The investor’s guide to Artificial Intelligence. July 5, 2024

In short

Podcast Summary: Motley Fool Money - The Investor’s Guide to Artificial Intelligence

Episode Date: July 5, 2024 Hosts: Scott Phillips and Andrew Page Description: This episode explores how investors should approach the rapidly evolving field of Artificial Intelligence (AI), discussing potential winners and losers in the investment landscape.

---

Key Themes and Discussions

Introduction to AI and Its Implications

AI's Current State: The hosts discuss how AI's capabilities have evolved from simple tasks to more complex functions, resembling human-like interactions.
AI's Future: Speculation about the future of AI and its potential impact on industries and employment.

Concerns About Scams and Misinformation

Scams: Andrew Page shares his experience with scams using his likeness to promote fake investment schemes.
Misinformation: The conversation touches on the prevalence of misinformation, particularly in the realm of AI-generated content.

The Growth of AI

Rapid Advancements: Discussion on significant breakthroughs in AI technology and how they have outpaced previous expectations.
Transformative Potential: AI is seen as a tool that could alter industries as profoundly as electricity did in the past.

Investing in AI

Investment Challenges: The hosts express caution about investing in AI, noting the uncertainty surrounding which companies will emerge as leaders in the space.
Historical Context: They draw parallels to past technological advancements (e.g., the internet) and the fluctuations in company valuations that occurred during those transitions.

Potential Winners and Losers

Companies Adapting to AI: Exploration of which companies may thrive by integrating AI effectively into their operations (e.g., data centers).
Caution on Hype: The hosts warn against the hype surrounding AI, suggesting that not all companies claiming to use AI will succeed.

The Role of Regulators and Ethical Concerns

Regulatory Challenges: The need for regulation as AI becomes more widespread and influential.
Ethical Considerations: The potential for AI to disrupt employment and change the nature of work, with discussions on social ramifications and the need for societal adaptation.

Conclusion

Future Uncertainty: The episode concludes with an acknowledgment that while AI holds vast potential, its implications for society and investment are not fully understood.
Investor Mindset: Emphasis on maintaining a long-term perspective and being cautious with investments related to AI.

---

Key Takeaways

Be Cautious: Investors should approach AI investments with care, recognizing the rapid pace of change and the uncertainty of outcomes.
Understand the Technology: Familiarity with AI's capabilities and limitations is crucial for both consumers and investors.
Watch for Hype: Not every company claiming to leverage AI will succeed; distinguishing genuine innovation from marketing buzz is essential.
Prepare for Change: The workforce will likely undergo significant changes as AI technologies advance, necessitating adaptations in skills and employment strategies.

---

Note: For further insights, subscribe to the Motley Fool Money podcast and follow their newsletter for regular updates on financial advice and market trends.

Written by AI. May contain mistakes. Listen to the episode to check what was said.

Hear the part that matters, and keep it.Open this episode in VO. Double tap your headphones to save a moment as you listen.

Get VO free

Transcript

Automatic transcript. May contain errors.

0:07Welcome to Motley Fool Money, the podcast that might, by this time next year, be done by robots. I'm Scott Phillips from The Motley Fool. He is Andrew Page from strawman.com. Although, in a year's time, maybe I'm not. Maybe he's not. I have talked to you about AI before. I've talked to you about being scammed before, and I'll start with that story. But before I do, welcome Mr. Andrew Page from strawman.com. How are you? Is it me? Is it a deep fake? Who knows anymore? If I was speaking with a British accent, it may well be a fake. I did mention it a previous pod about a month or so ago. So people are using my name and video and dubbing it to like promise people 60 % guaranteed returns, which is one very, very small point flattering that could be bothered using me.

0:54On the other hand, it's just really yucky and kind of awful and people are going to lose money from it. Just really annoys me. And the fact that they're actually paying the social media platforms, like the media platforms are getting money from scammers. And then when my followers are reporting, they're saying, oh, it doesn't violate our standards. So what do you, what need, anyway. Yeah, what standards? But, oh yeah, are you talking about me or the platform? No, no, no. Yeah, yeah. Just checking. Certain unnamed platform owned by a certain billionaire. It might rhyme with Facebook. Mate, we are going to talk.

1:22So, first things first, I'm away right now. You are having a well-earned break from my presence. But the penalty for that is you had to sit down with me and pre-record some episodes. And insert the usual joke about pre-recording episodes. We thought we'd talk about AI. and I was kind of half joking and half serious about the growth now. You are a massive, massive tech head. You love cutting-edge innovation stuff. You're just all over this like a rash, like a fat kid on cake as they used to say before that was politically incorrect. And yet, so there was - I can't believe you said that. Can you not say that?

1:59Not anymore. Really? Well, if you're going to go there, my favorite one was always like a pit bull on a poodle. I always thought that had a certain elegance to it. You win. So, and we kind of thought, what do we do? So, first things first, listeners, for the next few weeks, we're going to give you episodes every day as normal, Fridays and Sundays, but we have recorded them in advance because I'm away. And so we're not going to do news of the week. So whatever happened this week happened, and Andrew probably knows about it. I may or may not know about it if I'm in touch. However, we can't comment on it because we're not here, or I'm not here anyway.

2:32We are going to, though, give you some episodes, as I say, and we're going to try and tackle some of the more, call it evergreen, it's been one of those kind of fancy words that means just stuff that isn't time sensitive. But we did want to do stuff that was topical for at least part of the next four weeks. The mailbags, by the way, will proceed as normal. And Ram, I just thought AI, well, you mentioned it, and I kind of thought it was worth doing because it is huge. We've kind of touched on this before, but we thought we'd dedicate an episode to it just to really try and unpick it. And essentially from an investing lens, although as per our usual conversation, we'll kind of be all over the place and kind of going from there.

3:11I am fascinated by the growth of AI. I think on one hand, right now, it's basically just a natural language search engine in a lot of cases. Or they talk about the fact it hallucinates, it makes up stuff. So it's not even necessarily a very good search engine. but the idea of kind of i think part of what we're enamored by is the fact it can talk like a normal person doesn't throw up 15 google search results it says actually here's what i know here's what i can find or here's what i can make up because i've learned from other things i can recreate um but in a you know bespoke unique way some content that hasn't been produced before and that is on one hand a bit of a magician's trick on the other hand incredibly impressive and is the doorway through which we will go over the next 3, 5, 10, 15, 20 years to an amazing future that I don't think any of us can really properly imagine.

4:04And I've got two thoughts in my head. The first is we went from being on the ground to space in, what was it, 56-odd years? To the moon, sorry. Like, you know, Orville and Wilbur, before the plane took off, add 60-odd years to that and we're on the moon. Like, that is stupid. And the day before they did that, people, experts, scientific experts and engineers are saying, man, we'll never fly. We will never achieve powered flight. Yeah, yeah. And yet, I grew up as a kid in the early 80s watching the Jetsons and my flying car is still not here. So there is a range of potential outcomes here, right?

4:42So I'm going to start, mate, by letting you off the leash a little bit here from a tech nerd perspective, because I know that's your happy place other than Bitcoin, but maybe that's the same thing. Same thing. same thing um paint me a picture of 2044 what's oh i can't what's ai doing for us 2034 i can't do i don't think i can do 2028 i saw a interview the other day from a researcher in the space and um he was saying if you had gone back 10 years and said so it's 2014 and said by the year 2024 we will have these programs that you can ask any question of in natural English. You can speak it, you can type it, and it will give you an answer as good, and he used the term, as a bad expert.

5:33So he's just being real. You find a domain expert in whatever particular domain, they're going to be better than these AI models. But the answers they give are going to be much better than the average layperson who doesn't know anything about it. And it's actually going to be as good as what you might consider a pretty ordinary expert in the field. Which is, I feel like I'm downplaying it now, but the significance of his answer was like, 10 years, like, that's impossible. That's 50 years away. Now we've got it. We've got it. And by the way, for the average person, ChatGPT launched, not only without fanfare, but it was like one day we didn't have it and we didn't have any real, I mean, the experts probably knew what was coming or at least what was being worked on.

6:13But it just kind of arrived. It was like, oh, it's like, I'd be a present on Christmas Day. It's like, oh my God, this thing is like, it's a thing. It can do its stuff. So there was a computer science breakthrough called the advent of transformers, not Optimus Prime and Megatron. But these - The proper transformers. It's a process, which kind of opened the door to that. And this is how science and technology works. It's a series of quantum jumps. It's not this linear, a little bit better, a little bit better. It does happen for some, once we've sort of gone from zero to one, things get a little bit better, a little bit better.

6:46But you have that zero to one moment. And so AI has been around since Turing. They talked a lot about it way back in the middle of last century. But then we had all these what they called AI winters where no progress was made. It was like, oh, I was overhyped. And then there was a breakthrough. And so we had a breakthrough pretty recently. Now we've got these things. And the nature of these breakthroughs, they tend to happen more and more frequently. So past breakthroughs lead sort of like it's the shoulder of, you know, I've seen far because I've stood on the shoulders of giants. Like previous breakthroughs have laid the foundations for further discoveries.

7:20And it's moving incredibly fast. Going from GPT-3 to 4 is a mind blow. Yeah, that was cool. And the other point he sort of made was, and I make that first point just to sort of say, 2044? Oh, my gosh. You can't tell me how fast. This is moving so rapidly. Can I remind everyone the iPhone is only 17 years old. Right? Right? So the 20-year, I was being a bit facetious asking you just to paint a picture, but the 20-year thing. Yeah. And you know what? I've said this before so many times, but what I find amazing about the human race is we have gone from this didn't exist to it's just such a normal, natural, accepted part of life.

8:01We should every day look at our phones and go, oh, my God, I can't believe I still have this amazing thing. and now it's like well if the internet's down for when Optus went down for a couple of hours we lost our collective minds newspaper headlines parliamentary inquiries you know it's like and reasons for that aside it's like we just you go from this can't happen to oh wow to meh in such a short period of time Was it Asimov or someone like that who said that any technology sufficiently advanced is indistinguishable from magic like I went back to the middle ages with iPhone you know yeah just like it is it's yeah it's magic you can't comprehend it and and it's just and again you go back to 2014 and say that we're going to have these agents right yeah yeah yeah and they haven't been let off the hook yet i gotta emphasize that so the second point i wanted to make with with what this guy was saying um i'm so sorry for not being able to attribute his name um was that there are so now we've got we've had the zero to one here a big part of what made gbt4 so much better than gpt3 was just bigger in terms of it just had it had more um variables and it just so just by if we do nothing we get no more breakthroughs and we just dedicate more compute they will get more powerful okay right so we we can and again it's just out of the gate here so there's just general optimizations and tricks that we'll get to make them much much much better um so yeah i i i i know that when electricity first came along it was overhyped and dismissed yes yes when steam came along it was overhyped and then dismissed i mean every major breakthrough the model t for like radio uh the internet what are the cars were got perfectly good horses yeah yeah right you know and and it it takes longer to for the potential to be realized but then when it happens it moves faster than anyone could possibly imagine and and you have what they call these heart uh gartner hype cycles which well define um these these kinds of things so i think we are we are at the dawn of the sort of mainstream adoption on on some of these models um and i i've gone from playing around with it to it now being an integral part of my work processes i've always i use clog because that's the through anthropic it's one of the companies um i just tried a friend recommended i've tried it i pay a 30 a month subscription the best 30 i've ever spent like wow i got my money back in day day one yeah and it's like they could up the price to 60 tomorrow 100 i'd pay it because i just get so much use out of it and and do you remember when google first came along you and i are old enough to have this conversation but for the kids who are listening i remember when it came along yes there was an art to googling Yeah.

10:52There was, right? Like Google's gotten really good. So you can be pretty bad with your question and they'll probably find you the right answer. But there was a time when you had to be pretty cluey about it. It's the same with prompts in AIs. Like I think the people who I've seen who have dismissed it have, don't have a fair appreciation for what it is, how it works and say, ask it to do something. It comes back slightly less than perfect. And therefore it's a big waste of time and overhyped. Yeah. Not, not recognizing that the fact that it was just able to do it at all was a miracle. And in seconds.

11:24In seconds, right? So how does it, like one more little thing, and I'll throw it back to you, is that the best way to understand it is that they are basically statistical, big statistical machines. And so what they're trying, they take your prompt and they've got like a sort of a God prompt, master prompt that's sort of engineered behind the scenes that sort of give it guardrails and the rest of it. but it just tries to find what is the next most likely word in this sequence given every single bit of human knowledge that's been fed into me like since the dawn of time and it will and in its response it's not even word by word it's letter by letter what's the next most likely letter what's the next most likely letter and it just does that iteratively and it just so happens that when you do that you get answers that are really good which opens the door to like holy holy good grief is this how our brains work are we just large language models like how that should not work that well right but we may have i mean you know neurologists have been trying to figure out like what is consciousness what is sentience how does that work and it i mean it's just race there's nothing magic i would argue you know it's just physics there's just cells going on in our head there's no magic there it's just the processing is that we don't understand it Anyway, what I'm saying is that we're just getting really good at figuring out how to do this stuff, and it's accelerating super, super fast.

12:50And there are certain technologies that come along that, like, you know, there might be a new way to – we talked about Fastbrick before. It's a small ASX company. They've got a robot that lays bricks, you know? Okay, cool. Is that a game changer? Well, I suppose for a one-niche industry it is. ai is a game changer for almost everything and you can apply it to those individual industries in individual ways or use it as a generic tool across as much as you want yeah it's like electricity or the internet it's that big a deal and it's just like you point to me a company a business an enterprise that doesn't use that in some way shape or form you just can't find them you might find a goat herder in the back of mongolia somewhere who doesn't have it doesn't need the internet but that's about that's the that's like the only places left so this is going to be without being too hyperbolic one of those things where i just i think it changes everything and it most likely will change it in ways that we can't predict because we we couldn't we knew the internet was going to be a big deal but how we envisaged the internet turned out to be far different than what it actually was that's true and even when you say we i think there's going to be a little bit careful about we knew it would be a big deal right i've said before tom and david gardner appeared on yeah it might have been oprah one of those shows and they were mockingly uh asked because it was already their view surely people won't use their credit cards on the internet because it's not safe yeah now you think about the growth of online commerce as a tiny fraction we're about the internet most of us think it's the browser in front of us which is therefore the couple of you know newspaper website the shopping site i go to and email and it is all those things but it's just a squillion times more um so it's only a single example but you know they were derisively mocked of course people won't do that what the hell's wrong with you people you know and so there is some element of and this is where i think your analogy actually holds up which is some people did not be a big deal others of us caught on it reasonably quickly others took a while to come to it but now it's got to the point a bit like i said with the iphone where it's like well of course it's there and it all of the things that it does um and and it's you know to your point about it becomes so ubiquitous and so universal that it becomes part of workflow you know um this morning i got a notification from australia post that apparently i've got something else coming to my house being delivered sometime today um stupidly small example right but possible i can track that package um you know australia knows when it's been delivered the customer the company who sent to me knows when it's been delivered i know what's been delivered again stupidly tiny useless example given everything else the internet can do you saw a paramedicus before the technology behind the ability to capture store and share and retrieve images medical imagery um we don't have to go and schlep a physical plastic bit of x-ray sheet around the around the you know the state city whatever to go well have you got a broken bone or haven't you is that really brain cancer is there you've got a shadow in your lung um these things are possible because of literally the internet i mean you and i are chatting over the internet and then we're going to send out the podcast on a different part of the internet or different you know iteration of but these things that we just take entirely for granted.

16:03Netflix. And again, they're stupid consumer examples. There's so much going on behind the scenes that power just terabytes of data and processing and whatever else. Even things like, speaking of things we used to do, I had an iPod that I'd plug into my computer and it would download the new songs from my computer onto the iPod, right? And on my first phone, you'd go, I haven't updated the apps in a while. like going up there at the apps. Windows used to come on eight bloody three and a half inch floppy disks, right? Yeah. Not three and a half, the five and a quarter, three and a half hard, the little hard floppy disks.

16:38It's, you know, and I think, I only use those examples and they're horrible examples if you're old enough you remember some of them. Just to reiterate your point which is it becomes just so unremarkable once it becomes ubiquitous and that is the, I think, the thing we need to go through. You and I aren't technologists, mate. We're business people and investors and I mean, you love a bit of tech i'll have a bit of tech too although uh you take the cake for a tech nerdery uh between the two of us um so i am i am curious the while we can't know what ai is going to do in 10 or 20 30 years um maybe you should read some more as i guess i'm often find out um replicators may yet be a thing uh from star trek the the reality is we're investing in a world in which AI exists.

17:24And I've used the example before, David Gardner, again, a lot of our co-founder, talked about, you know, a century, two centuries, a millennia ago, when you were born, what your grandfather did, your father did, what your father did, you did what you did, your son would do. And I mean, I deliberately mean it male because women weren't working or if they were, it was unpaid work, you know, domestically and otherwise, you know, picking fruit and whatever else, grains and stuff from the family plot. But, you know, those things were designed or defined by your birth circumstances and there is still an element of class snobbery and limitations i'm not going to pretend it's utopia but we go from all you know if you're born um you know harvesting wheat on the nile you're going to do that your kids are going to do their grandkids are going to do that these these days there is so much more to something we'll promise individually but societally as well So I'll ask you the general question.

18:17We'll get back to AI in a second. How on earth can you invest in a world where things are changing so dramatically, so quickly, value is being created and destroyed? How can one be a long-term investor in 2024? It's such an excellent question. Howard Marks wrote a really good blog on it, I want to say a year and a half ago now. Just sort of like there was a stage where you could look at a certain industry or group of companies and go, I'm almost as certain as one can be that these companies will be in existence in 10, 20 years time and they will still be viable, profitable entities. You can debate the exact size and success, but, you know, they're not going anywhere.

18:53And I don't know if you can say that. I mean, look at all the biggest companies on the US market today didn't exist 20 years ago. Like just didn't exist. In fact, in many cases, less than that. And that's only accelerating. So, yeah, that's a really tough question. I often say that the first question I ask when I look at an investment is, is it going to be here in 10 years? Because if I can't answer that question, then that's it. I'll move on. What's the point of doing all the analysis and valuation if you can't get that one part right? So, it's a good question. I think some things will be easier than others.

19:31so you know i mentioned to you canva the other day on might have been one of the other pre-records we did yeah so they they do imaging uh the scrapbooking posters all of that kind of stuff and now the image models are so good i don't know if they will adapt or die they i note that they are now spruiking oh we've got ai we're doing this come with us and now i and maybe their ai is better and they that just it goes in a certain direction where the big players in that space remain the big players and they just use ai and that's just all the place to go but but maybe it doesn't maybe i've just got a general ai agent that can do all of that for a fraction of the price i don't need all your fancy stuff because it just does it yeah and it's existential so it's not that i have a strong view on which one it's going to be but i don't know and whenever there's something that's like i don't know i'm going to be a lot more reluctant now if you want to talk about supermarkets woolies gas pipelines, things of that airport runways that I feel as though they're going to be around for a lot longer, right?

20:33Or they're going to be much more difficult to disrupt. So, so you, I don't want to give a definitive answer because you really got to think it through and you can have a 10 hour discussion on every single industry. But I would start with that view is like, am I, what parts of this business could be done by an agent? And, and I actually, I need to go back a step that's one of the other things you've got to recognize with ai is it is a very general tool i mean a calculator is artificial intelligence we don't call it ai anymore but it is it's artificial and it can do addition subtraction multiplication and everything else better than the most brilliant mathematical mind that ever existed yeah right it just it it absolutely can but it's what they call narrow ai it does that and that's all it does it's it's not going to give you a recipe to cook butter chicken um but these these llms these large language models and the like yeah they they are extraordinarily general so it's going to be a it's going to be a very very difficult question um i do i do think one one thing that will change in terms of our daily lives is we're all going to have really good assistants yeah like i got a little bit overexcited when i don't want to say the word because my phone's going to turn just spell it you know you can say hey yeah x or you know siri or alexa or any and and they kind of they didn't live up to their expectation did they i don't know if you found that i try you know what can i can i see here's what i meant about ubiquity thing right it was amazing for the first six months right and then we kind of got frustrated that it wasn't better than we wanted it to be i think i never got it did plateau very quickly it did that's true that's also true but when you i use it at home all the time i set a timer i put some music on i and again very narrow very very those few things it does really well exactly other things it's just like oh my god it's like how many i'm just going to type it it's too hard exactly and i think and that to me is the i think that's the key difference and it's the point about ai is that again all these things is if you if take it back any number of decades and look what i can do yeah this is like oh my god now it's like well i'm a bit frustrated by that because it hasn't hasn't continued to grow and get better and those gaps haven't been closed it's fascinating actually just quickly um so i'm a google user and i have the google gemini and it's on my phone you can actually replace the google assistant with gemini as your assistant yeah and what's fascinating is it does more things that gemini that the assistant can't do but it can't you do some things the assistant could do yeah and so part of this part of this iteration is like how do you how do you how do you how do you port over those very definitive specific things that that i again your point about the calculator right the the alexis and series and google assistants are super narrow so they do these these few things they'll kind of program to tell me a joke or it's like yeah okay fine um by the way kids still love that so that's that's still like a crowd pleaser but when you then port it over to actually now do things creatively or iteratively that's a fascinating gap you see big companies like google and again i'm shawler for anyone who cares um struggling to kind of not even struggling but going through the process of kind of trying to meld those two kind of concepts of i can do this definitively or i do this iteratively and the definitive stuff is always going to be better at that specific task because it's probably to do that thing whereas you're relying on the ai to kind of develop an ability and then somehow apply that to a set of devices or or something else you know what you're trying to do you set a timer right it's like well it's got to have integrator interface with a device to do that um i find that part of it amazing for me mate i think the the growth in the growth in ai is going to be fascinating and you're right about the the grosses but even then i think the i i still come down on the it's the implementation of the tool not the tool that's going to make the difference yep and i say that because if you look at i think retail is a fascinating example for what it's worth because you start with a you know there were some online grocers who tried to do online um only and they they crashed and burned woolies and coles however have adopted that and their online commerce i'm pretty sure woolies now more than 10 so well the one dollar and ten woolies makes the transaction is conducted on the internet Now, you can't make the groceries appear through the old telephone 56.6 dial-up modem, but you can have the transit, you can get delivered, you can go and pick it up.

25:07The implementation of that has been amazing for these guys. And I still have a suspicion that the game will be or can potentially be changed, even in existing industries, by the better application of that stuff. The change from, you mentioned Google before, a good example only in the sense that, or not only in the sense that you take all the search engines why did google win because it came it found the algorithm quickest and then it built on that to stretch that lead just further and further you know it's almost like facebook's a bit the same that neither invented the search engine or the social network right they were super late in both cases but they went here's a better version and then we're going to use that and use that advantage to create this incredible snowball that rolls away from everyone else so incredibly quickly so for me the hardest challenge with ai is not so much what's going to be disrupted by AI.

25:56I mean, that's its own mind-blowing, I can't wait until it happens. But it's like, which of Woolies or Coles or IGA or Aldi or Costco, or by the way, it might be Amazon shares, selling groceries or someone else, kind of goes, here's a cool way to do this thing. People have been trying for years, for example, I used to work in the coffee business years and years ago. And even then, they'll try to get ways of kind of automatic replenishment. Like you have the naivety and kind of the, I love that they tried. One company had a scale and you would keep your coffee beans on the scale. And so when the scale got under a certain level, it would send a signal to your Wi-Fi, send to the company, say, you're getting low on beans, send them some more coffee beans, right?

26:36Now really, actually quite a cool idea, except that people are going to keep scales for every individual piece of food in there. And maybe it is coffee because coffee's special to people. There's going to be a way that we use AI in some way, shape or form, right? to do some sort of automatic replenishment at some point. And the company that gets there first and wins my automatic replenishment business is going to have hundreds of bucks a week from me just straight out the door, come back in the door, like, you know, whatever those things are. And so I think that's going to be fascinating. That's almost, I think it's always harder in some ways to really properly predict because you kind of go, oh, hang on, this company's in the software business, Canva, right?

27:11A really easy way to say, if they do it well, they could 10X the business. If they do it badly, they're going to get smashed and someone else is going to beat them to it. but groceries how does that play out or to your point you mentioned airports that's great but it's like well is it is it the is it the um experience of actually using uh an ai instead of having to go to a place is it is it when you the process of getting to there and from there is it the does it you know reduce the amount of waiting time because they know when you're arriving and they can do i i i can't and this thing i can't even imagine what the permutations are right i'm not a futurist but that's the bit where you're like i don't know how this plays out i think i will go back something you mentioned on a previous episode i'm getting kind of to the conclusions but i don't mean to get there too early the for me it's the it's the things that companies do well that earn your business and whether it's a monopoly like an airport which is kind of partly i think your point or at least your example uh where there is pricing power because a brand uh can i impact that i mean kind of a coke is coke and pepsi is pepsi and you've got your favorite and how much AI can you use to really change your preference, right?

28:17Those companies will always retain a meaningful lead. And I think I kind of go back to, there's nothing new in investing. I mean, there's a million new companies and great new sectors and amazing new products, but you could put Ben Graham effectively in the Australian stock market today. Maybe not Ben because things change the way he invested, but you could write his books today with some different examples and not really have to change much. And I think, you know, that there is something of that. The companies that were amazing, so you mentioned calculators, Texas Instruments or Hill at Packard, right?

28:51Almost don't exist anymore in their current forms. HP is apparently doing corporate stuff which they didn't realize until the other day. Gone as a consumer brand, right? The ability of those guys to kind of, in their day, be the innovators and the disruptors, and the whatevers. And there's the businesses that just use the tools to keep being bigger. American Express was a stagecoach company. Right? And you kind of look at AmEx today and say, that's an example of waves of technology. And they're just adapted and adapted and adapted and adapted. And maybe they're one misstep away from oblivion or maybe they're, you know, one great step away from even more success.

29:28But yeah, I think it's the implementation and execution of whatever the technology and tool is that I find both maybe strangely reassuring and challenge you as an investor because it's, you know, you can do the one or zero outcome for some of these tech companies. But for those who are like, I don't know how they'll use it or who'll use it against them, that's what's going to be fascinating. Oh, gosh. I mean, I agree. Like, as bullish as I am on the technology, I haven't made any AI-specific investments. I can't. It's too early. Too many potential futures. I can't figure it out. Is it something which is more democratic or – that's not the right word – something that's more widespread, like email.

30:09Everyone uses email. You just do because it's better, but there's no real competitive edge. Is it something like that? Or is it something that only a handful of companies control? And one of the things with AI is that the good AIs are good because they've been trained on lots of data. And so they've become better. And then I've got a better AI than you. And so more people start using my AI, which means I get more data, which means I get even better than you. Yes, yes, yes, exactly. And maybe that's the process. But then again, you know, I was playing around with one on the weekend called Venice, which is an open source one.

30:46So there are people in their garage launching LLMs now too because the know-how has been released. So maybe it's not anthropic, maybe it's not open AI that controls it all, or maybe it is. And then that's, oh my gosh, that's a whole other kettle of fish and potentially real 1984 dystopian kind of stuff where it's like there is one agent here. And again, it just runs ahead of everyone else so far, so significantly that there's like, I'm using open AI because why would I use anyone else? It's just chalk and cheese. And there's a real potential for that to happen as well. So it's super tricky. On the picks and shovel kind of play, I'm coming really around to the idea of data centers, I've got to say.

31:28Are you okay? Yeah, because I avoided them for a long time thinking because the argument used to be, hey, the internet's not going away. It's only going to get bigger. We need a lot of data centers. And I always thought, yeah, but it's very capital intensive. And ultimately it's a commodity play because your ones and zeros are exactly the same as my ones and zeros. So what's the difference? But one of the things, the other differences with AI is that they are very compute hungry. Very, much more, much more so than just plain vanilla Google server. Like the resourcing required to have these models answer everyone's questions.

32:03It's why on the free versions you only get seven questions a day or seven prompts a day. They just can't handle it. And I think that that build out will happen so fast that there will be a lot of pricing power for those that are early on where it's like, yeah, eventually we might have enough data centers. But that build out, the demand is going to outpace supply, I think, for a considerable amount of time. So that's an interesting play potentially. I don't know if you're going to break up the data center question between the real estate, the management, and the equipment. Because I'm still not – you make a compelling case.

32:37I've avoided it for a long time because at the end of the day, a data center is data center. It's a specialized property play in my opinion, right? It's a warehouse. You can have racks of Bunnings, potting mix, or it could have TVs, or it could have data service, right? And I'm like, well, why would a data center REIT attract a higher multiple than any other REIT? If it had pricing power that was beyond its, you know, per square foot, per square, whatever, that makes some sense, right? If you're getting the value of that almost as the management bit rather than just the pure real estate bit, I kind of get that.

33:13If you can charge more rent.

33:18It strikes me, your description hit me both ways. One is there might be a really good short to medium term opportunity. The other is kind of like, still feels like oil or lithium, where it's like it's big until it's not, right? You make outside profits until capacity catches up. And at that point, we're all back to single digit margins again. So that to me is kind of like, I'm not sure which way to go. I haven't looked at - It's only predicated on there being a very lengthy imbalance. Yeah, okay, yeah. Because if I've got, I'm Microsoft. Oh my gosh, everyone wants our new access to our model. We just don't have enough data centers.

33:54call up Australia, call up Singapore. We need to rent it. Okay, well, this is the price. Well, Google's over here. They're bidding this. And you'll pay what you can pay. Why is NVIDIA the number one stock in the world, or the best performing stock in the world? Because of the chips that they make, because of AI. That's the whole story with NVIDIA. It's just, wow, there's not a lot of companies that can make the stuff that we really need, and they happen to be one that does it and does it really well and have made the investment, and that's why they've done that. Again, eventually there'll be enough chips out there, I suppose, where it's saturated and supply will meet demand and that dynamic will change.

34:35I don't think it's going to change anytime soon just because of the explosion of demand. Eventually the unit economics will kick in. Right, right. But I don't think that's like – If we were talking about something that was like three, four, five years, then yeah, it's pennies in front of a steamroller. But I suspect it's longer. By the way, this is all navel-gazing stuff, right? No, I think it's a really compelling idea, mate. I mean, NVIDIA keeps growing. This is why the company makes the most of the AI chips that are being used. It's growing some stupid rate a year, like just for, I want to say 30%.

35:12People will pay anything for the chips themselves because I need it, right? Yes. And all of them, they all need it, right? So I think that's really interesting. Which is funny because you're playing chicken at the same time as taking the opportunity, right? It's like, I'm going to hope this continues and I hope I'm out by the time the music stops. And that's the hard dance because greed takes over. You're taking money too early. I mean, it's made for regret. It doesn't mean you can't make an absolute squillion dollars on the way through if you're right. Oh, yeah, true. Another phrase I'm very fond of is, I think this is a Marc Andreessen quote or whatever, but it's the future is here.

Read the full transcript

35:45It's just not evenly distributed, which I've liked. And I think you sometimes feel as though you need to be in the research labs to be early enough. And I just, I guarantee you, go down to your local main street and ask 100 people what they think of AI. Yeah. And I'm sure at least 60 people will go, well, I don't know, some hypey thing. I wouldn't have even turned, wouldn't have had any direct experience with it. Probably 80%. And then of the other 20 % that have had a direct experience, I had a little play when it was in the news and I haven't used it. And you've probably got three people who are using it.

36:17So I would make the case we are insanely early. Yeah. Insanely early. Yeah. And even still a reasonably limited use case. I mean, I'm waiting for what I'm hoping. Again, I'm a Google shareholder, but I'm a Google user more than that. And if I had to choose one, I'd choose Google user. I have Google shareholder. I'm so embedded in it. The ability for Gemini, and we all kind of felt icky when Gmail first came out because, oh, they can read my emails, which is always. The privacy stuff, privacy is really important that no one else can access my stuff. but if a computer knows the stuff about me and can better tailor stuff i have no issue with it i know a lot of people but so be it um but when if and when gemini can properly put not only use the algorithm the general algorithm but personalize it to me so that it knows who i am what i want how i use it you made the point about the prompts i actually think i would i would suspect let me make such stupid outlandish prediction in five years time you don't need to be a good at writing prompts no i agree because it will know what you want well no what people want generally but like like you said with google search but also it'll know you so well it's like oh you're back again um last time i said these things you work in this area um you've asked about these things these are your contacts this is what this is the source you use and it'll just it'll refine that response so it's the movie it's the movie her writ large right you'll have you'll have your own agent yes and it will it will be trained on everything that they're trained on exactly be trained on you i just i just know what you like i know what you're interested in yeah yeah and the ability just kind of go straight back to that so i was like okay well you're here again rather than you explain it right the prompt a certain way include these things exclude these things just just it just gets better in and of itself and then it gets more tailored to you i think it's going to be extraordinary step forward and i think that's that that's fascinating to me and again i may even be hopeful as a google shareholder but i think what's interesting maybe microsoft the same with chat gpt actually is the your kind of life's operating system becomes so much more important whether whether it is windows or what are they what is what's the um apple os called these days i don't know or um or chrome ios yeah i was thank you yeah um that idea of kind of you know it becomes the base layer of your life because you can't do without it through that but if it's going to be integrated with it that's almost the point you know the gemini will be more input will be more useful to me than it would be to an ios user because it doesn't have that that amount of context and information so it kind of to your point at 984 the single agent thing it's incredibly powerful because of that because google knows me it can serve up ads that are better for me it can you know make recommendations on youtube that are better for me it just it just it makes that's what the algorithm it's a horrible thing when it goes bad and we know though that from you know political advertising or or you know the the tuning of it to deliver you crappy social credit scores in certain countries it gets very dark man very quickly and very and very quickly exactly so So I think that is fascinating.

39:03Hey, just quickly, how long ago was NVIDIA's share price one-tenth of today's share price? Oh, like scarily short, two years? Not even. Something like, no? What was it, one year? So we're recording this, I'll give the date away, 7th of June, the share price$132, rounded up. I just scroll back really quickly. The last time, so I went for a price under$13. The first one I can see is October 21, 2022. Wow. Yeah. Things change. No, 20 months. No, 20 months. Yeah. Gradually, then suddenly. It's like this is always how these things go. Ten times in less than two years. Yep. Always. Always how it goes.

39:47Gradually, then suddenly. Nothing, nothing, nothing, nothing. Whoa. You know, it's just every single time. If I was going to another, all navel gazing. Yeah, go for it. I'm going to, we're going to, I'm never going to listen to this ever again. But if I was to listen to this in two years, I reckon I would be so embarrassed by what's coming out of my mouth. I've looked, I actually read some of the stuff I've written on other things before. Oh, I can't do that. Oh, bless your cotton socks, Andrew. Like you were so wrong, you know. And so I preface all of this. But I guess my other gut instinct is that people are too general in their application of the term AI.

40:31We have AI and we've done it already, both you and I. We'll go, oh, chat GPT. That's right. I suspect what we'll do is we'll actually have a multi-agent relationship. You know, I've got maybe my PA, but I'll have an agent that does my tax. I have an agent that helps me with cooking. I have an agent or more specifically, I have an agent that my agent uses. So what I will do is I will say to my assistant, hey, I want to go on a holiday to Southeast Asia. Can you suggest some places? We'll talk it through. Give me some suggestions. I go, yeah, okay, make it happen. Now, it's not going to go and book everything, but it might go to a booking agent who goes, yeah, this is what I do.

41:14I'm super good at this. Okay, cool. um can i get you to book a trip for this this this and this yep okay cool and it's going to contact a hotel agent that's going to do this this and this which will contact a car hire agent like i i suspect we will have they will be more general than what we're used to in terms of software applications but they will still operate in certain domains and you see that now already there are ones that are really good at video there are things that are good at images there are things there are ai that are some of the ai is really bad at maths i've noticed oh right okay When I've tried to get Claude to do some maths for me, it's got a lot of confidence in what it's doing, but it's often wrong.

41:51And it won't know that it's wrong. But there are AI models that are good at or better at maths and not so good at English. You know what I mean? And I suspect that is the relationship that we will have here where it will be bots talking to bots talking to bots to get us down. It's actually, just to take a little bit of a fun segue, it's where things get really dark. is because i think you would say fun segue and then go to which i think it's we think well well you know existential risk yeah there are i mean it's it's it's funny to talk about because it just invites the the terminator analogies and the rest of it but look when you've got i think i heard something like 70 percent of ai research is seriously concerned about something you go like if all the nuclear scientists said oh this is people can make a bomb out of this stuff yeah No, we listened to them, actually.

42:43We took them very seriously. And now the people who are closest to this tech are saying, nothing imminent, but, you know, we might want to put some safety guardrails around this kind of stuff. To be fair, I'm not absent at climate scientists, so we're in trouble too. I know. All that, so, you know. Well, you know, here's a Hollywood formula for you. Whenever the scientist at the start of the movie is going, bad idea listen listen to the person in the lab coat right like listen to them and all of the rest of the movie will be boring but you'll avoid a lot of like when they're saying don't clone the t-rex it's a bad idea we don't need raptors running around right like it's and and there's it's not it's not the guy in the tinfoil hat right it's it's sort of like oh the most yes that's right to the point and and so and and it's not because well it's it's because you just get some really evil people out there he'll go i'm just going to design a bot that it's going to max them like you're i'm going to i want you to optimize for human misery okay there's that possibility but it's as always with this tech it is the unintended consequence we're at a stage now where we don't know we know the the process in which the answer was derived but we can't tell you the exact path and thinking that that took so so what what you have the reason i got onto this is that as soon as you introduce give an agent the ability to create sub goals things get get pretty interesting so for example in my my example um i want to book a trip to hawaii okay cool that's all i'm going to say now to fulfill that the agent needs to come up with its own goals okay well well, I guess I will need to contact this.

44:24I need to think about finances. I'll have to look at that. So it's doing all these other things in pursuit of the master goal, but you never overtly ask for it. Now, don't forget, this is not a human intelligence. So a perfectly rational answer might be, actually, to get Scott to Hawaii, I could get Scott to Hawaii much easier if I controlled the global internet and all of the transport infrastructure around the world and energy production. Now I'm not doing that. I'm not doing that to be evil. It's just that, Oh my God, when I actually plan this stuff out, it's really convoluted and I have to go through all these humans and it's really hard.

45:02And then there's all these bottlenecks. If I take over the planet, I can, I can help Scott get to Hawaii much. So again, there's no nefarious thing here. And it's, it's kind of a fun example, fun example, but, but that's the classic I've spoken about this before too. It's the paperclip apocalypse where, you know, A factory owner makes paperclip said, I want you to maximize paperclip production. Well, if I just eviscerate Brazil so I can get to the steel much easier and I can do it. Again, when we build a highway and we plow through a big ant nest, we've got nothing against the ants. I've got no ill will to, I don't care.

45:43It's just, you're in the way. You're not even a consideration because I need this road to go there and you're in the way. and and there's no council of humans that like let's get the ants the answer you know we're gonna make them suffer we're gonna wipe out their civilization no i just just building a highway and they'll they're they're the the one of the mistakes i guess the argument goes that we tend to think that we're the top of the pyramid yeah that's right because we're the smartest thing that we know of yeah great but as if that's it that's the limit like there's nothing beyond us like yeah yeah there is like they why why would iq top out at 200 or whatever the highest human is that why why can't i have an iq of 2000 if i can have an iq of 2000 why can't i have an iq of 50 000 you know in fact once i get to even 180 well now i'm just going to replicate myself 100 million times and then i'm just going to focus on making myself better and then the smarter version of me will then focus on making an even smarter version of me and you have that they call it the singularity the runaway it's it's it's you know it's where um cyberdine systems you know finally goes aha and and it's so it's it's kind of it's it's um it's something that needs to be discussed without smirking too seriously because it's it it's a problem that happens instantly right and then again gradually and then suddenly yeah that's right and and the pace of the internet the pace of computing yeah is just phenomenally i'm still every now and again i catch myself i ask google for something and gives you the response really quickly yeah and it's i'm still like the ability to process the information in milliseconds something stupidly complex or difficult not even ai just just search but think about ai that's just to put all that together in that time frame is is phenomenal it's not fair it's not fair we don't stand a chance imagine if imagine if for whatever reason you woke up and somehow the chickens had got you in a cage right and you were the chickens prisoner yeah you're going to take half an hour to outthink them and get out of there right and there's the best chicken scientists in the world are never going to come up with a strategy to contain you like because you're not like you're not 2x the chicken's brain you're like 100x it's just like and and that's that's the thing because you you think well oh we'll do this or we'll do that and it's like you have no idea and and it's it's probably it could probably even achieve its ends through social engineering it's like i don't even need robot automaton terminators walking down the street i'll just convince everyone that this is going to happen and the humans will do it for me because they're so easy to can we are so easy politicians know this right we're so super stupidly easy to control my favorite my favorite apocalypse ai scenario isn't the robots it's just it's just basically the ai presenting to you the information it needs to present to you to get you to do what it needs you to do that's the beauty we've built the transmission mechanism it's called social media that's right literally like you think about the political ads okay apparently trump campaign had 10 different types of ads depending on the vote are they different messages sometimes completely contradictory messages to earn your vote now Now imagine you could personalize it per person because AI can now create the language and the pictures.

49:00You get the Andrew Page version of the, it doesn't matter what Donald Trump, whoever, right? Or whatever. You think the Russian hacking and the Russian involvement is bad when the computer side, they want to change the outcome of the election. Now imagine this, to your earlier point. Now your agent that has worked with you for the last three years knows everything. I know what makes Scott tick. and now I need a bespoke strategy to convince him. I was like, it is going to be, anyone who's had kids knows this, right? They reach an age where you can't do it anymore. But when they're little, you can convince them that your thumb just came off, that you pulled a penny from behind the ear.

49:39Like, you know, kids are easy to trick, right? And we're going to be potentially in that camp too. So that's, I just, the world, or we could go into another AI winter. we cap out we plateau at these llms that kind of get a little bit better but no no uh massive surge forward so it's just this is why this is all super interesting but it's also why i think we need to have some of these conversations now just to think through it and the other thing i'll throw it to you for your answer is that we haven't talked about is just the social ramifications when when a computer can do your job much better than you it's like what do we do with employment like I've said before, anyone working in a call center, your job is gone in five years, right?

50:23Anyone who works, who drives for a living, whether it's transport, courier, taxi, Uber, we've already got that. That's just a matter of time before that job is sort of gone. 20 % of jobs involved in transport, right? That's gone. Even if you say, well, we'll always need some humans. Okay. What about the rest? Which ones are the sum of you? I think, you know, that's more - So many directions that go in with this stuff, right? I know. I know I do want to get back to investing, but we'll indulge it for another few minutes and we'll get back to it. One of the things that's fascinating about your question, you mentioned driving, I like it because I was going to make the example of, I think about AI, and this is not nefarious even slightly, it's just the issues of AI to deal with, right?

51:03Tesla's self-driving needs to decide whether you plow into the school children or drive the car off the cliff and kill the driver. There is no option. That's literally the plot of iRobot, essentially. But that's literally, it's going to need to make that decision or non-decision. Again, remember, non-decision is a decision. Choosing not to make a decision is a decision you're abdicating responsibility and therefore. But think about that and think about the, now humans have the same choice, but we don't have to program for how many people, how old are the people, how many years of life do they have left.

51:36And by the way, computer could do that in a split second if it chose to. So it could make the air quotes right decision well and truly before it needs to. uh where you know who's more important is the drive more important because you're their assistant and again you talk about the the not even not even deliberate stuff let alone the deliberate stuff which is just in that circumstance what happens and so when you get back to the jobs part and it goes back to it comes full circle beautifully actually mate because throughout the pace of change as an investor the challenge i i am always optimistic and always be pollyanna right so my my and you're a bit the other way my default response is i expect that we will absorb ai and people will still have the opportunity to be wealthy and meaningfully employed enough because the examples you gave could have been about what happens when we have spinning looms in in warehouses and what happens to the all the horse and buggy drivers when people can drive themselves so there are there are there are huge historical precedents for over the last particularly two and a half centuries roughly almost um of of jobs changing because we have for example more leisure time you know the we we had slaves and then we had maids and then we did it ourselves and we had cleaners again because we chose to afford to employ some people as cleaners we don't want to do it ourselves we had the money um there's there's examples parallels where those jobs come and go it's brand new jobs invented um you know there were there were no there were no masseurs you know 25 years ago 50 years ago 100 years ago other than purely for sport for example or for the kings um you know things become more accessible and achievable and available australia is i don't know what the percentage is probably i think it's more than majority service oriented and service industries oh yeah didn't exist 70 years ago my biggest concern is not the and well we may never get there because the speed is the is the key right so let's assume ai has more leaps of improvement you you hollow out call centers then you hollow out drivers then you hollow out bricklayers because you mentioned fast brick let's go with that it gets good enough to be able to do that itself and these are all different versions of ai um where do you stop right and and at some point maybe hollow out stock picker so you and i have got to go on we'll do the podcast but we won't be employed anymore we'll just do it because we've got nothing else to do with our time while we're sitting on our backsides remembering the good old days when there were when and there were websites and jobs to be had.

53:58But, you know, that's... If you want to be really dystopian, the horror scenario in my version is if it takes long enough, we absorb it. The horror version is if it goes too quickly, then you have 20, 30, 40 % unemployment in 10 years' time. And the problem with that is that not only is it a... And that could be reabsorbed potentially, but society's pretty fragile. And if you break society by having 30 % of people unemployed all of a sudden, now we got through the Great Depression and it was kind of fine, but that was largely not... See, from our vantage point, it was fine. It was just really bad for a lot of people for a long time.

54:37But my point is we got back to something, right? So in a mass unemployment scenario, you can point to it and say, well, okay, sucked, but we got back to 4 % unemployment by 2024. You know, these things improved pretty quickly. Things are okay. But if you break it in a way, that was largely a financial depression. if you break it in a structural way by saying society actually no longer looks and feels that way and the jobs have just gone so it's not that the people were thrown out of those jobs and they never have demand for those jobs you can never go back to those jobs yeah and and the ability for that to break some kind of inherent um what's the right word uh ways of coexisting the societal bonds right that that's that'd be to me that's that's the dystopian outcome bit is happen so quickly so phenomenally that it kind of breaks our relationships with each other we have to kind of reimagine what society itself looks like that that's that's pretty scary um but potentially i mean like again it's maybe not but maybe yeah and it's just i i've i know well that's the thing right when when you first hear about some of these ideas it's just funny right like yeah right but that i i'm just it doesn't i'm not saying that you must automatically fear this and expect it yeah because all kinds of bad there's this great website called the pessimist archive or something like that of all these of all these things that were going to go wrong that didn't go wrong because of new technologies and it's a really good reminder that a lot of the time they don't come to pass yeah i still think there's value in discussing it and preparing for it and thinking it through because when they do change if they do change in certain directions they change pretty fast so you you definitely want to sort of think about it here's something else there's another thing that didn't even occur to me but i heard it on a podcast and it actually reminded me that i was doing this which is the more i use ai the less i visit websites yes it breaks it breaks the monetize so anyone online generally speaking the monetize i want eyeballs on my site so i get ad revenue that's the business model yep that's the business I mean, there are variations on that, particularly for e-commerce and stuff.

56:48We're actually buying stuff. But for a lot of the internet, it's just like, come to my website. I want page views. That's supported, correct. Yep. That's how I get Google AdWords to give me the ads. That's how I get the money. That's how the whole damn thing works. Now, when I say I don't go to websites anymore, it's like when I need to look something up, I just open up my Claude app. Hey, Claude, what's this? Or how do I get that? And just it searches the internet. But it's not distracted by pop-ups. It's not reading the ad. Sorry, I'm just going to go and buy a, yeah. I'm going to go investigate this new one crazy trick to lose belly flab, yeah.

57:22It changes. It changes. So the internet changed the business models of so many businesses. Yeah. And not that the internet itself will be disrupted, but the business model of a big part of the internet potentially is on the verge of being disrupted. Yeah. And then you've got to say, well, it's all great and well that we've got these AI agents figuring all this stuff out, but where are they getting the data from? And where's my incentive to create my blog or maintain Wikipedia or do all this kind of thing when the bots are doing everything? Like where's the source of truth? Who's the auditor of all of this stuff?

57:59Motley Fool Money. For more, subscribe to the free newsletter at fool.com.au forward slash listener.

58:10Just for fun, so I was going to write an article about the Berkshire Hathaway annual meeting. I might have said this before. And I actually wrote it. I wrote my, made my dot points, actually. I thought, I'll go and just ask you to summarize the meeting and just make sure I haven't missed anything kind of any of the big pieces. Yeah. And it said, at the Berkshire Hathaway meeting, Warren Buffett and Charlie Munger talked about this and this and this. Munger had been dead for four months. Yeah. And so it was, and it was, there was no, and I said, did that actually happen? Yes, it did. But Charlie Munger said, oh, you're right.

58:40My apologies. Yes, Charlie Munger did die, but what he would have normally talked about is this. I'm like, that's not what I asked you. I literally asked you what was discussed at the 2024 meeting. And so it's kind of that, that's the, you know, it's a source of truth. That's the other part of the, you know, we think Wikipedia is bad, right? And it's not, by the way, as bad as people think. But you imagine, you know, ChatGP said this, had I not, I mean, again, I would have been fired if I didn't make sure my stuff was based in fact, regardless of what it was. But, you know, had I said, I asked AI, I got this, okay here's the here's the dot points i published it you know it's uh look stupid for me again my stupid fault if i did but there is there is some element of that deliberate or otherwise the so-called hallucinations which again i think i i back to the ai thing it's a consequence of how the tech works again it's not right there's no malfeasance there it's yes except that we are you know what's amazing about tech and and i think it's the early adopter thing i think it's a bit of the x-files i want to believe stuff is the defense of that so on one hand totally reasonable totally okay it's the way it works i know that you know that everyone knows that right so on one hand it's fine on the other hand when we let an ai say this thing is real and then dismiss it with a oh yeah but it you know we don't if if you know if a kid says this is real it's like no you made that up mate it's okay but you made that up and it's not okay you can't tell me that you know your teacher said this or your mate said that or what we're in our house but mum said i could no she didn't that's not okay you know i was like mum said i could no it's all right it's it's cool and it's fun it's interesting and it's new of course gonna make mistakes it's all right and i think i i the the one thing i do think is worth just just having half a half a break on so about the social media stuff it's like uber i mean we we let tech get away with a whole lot of stuff because it's new and hard to regulate and different and whatever and there is something to that it's not it's not a fraud necessarily by the same token if i made those statements of fact i would be and it's not the fact that it's wrong it's the fact that it says it's right that's my issue and so there's there is something to that i don't really know what my point is um it's no point you know yelling at the clouds but there is something about our acceptance of a statement of fact as if it's a statement of fact rather than you know if it i don't know but it might have been this we don't say that it says this happened that's that's my issue that is fraud any in any other format of of communication and yet we don't kind of require it to have those stand that's the bit that you mentioned kind of the the some of the dystopian stuff or some of the you know when the bots just choose to say that and there's no conscience and there's no that's a bit you that's the that's the thin end of the wedge if we're not careful and not not that it's necessarily a slippery slope argument at all it's not but there is some element of that already playing out to your point which is when it chooses to just you know yeah that's an excellent point the social media example is a great one because facebook never set out to be evil yes exactly yes they just said we just want an algorithm that just keeps people on our app like we want people to use our app and they'll use more if we do this yep they do and you know like so it didn't start it wasn't i mean a moment of reflection might give you some pause for thought but but it didn't it wasn't for evil it was for commercial reasons and this is why when you speak of some of the scary stuff it's you've got to remember that the institutions that are pursuing and developing this technology are for-profit commercial operations, you know?

1:02:07And there's a game theory. I know I'm mentioning that a lot lately, but there's a game theoretical element at play here. It's like, no, no, no, move fast and break things. That's the mantra of Silicon Valley. I am not going to sit back and leave this thing in the R &D lab when Google's out there sucking up all the market share. No, no, no, no. Let's just get it out there. We'll figure this out later. It happens all the time. Uber just launched in Australia before they said, can we do it? They just did. And then the regulators scrambled to catch up. You know? And I think... Not an unresomely operator.

1:02:36I'm pretty sure it was illegally operating in the context. It wasn't even like... It was in this grey... It wasn't even grey space. It was like, I wonder if that's okay. It was like, no, no, you literally need a license to do this thing. Well, we're not going to have one. Okay, then you should be shut down and thrown out. No, no, no. Afterpay, same thing, right? Yep. Same thing. Always the same thing. Airbnb. Yeah, exactly. And so that's the thing to be mindful of because the people who really wield the power here are just driven by profit. Again, not to make that a bad thing, but it's just like we've got to – they are not there thinking long-term existential considerations.

1:03:12So it's – yeah. The other thing I wanted to make too is I lament a little bit of this with – I know our kids have discovered it. And, you know, there's impacts here for homework and school and that. And it seems as though the knee-jerk reaction in many quarters has been the wrong one, which is, oh, we've got to ban it. Yes, yes. And I, again, because of my vintage, I remember that. It was, oh, with the internet, everyone's just going to cut and paste and no one's going to learn anything anymore. Could you imagine not teaching a kid how to use the internet now? Like you, like that would be like saying, well, we're not going to teach you edition or something.

1:03:50It's like that is not equipping. Or you can't use encyclopedias. in my day we had a 27 volume uh funken wagnalls encyclopedia you bought week at a time at coals and every assignment i did up until whatever time i did on the internet right through high school was was done using the funken wagnalls i could i copy and paste out of there of course i did sure sure is this any difference because it's new and easy to access no it's same thing yep so i think i think educators need to embrace like the genie's out of the point right it just whether you ban it or not, kids are going to use it. In fact, if they don't use it, they'll be competing against kids who do use it.

1:04:25Does that mean that, oh, shut up, Andrew, every kid can just say, hey, chat GPT, write me an essay on Federation. No, I'm not saying that. But it is a tool and it exists and it will be used competitively out there. And those who know how to use it and have an understanding of it are going to be at an incredible edge to those that don't. It doesn't mean just cheat and outsource everything, but it means know how to use it. So that's yet another complication that we have to contend with. And that's what's so big about this and why it is talked about so much is that it's one of these texts that's so general, so pervasive, so all-encompassing that it's hard to avoid.

1:05:07I'll just end on that, not necessarily, but just on that point of despite all of my optimism, excitement, that. As I said, I'll re-emphasize. I've not made any investment based on that alone. It's too – and what I will also say is that 90 % plus of companies in this space will go to zero, if not 98%. Like, again, that's just – most steam engines companies went to zero. Most automobile companies. I was going to say, 100 % car companies in the US nearly 900s. Yep. And so it's – two things can be true at once. You can have this sector that it's going to grow insanely fast and be all-encompassing and still be a terrible, terrible investment.

1:05:51And I'm already seeing hints of it now where every company is talking – they've got a slide in their pitch deck about AI, how we're using it, how we're different. It's cool. I'm glad they're talking about it. But just don't ascribe too much of a premium to it at this point in time. Don't get caught up on any of that. I don't own NVIDIA shares, right? And I actually think that they're a wonderful business, but I've just got, it's just too hard to value it. And whatever value is imparted by the current prices really accounts for a lot of growth already. So it's just too hard basket for me. So don't feel as though you need to.

1:06:29Warren Buffett avoided technology famously. Yes. Apple is now his largest investment listed. Yes. So, and you know what? That's not very well for him too, by the way. He's done incredibly well, right? And you could have done it earlier. Of course he could have. Did he miss out by not doing it? Sure. doesn't mean you shouldn't have done it later or the only way to do it was to be there at the birth. 100%. So, yeah. I'll add my final thoughts to this, mate, in not too dissimilar a way to you. I... As much as you love keeping up with the tech and you kind of love some of the stories around investing, and normally you're the resident curmudgeon.

1:07:07When it comes to stuff, I'll take that title. I'm just kind of probably jaded is probably the right word in the investment sense that we've seen this movie so many times before and we've referenced some of the number of new things that have come up that have come to nothing or come to something but not in the way and we use the internet as a great example not in the way that you might have envisaged or with the specific value creation you otherwise might have imagined has the internet created enormous value? Yes 100 times yes but have you been able to kind of capture that which company were again in hindsight you can absolutely see which ones did well which ones didn't we went through a dot-com crash when the nasdaq fell 80 percent in the meantime um the again to your point about the wall going to zero i think we don't know how it's going to play out and that's kind of the point you know you there's so much fomo in investing and ai is huge it's going to be huge and i hope we've been pretty clear with our listeners about different ways to think about ai some of the executions some of the risks as well by the way but um the the reality is that the uncertainty in our conversation is exactly why you can have a large too hard pile as an investor oh yeah and not only do you have to know which ai companies are going to do well but also at what price i mentioned video is up tenfold in less than two years now maybe it goes up another tenfold from here or maybe it crashes back and never again hits the heights it was yeah and if you don't know the answer you don't have to play the game this is not a case where you have to say well sorry for your life bet on red or black you say oh that sounds pretty risky i'll just won't play thank you and that's completely okay i i own fortescue shares now i saw something the other day if you'd invested if you put um 10 grand worth of fortescue shares at on the day of listing it's worth like four million dollars now right like oh yeah because it was two cents or whatever whatever it was and so you know do do i regret that i mean you you can't right you really honestly you got to work on yourself here guys because if your if your view is i'm going to beat myself off missing fortescue i was the i was around i was there i mentioned the other day um that almost went broke over a weekend that took you to go and beg the lenders to extend the lines now had that gone the other way it's all over i don't talk about fortescue at all other than the amount of money lost and that bloke andrew um uh what's his name who who did that mining thing and now he's you know one of the richest men in the country and i again could like i'm glad you did good on him i hope particularly he does well from now because I own some shares.

1:09:30But you don't have, you can't, you cannot try to make sure you make every single winner. You just can't do it. You'd lose too much money trying and you'd drive yourself nuts. You're going to miss winners. You're going to miss lots of winners. You're going to miss really big winners. That's just the way the game is played. And so with AI, I'm going to be watching it closely. Andrew will be watching it even closer because he loves the tech. I'm more a consumer user. He's just like, you know, I'll have this thing. I had to explain who Claude was. I know Claude apparently is a new one. There's others out there too.

1:09:56I know about three and Andrew probably knows about 15. So, you know, but the thing is, there's no need to put your chips on anything right now. And as an investor, we talk, let me take it back to Buffett because we're obliged to. Warren Buffett talks about his circle of competence. So he says two things. Firstly, stay within your circle of competence. If you don't know, don't start playing outside that circle of competence because you don't have an edge. You have no, there's no reason for you to rescue your money on something you don't have a strong conviction in the outcome of. Makes no sense.

1:10:24But second thing he says about circle of competence is you should try to widen it. And so for me, for AI, it's like staying across it, being really cognizant of the executions, making sure we see how those things are happening. And even when there is real value being created, you can absolutely go and jump all over it. Cars, 100 car companies, actually turns out this General Motors mob and this Ford mob, they seem to be doing reasonably well. Okay, but actually, okay, well, now's the time to go on 1930, 1940, 1950. Time to go and invest, right? You just don't need to be there at the birth. And if you are, there's a very good chance you're not going to make a lot of money because you're going to pick the wrong ones or spread your money so thinly, you can't possibly be across them or that's the other part of it.

1:11:07Yeah. Yeah. If there hasn't been, I know there is. EDFs come up all the time on the pod, right? So there'll be an AI ETF. And even with my enthusiasm, I don't know if I would because just for that point that you just made there, it'll be spread too thinly. And, yeah, I feel as though this will be a game where a handful potentially win very significantly. And there's just a lot of losers after that. Even if it's something that every company is employing in some way, shape or form. Like email. Like email. It's just such a great example for it. Like no one's got an edge because they use email. And yet no one's getting rid of it.

1:11:45It's vitally important. And you can't invest in email. You can't invest in email. It's not possible. Yep. Yep. So, yeah, I mean, stay across it. I think it's something that, I mean, I remember when I moved out of home, well, we had to get electricity, right? Because I want hot water and, you know, flights. And, you know, to have the internet at home was a luxury. It was. It's not a lot. When you move houses, it's the first thing you get on. I turn the water on. I turn the electricity and gas if it's there. And I turn on the internet. Yep. It's a not. I'm not even in that order. And I'm not, it's not even entertaining the idea that I'm not going to have the, even if I just say, well, I'm going to use the 4G or 5G on my phone, right?

1:12:34It is a must. I make sure I have access to it. Yeah. And I suspect, you know, this going away was just sort of like, you don't have AI. It won't even, it won't even be a thing. It'll just be like part of the background of your life that it'll just be like a calculator is now. And like, And our kids won't be that impressed by it because it's just like, yeah, we've always had that. What's the big deal? So it's, well, just stay on top of it is all really. And that doesn't mean reading research papers or anything like that. It's like just following some interesting follows on Twitter is get you a long way or some good podcasts that are out there or read the odd thing.

1:13:12And I think you should do it because it's just interesting, right? But I think they are the people that really do have a bit of an edge because the future isn't distributed evenly. And you can be in the 30 % to 40 % bracket of people who discover it and realize its significance and still be ahead of vast majority of people without having to be on the cutting line. And who is an investor, particularly one that's got other jobs and priorities and stuff in life, can stay across every breakthrough that's sort of happening in the world of biochemistry, computer science, engineering. It's too much for any one brain.

1:13:51But you can go a long way, I think, just keeping your toe in the water with this kind of stuff. And I'm here for it. It's an exciting slash terrifying time to be alive. I've got one more quick one, mate, which is just a bit of a watch out, which is you mentioned every company. is mentioning AI right now. Five years ago, everyone was talking about blockchain. And again, what you think of Bitcoin, what happens next to Bitcoin, is not really, not at all the point, actually, in this context. It is just throwing it into a press release. And it'll be something after AI. Everyone's going to have to cover it because, frankly, and this is the other thing, by the way, people get over-obsessed with it.

1:14:26So imagine we spend a whole episode talking about AI, right? It's probably not worth 150 tooth, that's not really a thing, 150 seconds of our Friday podcast this year, right? Because it's not that important at this point. given what we know and given the amount of time there is to go to make it worthwhile. But every analyst on the call wants to know. Why? Well, because firstly, they get excited about it. Secondly, their clients want to read about it. And the company wants to look like they're on the cutting edge. So everyone is, and there's no grand conspiracy, by the way. Everyone just knows it's in their interest to mention AI in passing so that the conversation has been had so that everyone feels like they're cool and trendy and up-to-date and doing things and not getting left behind and all those things.

1:15:02And they should be, by the way, companies should be aware of AI. But don't – just be really careful of buzzwords. We'll talk about maybe meme stocks another day. But just be careful of buzzwords in presentations because everyone mentioned blockchain. Everyone mentioned.com. Everyone was doing a.com, this is a.com, that. Everyone is now talking about AI. Not that they're not a thing. It will be a thing. But it's not a differentiator for most companies. And including it is not even because it's a central part of their strategy. Just because they have to talk about it because they have to talk about it.

1:15:31But frankly, if Woolies is too excited about AI, that's a red flag for me rather than a green flag. It's like, just price the groceries well, get the Senate off your back, get customers coming in the door, grow your sales and profits. If you can use it to do it, sure, but don't tell me about it at the press release. That's nowhere near, it's not the 50th priority you've got. Like, yeah, do your job, do your thing. So, you know, just be careful with companies being too focused on that and making it too big a deal either. Don't get fooled yourself and just be careful. They're not getting carried away either.

1:16:01Yeah, I mean, it's just a tool, right? Woolies doesn't have anything in its press releases about electricity. Exactly. But they use it. Yes, yes. You know. A new air conditioner they've installed in the Campbelltown store in Sydney or something. Yeah, go. Yeah, okay, great. I spoke to, I don't own any shares. I'm not spruiking in any way, shape or form. But we spoke to the CEO of Frontier Digital Ventures. They kind of run the car sales in the REA group of Southeast Asia and stuff. Yeah. Yeah, anyway, Sean DiGregorio, the CEO and founder, we're having a chat to him. This is an online business.

1:16:38And I said to him at one point, I was like, I could only find one little passing mention of AI, but it kind of behooves me to ask because everyone's talking about it. And he kind of went, well, you know, we've got our eye on it, but really it's not something that's front and center for us. And he really downplayed it. And honestly, to your point, it kind of, it was a tick in the positive column because here is a company that if anyone could talk about the potential of AI, these guys couldn't. They got lots of user data there. They got lots of traffic and analytics that an AI model could suck up and that he could spin into something that gives them a competitive edge and run.

1:17:17And he didn't. But it wasn't like, oh, it's all non-the-art. It wasn't Jerry Harvey to pick on poor Jerry. You know, oh, bah humbug, the internet, no one will ever use it. It was just, it was a very reasoned, sensibilized, like, yeah, we've got an iron. It's got potential. Certainly, we're actually using it a little bit here with some of the marketing copy to sort of help with this and that and rah, rah. But, you know, it is what we're just at this stage yet. And I think that to me shows someone who's got a very level-headed, pragmatic view of things. And not, these are rules of thumb, but if ever I hear a CEO really laying it on thick and they're just lots of buzzwords.

1:17:53I run a mile. I run a mile I remember again you would too the year 2000 every company added dot com to their name they all changed their name why well because when you did that you got a 20 % pop on your share price so you're gonna do it wow that's amazing yeah exactly yeah I was like oh such and such dot com and I had to have some stupid name you know and it was just it was ridiculous but it worked and eventually that insanity comes to an end but it runs pretty rampant for a while So I'd be very much bracing for that to come onto the ASA. It's already here. I think it's only going to get worse because you will find that some companies legitimately come and go, look what we're doing.

1:18:36Like, whoa, they're doing, we're doing that too. And then it's just the halo effect is going to be very difficult to ignore. So just, again, tread very carefully out there. I like it. And with that, we will let our AI overlords take over and hopefully either make this podcast appear on the feed or if you never hear from us again, the robots have got us and they won. So we'll see. But if you are listening to this, thank you for spending a bit of time with us. We hope it's been useful. Just dissecting a little bit of what's happening in the AI landscape. We are not experts. We don't pretend to have either technical expertise nor perfect foresight.

1:19:10We're just giving you our thoughts about where things are up to, what's exciting us, what's worrying us and maybe some things to think about when it comes to what the future might look like. All right, well, that's it for us. Mate, will you come back on Sunday? Yeah, you know I will. Well, I do. I think I already have. Yeah, you already have been here. But it hasn't yet made it out into the interwebs. So when it does, we will talk to you then. Until then, enjoy the first half of your weekend. And Fool on. See you later. The Motley Fool and people appearing in this program may have positions in the companies mentioned.

1:19:40General advice only. Please speak to your financial professional to understand how it may pertain to your situation. Subscribe to the free newsletter at fool.com.au forward slash listener. The Motley Fool operates under Financial Services Licence 400691.

From the publisher

– How should investors be thinking about AI? Where are the winners and losers likely to come from… or can we know at all?

See omnystudio.com/listener for privacy information.

Ask about this episode

In short

Transcript

More from Motley Fool Money