Cash Lads
Welcome to Cash Lads - a show that demystifies finance and money with plenty of stories, history, and banter!
Hosts:
Paul Molloy - https://www.boxclevertax.ie/
Marcus Doyle - https://clearfinancial.ie/
Warning: This podcast does not constitute financial or tax advice. Please contact a financial advisor or tax advisor to discuss your own individual circumstances, taking into account your needs and objectives, knowledge and experience and financial situation.
Warning: The value of your investment may go down as well as up.
Warning: If you invest in this product, you may lose some or all of the money you invest.
Warning: This product may be affected by changes in currency exchange rates.
Warning: Past performance is not a reliable guide to future performance.
Warning: The prices quoted assume standard rates and may be subject to change or special terms pending underwriting of a proposal by the life company.
MB Planning Ltd. T/A Clear Financial is regulated by the Central Bank of Ireland. Registered in Ireland No. 494837 Registered Address: Rathcoole Premier Office Centre, Main Street, Rathcoole, Co Dublin. Directors: Michael Bradley (MD), Michelle Farrell.
Cash Lads
The Artificial Intelligence Takeover?
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Today Marcus and I look at AI and the constraints the industry faces as we all jump on the bandwagon. Not everyone loves AI and we look at some of the impacts it could have on employment and even on our existence!
Warning: This podcast does not constitute financial or tax advice. Please contact a financial advisor or tax advisor to discuss your own individual circumstances, taking into account your needs and objectives, knowledge and experience and financial situation.
Welcome to Cash Lads. My name is Paul Malloy from BoxcoverTax.aue and this is Marcus Doyle from Clear Financial. We're here to take the mystery out of money and finance with some history, some stories, and plenty of humor too.
SPEAKER_01Welcome along to our our show today. You doing well? I'm doing pretty good, yeah. How are the Kildare footballers doing? Are they gone yet? Not yet. I think by the time this show comes goes out, the Dolph footballers will be gone too. But uh next up, Kevin, Brefney Park, terrified to even look. Well at least we've been in the doldrums for a long time. You you've actually been quite good for the last few years. Yeah, yeah. Well, today we're going to talk about uh AI or artificial artificial intelligence. Uh you know, no matter who you are, anywhere, you've been hearing about this, you know, uh for for for you know and the intensity of it, you know, it's being mentioned almost all the time. Like uh billions and billions being invested in it. So just to be uh a courtesy of a course I did recently in UCD, I had thought, Marcus, that what AI was doing was it was it was you know when you put something in like say, can you tell me about you know how the Dublin football team fared in the last four years? And it comes back with something incredibly accurate. Yeah, each year details. So what I thought it was doing, I thought it was scraping the internet for every website you could find, and that that's what it was doing. And it was just doing it really, really quickly. Which is quite logical. Yeah. That is really not what it's doing. It is doing a little bit of that, but not much. What AI actually, what they do, you know, anthropic or open AI, they they they build these AI models and they essentially send them to school for about six, nine months. They teach them how to think, how to react, how to answer questions, what type of questions to anticipate, they give them boundaries and all that. Like I said, a really intense period. So it's skewered in the human direction as well. Very much so. Very much so. And obviously, look, it's gonna pour things like databases downwards and things like that. It knows it's gonna get asked history questions and things like that. So it pours all this stuff in. And then it just it just lets it off. And it's why sometimes, if you want something new, uh you know, recent information, you know, tell me how in detail how Dublin have fared in the last year, say. Not as good at that. It struggles because it doesn't, it hasn't learnt as much. So better at the old stuff than the new stuff. But anyway, it's it's it's still quite impressive. And I think it's with Calderdo rubbish, I think, for the last many years. But it it it is really quite something, and you know, I I recently uh my my partner Gronia um I I I put something in about you know the the film industry because Gronia's the director of film festival on in February, folks, please go. Um and you know, very specific to to her in business and industry, and she she just couldn't believe what came back. You know, because when we put in a Google search, all Google is doing is saying these are the most commonly searched, this is what other people have done. Yeah, this is what they've looked for, and that's the most popular one. So this is this is quite quite quite different, quite quite different.
SPEAKER_02Whereas I'd say people would have thought Google was kind of AI, that it was searching what it thought was the best, but it's actually more popularity than anything else.
SPEAKER_01Yeah, yeah. So I think Google says, oh god, is it Gemini? Um the Google is Gemini, yeah. So when you put in a Google search now, now very often you're getting that that kind of AI prompt that's that is absolutely its own.
SPEAKER_02Yeah, that's it. Once you see it, that it starts nearly typing it out.
SPEAKER_01You know, some people are like, I don't want that at all, you know, and they they skip down to that. But it's it's two very different ways of of of of getting at it. And the thing of it is if you put a some search like that into AI, a prompt. So a prompt is where you you kind of said, you know, tell me about the Kildare football team over the last five years, or you know, we're on the topic of football, so it's just in my head. Um there is a cost to to do that, and the the AI is is is is is gobbling up what are known as tokens. These like little micro tokens, when it spits out content, there is a cost for the company that's providing that. Electricity, servers, all the codes, water to co to cool down. The uh so there is a they go into school, people have to do that, they have to be paid. Yeah. The the the weekly token consumption between January and March of this year, so that's just two months apart, quadrupled. Wow. In terms of the AI use. So the worldwide AI was being used four times as much. Um Palmoloi's massive shares in AI went quadrupled. Well, I mean, uh we don't know who's gonna win that race. Yeah, you know, so so ultimately in terms of search, I think it's fair to say, now maybe this is Irish biased, but I think that Google won that race, you know, in in the way I I was at kind of an investment seminar, and it was kind of like it's been compared to the kind of smartphone.
SPEAKER_02Whoever gets there first is gonna reap the words. So obviously, Apple won that race, they reap the rewards at the starts. And the whole I think at the investment seminar I was at was like, well, the first company to make it there, but then all the other companies invested so much billions and billions in it. Are they gonna suffer big time? So I think they've predicted that okay, there's still a lot of billions being invested here, but it can't go on forever, and some people are gonna lose out with all this investment and not have it reap any rewards for.
SPEAKER_01We don't know who's gonna win. You know, so I remember reading, uh listening to Warren Buffett uh recently, and uh he was talking about the airline industry, you know, and he said clearly this was a technology that people needed, people wanted, um, it changed lives, it had a huge, huge, huge impact on on us. Make very, very little money, you know. But a lot of money's a lot of airline companies made very, very little. Seb Rainer. Seb Rainer. Yeah, yeah. They've just found the right model of you know, uh their hostility to their customers is always made, always made. We keep flying with them. What yeah, because the f the service is very good and it's much more expensive what's put somewhere else. Yeah. So um, so that is an example of an industry where there were very, very few winners. The the railway industry, there was some big losses in the railway industry. Uh, I mean the 19th century, the money was poured into, particularly, you know, America, Argentina, and places like this. Well, there's sending the rail everywhere. Sending just over over investment, you know. So so we we don't know who's going to win the AI race, who is going to come out with the technology that's going to be better um than than anybody else. But there it's it's very costly to you know, there's a lot of barriers to entry. You can't just say I'm you know, I'm gonna build an AI machine. You can't do that. Like you need serious money, money behind you. Um and this is what the chips as well.
SPEAKER_02Those famous chips.
SPEAKER_01The chips, the chips, absolutely. Um, you know, not not not potato chips now. No, not potato chips, no, micro chips. No, no, no. But there's a specific kind of AI chip that is made particularly well by NVIDIA. Um, it's one of the world's largest companies, a company that I still think the average person hasn't even heard of. Um they're they're like if you if you think of the guys in the in the in the California gold rush near San Francisco, um, it was the guys at the bottom selling shovels and sticks and pikes, you know, they're the ones that made the money. Not the poor lads that went went up up looking for gold.
SPEAKER_02So that that's kind of NVIDIA and TSMC, which is another company I think very few people know about, but it is they're they're they're supplying all the chips, so even if they're gonna get their money anyway, unless they're old unless they're owed money from some of these companies and they can't pay them back.
SPEAKER_01Well, they they're going to the the these two these two companies, especially C SMC are in Taiwan, so you know Xi Jinping is very keen to get closer to Taiwan. And if you look at some of the recent statements by uh Heggseth, you know, the American Secretary of Defence, I think they are going to cut Taiwan loose very soon. Oh God, leave them to the Chinese. Yeah. Oh dear. I think so. Now I was also reading that we're we're on a little bit of we're skewing a little bit, folks, but I was reading that Taiwan, uh it's quite a big island. If you look it up, uh it's it's no small island. But it seems small beside China, but I think it doesn't have if if if you look at the the distance between China, the direct straight line between China and Taiwan, it doesn't have like a deep port or deep water. It's actually very, very shallow in in Taiwan, which actually is going to make that very difficult for Chinese ships to come in and do it. So I you know, and there would be a mass of drones ready to blast the head off all the Chinese troops. So I think Taiwan could be tougher than than Ji uh realizes. He's a smart man, so I'd say he realized it. But I I do think the Americans have cut the body. They'll probably build a bridge like they always do Chinese, they'll build bridges everywhere. They might see that, yeah. Yeah, building on the Maldives when I was there. Yeah, there was a uh the Persian Emperor tried to do the same against the Greeks uh two and a half thousand years ago. It didn't work out well. But um anyway, so so you were mentioning just you know some of the some of the chips, and you know, these memory chips, CPUs, they're in short supply, you know. So the companies that are making chips that are good enough for AI are good enough for you know the gear, you know, the switch gears, transformers, all the stuff that's needed. You know, you can't you don't just say, you know, here's a data center and just you know put one lad in charge and you know pay service there you go give them cigarettes for the AI. You need there's a lot, a lot of investment that goes into it. And there are five what are known as hyperscalers, which is Amazon, Google, Meta, um, Microsoft, and Oracle. And in the next 12 months, according to the economist, they are expecting to spend $750 billion on cap in in capital expenditure. So so there is a major constraint in terms of the data centers, in terms of the chips, in terms of you know, um uh access to quality people. Um, you know, you need land for data centers. And there's already been protests in Ireland.
SPEAKER_02We're kind of perfect now because uh uh temperate, cool climate and uh a good electricity supply.
SPEAKER_01Yeah, yeah. I want I mean this is one of the things that we were talking about beforehand, uh, was about Greenland. You know, that we've I suppose anybody who's ever thought about Greenland just said it's just a big lump of rock. With loads of ice on it, loads of ice in it, um, some seals, not many people. Um the edges, aren't they all around the edges? Who would who would want that?
SPEAKER_02But wouldn't that be a great place to put some baby centers? True. Maybe maybe Trump's pulling a fast one saying he wants it for military purposes, but maybe he wants it for other purposes.
SPEAKER_01He might have better reasons than we realise. And and there's uh a word you're gonna hear more and more. It's called rare earth minerals. And these are not the typical iron. It's not the stuff you wash yourself in, but it's not no. No, no, I don't uh I don't have that kind of you get from the special shops. This would be yeah, special shop, you could go anywhere now with that. But um so far at times. So um, you know, these rare earth minerals, China has a lot of them. Has a lot of them. And you know, they go into kind of electrical vehicles, they go into the solar, these are little tiny pieces of metal that the world hasn't got a lot of. Like the lithium that's hard to get and expensive to get out. Yep, hard to get hard to get out. And so not only the China have a lot of these, but they also uh have have done deals and you know built plants in Africa where a lot of them are are because they're building loads of trains and railroads and like that.
SPEAKER_02And I was like, oh, they're building them for free.
SPEAKER_01What are they getting out of it? So that's it's the rare earth metals they're getting. So we uh we have you know red hot demand and we have super uh you know supply constraints. It's very, very few people. There, there's other people doing AI chips compared to NVIDIA and um D SMNC, but they're not as good. Not as good at all. So um prices can only go up realistically, you know, because you know, like I said, from January to March, just quadrupled in two months. I just can't get my head around it. That kind of those spending. Then we'll have the terminator is coming. Well, for those of you who haven't seen the terminator, and I Marcus and I really quite shocked to learn that our our our sound engineer Erica has been too busy not to watch the terminator. Hasn't seen the terminator.
SPEAKER_02But for for for those of you maybe after number two, you don't have to watch them, but first two was even was where a film was better than the first film. What a classic.
SPEAKER_01If you haven't seen them, um and or if it's been a while, the this is the the the the takeover of of the world by machines. Skynet. Skynet, yeah. So it's a a piece of technology um became self-aware and has realized that humans are its enemy, and takes control of the nuclear weapons of Russia and uh America and fires them at each other, and you have what's called in the film Judge Judgment Day. And are we are we getting to the point where that could you know could be a thing? And like I said, I went on this course with uh with UCD for a week, and I think it's more likely having gone there. I now think it's more likely than I did at the beginning because it you know, we have um I think it's and Anthropic has released a a particularly powerful piece of AI technology called Mythos that it restricted access to because the technology was so good at coding and breaking uh you know, you know, yeah, they thought it could be hacking government systems exactly hacking it. And apparently they have expanded access to that. So having been very, very sensible, they're now you know, like tempted by the cash or have been given enough assurances. But you know, how do we know this technology isn't going to go out of control? Yeah, if it thinks for itself. We don't know that.
SPEAKER_02And I can think faster than we can and all this. That was always the thing with computers and when we were younger, they were they were getting smarter and smarter, but it was still we had to give them the code, they were totally dependent on us. Um but now if they can do it all themselves and improve themselves. Yeah, we're kind of we're we're kind of it's kind of survival of the fittest. Darwin's Darwin's uh theory of evolution's gone gone on to techno onto technology next.
SPEAKER_01There was you know, there was a recent um there was a recent uh uh speech at a like you know people getting their diplomas, you know, one of those events on stage. You'd see you'd see this online, it'll take you a second to find it on YouTube. And the lady uh giving speech was was kind of saying that you know AI was fantastic, and she was saying brilliant is going to create lots of opportunities. And then you can very, very quickly hear booing in the crowd. Yeah, which is spies. You think the younger generation would be all for AI. Yeah. You would think, but I I I think, you know, um, you know, I think Erica was telling us beforehand that you know that people coming out of computer science degrees, which I mean I I know for a fact I'm not smart enough to do a computer science degree, um, and the jobs aren't there.
SPEAKER_02Which I think you would be there.
SPEAKER_01You would think that they would be.
SPEAKER_02That was that was a few years ago, it was the thing they were saying, get get kids into coding because the next thing is coding, but now is AI gonna take all that over.
SPEAKER_01I would have thought that if you got a software degree, you could gr you could pivot to AI quite quickly. And I think that's what some of these students are gonna have to do. That you know, it's it's we we can't get away from the fact that this thing is here and it is I mean, I I was saying, I think, Marcus, that I'm I'm a tax consultant myself and I work with accountants quite a lot, and I have noticed an awful lot less what I would call easy queries coming in, the kind of things like I never you know, I know I know straight away without needing to check uh the legislation, they're falling away, and I think what's happening is the accounts are finding stuff on.
SPEAKER_02Yeah, but I think the problem is I I have it with myself with clients, they're obviously putting the query into AI first and then they're coming back to me. Like I um like I don't one client, he obviously put it into AI to AI, yeah. And he was kind of saying to uh saying to me, Oh, this is the way the markets should have been for last but but I'd kind of well that's AI pumping that out, but that's if you're in high-risk stocks for all that period, but I said, but you weren't in the high risk, that's the kind of maximum you could have got over those years, but obviously you weren't in those, so that's why AI is pumping out a general figure. Well, if you're in all the top stocks for the last whatever it was, 10 years, you would have got whatever, whatever pumped out 10% return. But you aren't in the top stocks, the risky stocks, so that that's why that didn't happen. So that's a generic answer from AI, but it doesn't take into account all you've diversified it to this, that, and the other. So yeah, that's uh I I get people they've done their research through AI when they come to see me, but then it's it's not exactly correct. It's kind of taking the optimal, what they think it should be, but that it doesn't take into account like your spread on investments and things like that.
SPEAKER_01No, and and if people don't know what's wrong necessarily, as in they're not able to see what's wrong. Yeah, but I would also say, and I think it just it's just it's just wrong not to say this, but it's getting this stuff right much, much more than it used to, and it's getting much, much better at giving answers. And I I teach in a college in in town, ICD business school with Jason and Healy, and uh, you know, we really can't give essays as as assignments anymore. So they're putting the true AI. Um so we're having to go to more we actually did an oral test uh where we gave the students had to had to you know starting. Open book oral tests. Well, they literally were just sitting there with nothing in front of them. All right, and I was asking them questions, and uh Jason was a lot of a lot of fear in the room. God love them.
SPEAKER_02One of our teachers is that in secondary school, yeah. Go around you like, hopefully, hopefully it was physics, hopefully I get an easier question. But you and you know you're next because he's going around the room. Very good though. People have to know their stuff.
SPEAKER_01Well, look, that's maybe you know, uh people like you said are are buying things into AI, but they but they're not reading what they're not remembering what they've found. So they're not they're not learning what they found. They have found the information, but it's not gone into the I know what you're saying.
SPEAKER_02The problem is people are now dependent on AI, whereas we had to learn stuff from we're ago. Yeah, yeah, and now they're just getting all the answers given to them, so they're not and they're not, or they're not doing problem solving because they're getting the answer themselves from AI.
SPEAKER_01Sweden used to have the highest literacy rates in the world, kind of associate you know, the the Nordic countries with that. It's really fallen away the last five years. Really? They've really fallen down, yeah. And they blame too much technology, exactly. So they're they're taking not entirely, but they're looking to take the technology out of the classroom most of the time, and that the students would just be reading again, uh, and things like that. So, you know, jobs like sales come to mind. I mean, if if if you know, mo a lot of people still want to deal with a human being, you know. So you need a human for that. So finance, your your your job, Mark.
SPEAKER_02Well, yeah, I've been that seven hours before, and it especially it was said that Japan who love technology, they always like taking advice from a human for finance. It's it goes back to the cultural thing you're on about and things and things like that. That's still so hopefully my job is safe. But I did do a bit of computer science very poorly in my science degree in college, so maybe I can switch to AI if if needs be.
SPEAKER_01I think the reality is we're all gonna have to get used to to AI. Uh, I think anybody who starts using it, um I I find it very, very good for research because it it reminds me of things I may have forgotten about as opposed to things that I I you know didn't utter, you know. But you know, it's it's it's here to stay. I think it's gonna be as you say an assistant tool. It's a t it's it's a tool, it's an important tool. And I I I whilst I can understand people being nervous about it and and finding it very, very concerning that it's going to take away their job, I think the thing is to uh take the fear away, start using it, get good at it, and then employers, if they hear that you're good at it, will be more interested in talking to you.
SPEAKER_02But if the if the T2000 turns up outside your door, the radio's gone bad.
SPEAKER_01That it well then then you're just cooked, yeah. Yeah, literally. Um, but on that cheery note, uh let's uh wrap for the day. Uh so thanks very much, Marcus, and uh we'll uh we'll talk to you next uh next time, folks. Cheers. Talk to you soon.