most of us and it is an attempt to replicate really the human body, the human experience. And the field is broken into four areas and each area has had application areas for decades, uses in business and in government. The first area is robotics, the attempt to replicate mechanical movement. The second area is computer vision, which is the attempt to replicate the visual system of the human body, image recognition and so on. The third is communication or natural language processing. Siri, Alexa, Google Home. That's a big branch of A.I. And this cognitive systems or cognitive computing, which is an attempt to replicate certain brain functions like inference, deduction, decision making and so on. So if you think that these four very big fields, you probably have an iPhone, Peter, you use Siri. I know many listeners will have engaged with natural language processing devices. If you go on Facebook or any of the big social media sites that we use, vision recognition algorithms to recognise people's faces or even the iPhone does that, robotics has been around for decades, and cognitive algorithms have also sat behind the scenes since the 80s and systems and in government, if not the 70s. So what's changed is it's really become popular recently, but it's been around for a long, long time.

[00:08:35.850] - PS

It's funny, you know, when you were in the USA as a young man and I was in Australia as a young man. This might be a stretch because there are a few years difference between you and me, but there was a television show called Lost in Space, and the robot there used to talk to the bad scientists who had been lost in space. It was Gilligan's Island basically in space. Yes. People used to watch that or have forgotten that, the father in that as well, Marvin Robinson's as well. I think about the father was actually the guy who played Zorro in the great television series of the same name, which always was a credibility problem for me. I love Zorro. How he ended up as a father in the Robinsons on Lost in Space was beyond me. I couldn't take it, but that robot was talking to the and the robot could really have a conversation with the bad professor who was, you know, always plotting against the family, that was artificial intelligence, I guess.

[00:09:44.420] - MM

Yeah, yeah, it's that's the ultimate goal of artificial intelligence, is to really replicate a structure that you can interact with, that you wouldn't be able to tell that it's not a human. That's even the essence of the Turing test, that you're engaging in dialog with a computer and you can't tell the difference between that dialog and conversation and a real human being. This is the ultimate goal. And I think we're some ways from that.

[00:10:09.470] - MM

Yeah. So so let's start with that then. And I kind of presume that that's just like the entertaining aspect of artificial intelligence. But there would have been I guess the CIA would have had an interest in artificial intelligence. I would have thought, you know, the big corporations of the world trying to work out everything they can to maximise their success in a competitive world, they would have been using AI for a long time. Is that true?

[00:10:39.870] - MM

Absolutely. And I think there's a point that comes from what you've just said, usually advanced technology, things like artificial intelligence and the like emerge in military divisions of government. And there's a number of reasons for that. The military arguably has the most advanced technology in the world today. They print their own money. It's classified. We won't even know about it until sometime in the future. Once it moves into the mainstream, into business, there's usually a very natural progression of how this technology filters into the general market.

[00:11:15.620] - MM

It first starts with at the very big end of town with very big companies and it starts with financial services. First, they print their own money as well, and from banking insurance, it moves into telcos, big retailers and so forth, and it keeps moving through these industries into it finally ends up in distribution, wholesaling, manufacturing, and it moves across the industries and the down in size at the same time. That's how technology moves or is adopted in the business marketplace.

[00:11:46.310] - MM

What's changed over the decades past is the speed of adoption is faster than it is before technology: moves through the markets faster. But the curve and the way it moves has remained unchanged. And the reason it makes perfect sense all the money is in government, then it's in financial services. So it's progressing where the resources are. And companies that are big have more resources than companies that are small. So it's a natural progression of how technology unfolds in the marketplace.

[00:12:16.080] - PS

I will come back to that because I think that's going to be a really important part of this interview is to understand how our Artificial Intelligence and all the robotics and all the things that you talked about will actually filter down to medium and smaller-sized businesses. And how long you think it might take longer. Before I go back to that, I just want as I was listening to you, it reminded me of a course I did at University of New South Wales many years ago.

[00:12:42.320] - PS

And it was a great course. It was actually about the role of science in economy and economic history. And the point that the lecturer who was a guy called Ian Inkster, who was a guy who came up with the idea of the multifunction polis in South Australia, which never really got off the ground, but was a great idea. But I'm sure you would have come across the multifunction polis suggestion or innovation that was suggested. But Ian made the point that so many of these great earthshattering scientific and technological creations actually started in sci fi, and a lot of sci fi writers are, in fact probably second-rate scientists, but were so preoccupied with what is possible. They then wrote about it and then real-life scientists made it happen. Is that pretty well true?

[00:13:40.310] - MM

I think so, absolutely. And one of the best examples of that is Star Trek. So if you look at Gene Roddenberry and how he wrote Star Trek and the things that were in it in the 60s, anti-matter used for propulsion, whether it was neutrino particles, whether it was even the mobile phones, you know, they were on another planet and they could talk to one another with the tricorder where they scanned the body and said, what was on - all of these things are scientifically possible, and even from a physics point of view, they're possible: wormholes and so on. So I really enjoyed science fiction that has been based on the edge of science rather than completely fabricated realities and physics fabrications as well. So we in many ways, the world has caught up to what Star Trek first displayed in the 60s and 70s.

[00:14:30.110] - PS

Okay, let's go back to that issue that AI and all its associated offerings will eventually work right through all sizes of business and probably a good way of explaining that. I think people can get their heads around why big businesses might be using artificial intelligence and well, my favourite one for my area of expertise I know I share the book with The Principles by Ray Dalio. Now Ray is one of the biggest and best hedge fund managers in the world who was always able to use his brain, his fantastically smart team of his academic-type people to try and work out what's going to happen to a market, and he just, quite frankly, says brains just can't compete with what you can get out of mass computers to handle big data to come up with all the possibilities I need to consider when I'm placing billions of dollars of bets on gold or or ETFs or whatever it might be, he said he does a better job because of artificial intelligence.

[00:15:35.790] - MM

Yeah, I agree.

[00:15:37.770] - PS

Okay, so you understand that. But you first explain to me how this can be applicable to lots of businesses when you start to describe to me what your new business was, your ultimate goal is to float - how it started, what was the first customer? I think when the first customers you talked about, I think was it PFD, one of the suppliers of food? Can you talk us through that story? When you first started this business and what you tried to do -it just seems so logical that businesses would like it.

[00:16:13.580] - MM

Yeah, the general philosophy we have and there's two points that you've made, no, I will address both of them. I always view artificial intelligence and I make an analogy to any technology that it will be commoditised over time. It will become more accessible, will filter down in the marketplace, and sooner or later everyone will have it. And the analogy for that is to look at the calculator. When the calculator came out, whenever that was 60 years ago, 70 years ago, it was prohibitively expensive: $10,000-$20,000 for a calculator. And if you had a calculator, you had a competitive advantage. You could do calculations faster. Then everyone had a calculator and then the computer came about and the computer was prohibitively expensive. But if you had one, you had a competitive and so on and so on, spreadsheets etc. AI is the latest calculator, the latest computer, the latest spreadsheet. It can calculate, just like your comment about Ray Dalio from Principles, it can make those calculations faster, look at the probabilities, analyse scenarios, recommend courses of action better than anything before. And today it's still expensive, not as expensive as it was 20 or 30 years ago, but it's still expensive. And hence, those that buy it usually come out with a competitive advantage versus those that can't afford it at the moment. But over time, it would be like a calculator. Everyone will have it and and it will go into every area of business and in many ways, many areas of our life as well, because when you are able to calculate the probabilities and different scenarios and different outcomes and recommend courses of action, what that leads to is improved decision making. And if you think about business or even life, the quality of our future is often defined by the quality of decisions we make for a certain right way. Whether their investment decisions, which you're an expert on, whether career decisions, relationship decisions, these decisions we make impact the quality we experience in the future.

[00:18:09.380] - MM

And so in business, the same principle applies. The quality of a business's future depends on the quality of decisions that that business is making today that its employees are making today. And the world has become so complex and so noisy and dynamic and fast, for lack of a better word, that you really need some kind of more sophisticated technology to interpret all of this and help these companies and people make better decisions. And the one of the first customers of the company, which you will reveal, I almost said the name, but now I will give them the name, the name of the company is Complexica. And our first or one of the first customers was a business called PFD, which is about a two billion dollar company in Australia that distributes food products to cafes

[00:19:00.160] - PS

We see their trucks everywhere - PFD

[00:19:02.330] - MM

Everywhere, a white truck with the red logo. It's one of my favourite companies, but think about all the decisions that they need to make every day, how much stock to order, what kind of products to offer to customers, how to price those products, where to send their salespeople, website decisions when people come to the website to order what should be offered, how they should be bundled. Continuously, thousands of daily decisions. And if you could just improve these decisions every day, you can really improve the outcomes that these businesses achieve. So I'm a big believer that in the future everyone will have it and AI won't be the exciting technology is that it is today, it will be a boring technology, there will be something new and it will pervade every element of our life and business as well.

[00:19:47.510] - PS

Yeah, OK, so when you when you started work with PFD, when you first pitched to them what your artificial intelligence program could do, so you've been talking to a normal person, wouldn't you - it could have been the CEO or CFO or could be someone's making big decisions, it could have been just the guy who had to make a decision about ordering on whatever. How did you pitch it to him or her so they could get the potential of what you're offering?

[00:21:03.590] - MM

I think the pitch is quite simple in whether it's you or any company, you usually can identify a commercial opportunity that they're not realising at the moment that might be improved margin, improved revenue, revenue growth market, whatever that tangible business KPI is. And then the next thing that you can say is why aren't we capturing this improvement in our performance at the moment? And the reason might be, is that, God, you know, we have 100,000 customers, we have 100,000 products, we have millions of pricing files. We have 70 warehouses, thousands of trucks, all of these moving pieces, and this is really no brain in our company capable of computing all of this in real time to inform us what these decisions should be to achieve the commercial outcome and then the final part of the picture. Well, here's an engine that's capable of making those calculations 24/7, 365, so that you can capture the commercial value that isn't being realised at the moment. So I think I think all of these conversations should start with the business outcome that you're achieving. Then talk about the business problem that is preventing the outcome. And lastly, the solution that's going to solve the problem.

[00:22:17.960] - PS

OK, so listening to that, because I'm a person who does think in pictures, I'm thinking of it like a really big machine with a spout at the top. And so I can just pour all this information about this business from everything from trucks to drivers to stuff to buy the stuff, to sell and throw it all in. And when you put in there, crank it up. And then I always this is really a complicated machine to have until I can type a request and say, how can I make sure I'm getting the best of all my trucks? And by posing that question, are you saying that potentially an AI program could possibly answer that question for them?

[00:23:10.370] - MM

Yes. So the analogy that you've given in terms of a big machine and you crank that, that's a nice way of describing Amazon Web services, the AWS environment. It sits in the cloud. There's lots of heavy duty servers that it's using for computational power. It's plugged into a variety of data sources. The data that comes from a company, external data, weather might impact what happens tomorrow. It might be a census data or whatever the case might be.

[00:23:39.950] - MM

And it's constantly calculating. And instead of asking it, it just tells the people that are making these decisions what the recommended course of action is. They think that, you know what analogy I love using. Have you ever seen in whether it's in politics or in really big business, how these politicians who were executives, they have an assistant and the assistant is almost like a data analyst. They're like analysing everything that's happening and giving them suggestions for the meeting and so forth. Imagine a digital analyst like that. Imagine everyone in a company having their own analysts, their own assistant that's populating and evaluating everything that's happening and then feeding them recommendations when they need it. I think that's future and it makes perfect sense. And there's no technological hindrance in achieving that. Wouldn't it be great on all of our phones, Peter? We all have a digital assistant, digital analyst that really analyses everything that's happening and gives us recommendations.

[00:24:37.590] - PS

This is funny as I was listening to you I was thinking about when I was a young man in the newspapers. I used to have a Mr. Magic computer selection for racehorses. And I was also I've interviewed Tom Waterhouse recently, talked about how his father was. And they actually have a tipping service now. Now, in the old days, his father, Robbie Waterhouse, famous Waterhouse, we've done all that calculations in his head, he would have written things down, you know, like the form of the horse, whether guys can win on the mud, the quality of the jockey. And he would try in his head to put it together and rank the horses from top to bottom. Who's likely to win, but gee! A computer could do that much better than any individual could even come close to because there's so many variables out there.

[00:25:31.190] - MM

Well, yeah, I couldn't agree more. You need to look at the famous movie Rainman with Dustin Hoffman and you know, it's some people through abnormalities in the brain can make these incredible calculations that are very difficult, like, you know, the counting cards and so forth. But most human brains can't and that's why we use spreadsheets and calculators and different computational devices to aid us.

[00:25:58.850] - PS

Isn't it ridiculous that in a pathetic old age, those sorts of people were called idiot savants and of course, you know, the much better world we live in and just call them savants. But in those days, if someone was not what you might call normal, but had this unbelievable skill, they were then tagged idiot savants. And it's like, what kind of world do we live in? But that's another side issue. All right. So, sorry about my observations, but I always like to know that mankind has actually moved ahead. Let's let's now look at Complexica. So it's going to be a listed company when?

[00:26:39.930] - MM

Next year, mid next year is our target.

[00:26:41.640] - PS

Okay. And when you go to market. What are you, going to tell the market Complexica does?

[00:26:49.820] - MM

It improves decision making within companies in really three key areas: sales decisions, marketing decisions and supply chain decisions. And in our experience, these are the areas where complexity resides and better decisions create commercial outcomes in these areas.

[00:27:05.970] - PS

OK, so at the moment, most of your customers would be big to big, medium sized companies, I presume?

[00:27:14.780] - MM

Yeah, the smallest customer in revenue we have is just under 100 million in annual revenue all the way to businesses like Metcash that are in tens of billions or the Pfizers of the world and so on. So it's a big range. But what they all have in common is some level of complexity within one or all of these areas that needs to be addressed with sophisticated algorithms, a.k.a. Artificial Intelligence, to achieve a commercial outcome.

[00:27:44.500] - PS

Hmm. So how long do you think before Complexica is also dealing with medium and small or medium and then ultimately small business? Do you see that eventually happening over time?

[00:27:58.550] - MM

Yeah, definitely. I think it's inevitable that will happen. And it won't just be Complexica. I think it will be lots of companies that provide software and technology to the middle markets in the small end of town. How long - those kind of predictions are difficult, but I think it will be another 10 years before the really small end of town, you know, companies that are turning over a million dollars, half a million dollars can really benefit in full from this kind of technology. And it's not like there'll be nothing and then there will be everything. It's a progressive process. More and more becomes available, AI becomes embedded in Microsoft Office suite and all of these products. And in 10 years, the small business, without making any significant investment, will have access and capability.

[00:28:42.290] - PS

And I guess there'll be a lot of service providing businesses that they will be like, for example, a company like Xero , they might be in bookkeeping and HR, but over time they will see the good reason to add those AI type functions to what they do as well. And they are catering for the smaller end of the market.

[00:29:08.960] - MM

Exactly. That's absolutely spot on. So, I mean, Xero is a good example if you think about it touches the epicenter of small business, which is financial records and transactions and so forth. There will be so many obvious places where some intelligence put into that application that can provide guidance, recommendations and insights, et cetera, would be of enormous value. And then you think about every company doing that over the next 10 years. It empowers small businesses with the technology.

[00:29:39.950] - PS

And I guess then the fact that there's a potential demand from those sorts of companies to provide for small businesses, then the technological research to try and reduce the cost. As you pointed out, calculators were very expensive and all of a sudden calculators became cheaper. AI over time will become cheaper because something discovered for a big company advance will be rolled out for a whole bunch of companies, economies of scale and so on and so forth.

[00:30:07.820] - MM

Exactly when when I was graduating from the university in the mid to late 90s, a website for a big company might cost a million dollars to develop. And it was html programmers and it was a big deal even when we were a startup and we got a basic website built and this was the year '99 Peter. It cost us USD100,000. How much can you build a website for today? You commoditiescould build it for $500 or $1000. Everything is commoditised. Everything becomes accessible to mobile phones. Remember when they were in a briefcase. They were brick and cost ten thousand dollars. Everyone has one now. And the famous quote is that the mobile phone today has more computing power than when NASA put the man on the moon more computing power than all of NASA when they put the man on the moon in everyone's pocket around the world. So AI will follow that trend because all technology follows that trend.

[00:30:59.960] - PS

Okay, this is there anything else you think we need to know about Complexica?

[00:31:07.550] - MM

No, I think you've covered it. It's a software to improve decision-making and achieve commercial outcomes. I think one interesting side note just to mention on AI, however, is there's an enormous amount of hype and an enormous amount of misinformed people because ultimately newspapers, whether they're digital or in print, they'll sell more of them with stories like AI will automate half of the jobs, AI will take over the world, AI will do this - that sells more newspapers rather than I might never happen or I might happen in a hundred years. I think that's just an interesting point, that the reality of AI is a lot different to most people's perception from what they've read.

[00:31:49.480] - PS

Yeah, and you've opened up a Pandora's box for the media, but when we open this box, is AI as scary as the media portrays it in terms of they're going to take away all these jobs, everyone will be unemployed?

[00:32:06.980] - MM

First, I don't think AI is as scary as the media portrays it, because I've lived inside of universities and kind of seen the state of the art of where it's at. And we're an enormously long way, Peter, an enormously long way from even beginning to approximate the human brain and it being a real concern. And in terms of taking away jobs or that is the fundamental goal of a corporation of any capitalistic market and corporations in particular, the directors have a fiduciary responsibility to increase the value of the enterprise, become more competitive, increase profits, increase market share and so forth.

[00:32:46.230] - MM

And hence, because of this nature of capitalism, they will always look for ways of automating jobs, becoming more efficient. And AI is just one of many, many technologies that will be considered, from 3D printing to machine data to all kinds of things. So I think AI has unfairly gotten a lot of, let's call it, bad publicity for being this evil technology that's going to automate everyone's job away and everyone should be afraid of it. Where we're automating jobs away without AI every day and have been for the many last decades past. It's just the nature of capitalism.

[00:33:21.870] - PS

I do think that if I recall there was an economic theory that as technology improves, we actually will end up with more income and more leisure time. I think that leisure and leisure time we'll go to hotels and resorts and fly. Yeah, take the coronavirus out of the way. But the actual fact, the trade-off between how many hours we work and the lifestyle we lead could actually be enhanced by AI rather than dis-enhanced if such a world exists.

[00:33:55.680] - MM

Anything is possible in terms of the future and how it unfolds, and I remember I think it was in the 50s when they talked about the 80s, in the 80s, you were supposed to work 20 hours a week or 10 hours a week and there would be one day a week working and there'd be all this production that happens in automated ways through robotics and so forth. And obviously, none of that has really come to be other than quality of life and life spans kind of extending. So I think it comes down to government policies, to how societies are structured, redistribution of wealth and many complex issues like this that ultimately dictate the kind of lifestyle we'll enjoy as human beings it is. There's a reality that might be in the future that one percent of people own everything and the other 99 percent are working for minimum wage. For that, anything's possible.

[00:34:50.520] - PS

And I've got a really good idea that we could use artificial intelligence to find out what's going to happen and when it's going to happen. Now, before we go, I can't help but ask this question because you are an AI expert at a company that I've liked ever since its share price fell. This company called Appen, a company that did very well before the coronavirus. And because of the coronavirus, a lot of the big companies that it works with were cutting costs and people weren't going to work, people working from home. So there's a thought that maybe the market has marked that down too heavily because of the short-term problems for the company. And we have also learned that some of the big companies might start trying to do some of the stuff that Appen used to do externally for them could be brought in-house. And if that's the case and you and I were talking earlier, it had an exposure to a limited number of big companies. And my question to you is, if what you're saying is right that AI will become more accessible to medium-sized businesses and ultimately small businesses, a company like Appen may well be able to grow their customer base over time and not be so dependent. Is that a fair, possible call on that?

[00:36:14.670] - MM

Yeah, absolutely. I think Appen is a fantastic company, great business, and one of Australia's great success stories. And they provide the service really to train artificial intelligence algorithms. One of the shortcomings of AI today is, is they require a lot of training. And some of the more complex algorithms really require extensive data and training and processes which Appen has been an instrumental provider of. So given the extent of algorithmic development and consumption by big companies, Appen has ridden that wave because they've been providing the service to those businesses to train the algorithms and provide training data that the weakness of the businesses, the huge concentration of their revenue in a very small number of businesses. However, as the technology commoditises, becomes more accessible for more businesses, it should open up marketplace opportunities for them to diversify revenue streams through the same kind of service.

[00:37:18.120] - PS

Well, this is in no way financial advice or a tip or whatever, but I certainly had a great opportunity to talk to an AI expert about a company that is in the AI space. Matthew Michalewicz, just before we go, the name of that book, one more time,

[00:37:32.760] - MM

The Rise of Artificial Intelligence.

[00:37:34.830] - PS

And now I try to look it up on Google. And you said there's a short-term problem with the way you find. Like the hardest thing the world is trying to spell, Michalewicz even pronounced is hard enough. Try spelling. It is the E before the W or the C or the Z. It's a tricky name to spell.

[00:37:54.970] - MM

I know. I know. But it's not easy. It's not as tricky as Schwarzenegger.

[00:38:01.200] - PS

Well, no, it's not so much as the rise of artificial intelligence by Matthew Michalewicz. And it's great to see you.

[00:38:09.030] - MM

Likewise, Peter. Thank you for having me.

[00:38:10.710] - PS

And that's Matthew Michalewicz. And if you are interested in getting rich in artificial intelligence is going to be one way you can do it, you can also watch my TV shows, which come out on Monday and Thursday. Just go to YouTube and punch in Switzer investing and you'll get there. Thanks for joining us. See you next week!