Dire Straits

---

The Money Café is proudly brought to you by Intelligent Investor, Australia's home of value investing. For a limited time, you can get Alan Kohler's Weekend Brief, in-depth economic analysis, CEO interviews, the latest market insights, a limited edition Money Café cap and more for just $297 dollars. Find out more at moneycafe.com.au.

[Music]

Hello, I'm Alan Kohler, Editor-at-Large of Intelligent Investor and the Finance Presenter and Columnist for the ABC.

And I'm James Thomson, Senior Chanticleer Columnist at The Financial Review.

And we are The Money Café.

Yay.

G'day, James. Well, we've still got the Strait of Hormuz closed, they're not even talking - well, Iran won't talk. Where do you think we're at now?

No pun intended, but we're in dire straits. We're two months into the war now, the two-month anniversary was yesterday. If you cast your mind back to that time, we were told by all the experts, if this isn't over in four weeks, this is a really dire position for the global economy. Well, we're now eight weeks in, the strait is completely shut or only a trickle of ships are going through. Just to put it in context, 135 tankers used to go through the Strait of Hormuz every day. I think we're down to less than 10 getting through, so that is a pretty dire situation. The world has done a reasonable job of adjusting to this, but the adjustment buffers are now running out, there's no other way to put it.

Donald Trump might feel he's got time to let this play out because of America's vast oil production and oil reserves, but gasoline prices keep creeping up in America and the pain is spreading through Asia, it's spreading into Africa. You've now got Europe talking about jet fuel rationing and other rationing. Australia's done a good job of getting around all this but this is the time where the crisis is going to start to bite, I think.

I reckon the core problem is that they're supposedly at war, but they're not actually causing each other much pain at all. Iran obviously has taken the Strait of Hormuz hostage or it's closed it to the extent that it can, but that doesn't affect America at all because it doesn't actually need oil from the Middle East; and the American siege of Iranian ports is not that painful for Iran, it's not sending Iran shovelling back to Islamabad to do talks, it doesn't care, it's fine. The problem is, the whole world is feeling the pain, except America and Iran is not and they're the ones making the decisions. I think for that reason, this could go on. The only way it won't is if America starts inflicting more pain directly on Iran by bombing them or something.

That would be deeply unpopular for Trump at home as well.

Yeah. That didn't work at first anyway. Sure, they killed the Ayatollah and quite a few of his family and leaders of Iran. Trump declared victory I think after a week and said they've killed everybody and said their military was completely destroyed, well that's clearly not the case.

Yeah.

I think this could last a while and the problem is, of course, it doesn't just affect oil, it's affecting plastics, fertiliser, LNG, lots of things... Prices everywhere are going up. I spoke to a big food manufacturer the other day, the caps on his jars of peanut butter have just gone up 40 per cent, that's going to get passed on.

I spoke to someone very senior in Australian oil and gas and who's traded petrochemicals up in Singapore. Their point is that most people have no idea of what's going on in that petrochemical space and your example is a perfect one, but there are price hikes of up to 80 per cent in some of these lesser known oil by-products and that will flow through to inflation. What strikes me, Alan, is the Australian Government's done a good job, from what I can see, or securing supply, albeit at higher prices. In doing that and in cutting the fuel excise, we're not sending any signal to people that, hey, fuel is a precious resource that is going to be in short supply in a very, very short period of time. Certainly in Australia, we haven't seen anything like the demand destruction that we're going to need as this war continues to roll on. Even if it finishes, it's going to be 6 months of reconstruction in the Middle East to get things back to normal. I think we're just totally underestimating how big this pain could be.

I spoke to Paul Bloxham of HSBC yesterday and he said that there will be a downturn, that the second quarter that we're in now will be negative GDP and he's not sure if there'll be a recession or not but there'll definitely have to be a downturn in order to get inflation back to target of some sort.

The RBA's in a seriously tough spot, I think. We came into this with inflation too high. Clearly, the inflation that you're seeing in everything from fuel prices to the lids on peanut butter jars, that's coming through next. Does the RBA have to really crush growth to prevent, not only inflation getting out of control, but inflation expectations getting out of control? I think that's what Bloxham's sort of saying, isn't it, Alan? Is the only way out of the recession via a growth shock or a recession via the RBA causing the economy to slowdown to tackle inflation?

Well, we haven't had stagflation for a long time, certainly not a supply shortage stagflation and that was obviously the 1970s and what happened then was that at first, the bloke who was running the Fed, Arthur Burns - in 1973 when the oil prices quadrupled, he didn't do anything, he basically was focusing on employment because he could see that the increase in the oil price was going to affect the economy, going to slow it down and increase unemployment. He basically let it go and that got embedded then in the expectations.

By the time that Paul Volcker was appointed Fed Chair in July, 1979, which was just after the Iranian revolution and around about the time the oil price doubled again, he had to start increasing interest rates, which he did in August, a month after he started in the job. Obviously, he increased interest rates to 18 per cent and caused massive recession in order to deal with the inflation. The thing was, what he had to do was reduce demand to the level that was appropriate for supply and that's kind of what you're talking about the RBA is now faced with. Do we have to retrench demand to meet supply or can we kind of get by?

Yeah, I've got no idea, but I wonder if we'll look back at the last two months and think, policies that allowed us to just burn through petrol at the same pre-war rate, that wasn't sensible.

No - and the other thing of course, is we're in the final two weeks of the budget preparation.

Yes, you wrote a great column on this, Alan.

I wrote about that on Monday in the ABC, saying, "What are they going to do is the Strait of Hormuz is still closed on May 12^th and we don't know what's going on?" Which is looking likely.

Totally, it's looking almost certain, isn't it? There's not going to be a resolution to this.

Even if it's just been opened by the time the budget's brought down, no one's really going to know how it's going to unfold. They're not going to know what inflation to put into the budget papers, what nominal GDP to plug in, what oil price to plug in, they're not going to know anything. I think that it means that this budget is probably going to be irrelevant. They'll have to do another budget as soon as they know what's going on.

Yeah, it's a great question. But the speculation about all these budget moves and I know we've got a lot of questions on all these balloons that are being floated about CGT and tax breaks on this, tax cuts there and tax discounts there. It does feel like the Government wants to move ahead with some of that stuff, but you're right, how do they put it in context? It's very difficult.

You're right. There's a structural deficit of $35 billion dollars which I've been reading various economist predictions and they're all saying that even though there'll be more revenue from gas LNG exports, that's going to get spent on cost of living measures and energy security and also defence spending and so on. No one's actually predicting that they're going to deal with the structural deficit at all, which is amazing, really.

And back in the real world, Alan, it's not just Jim Chalmers and the team inside Treasury that's worried about this, you can feel Australian investors starting to get worried about this too. The ASX has been down for six days in a row. We're clearly losing steam compared to overseas markets. You look at the last month, obviously on Wall Street the market's up about 13 per cent there, but you look at places like Taiwan and Korea, these hot AI trades, they're up 30 per cent. The poor old ASX suddenly feels like we're being left behind. Part of that is flows, but I think the other part is this domestic economy story. People are starting to realise, we're pretty exposed to this, it's not good for a domestic cyclical company. Your food manufacture again is a great example.

Can they pass through a 40 per cent increase in the jar lids, I don't know. Some companies will be able to, but others just won't. There's a margin squeeze coming, so we're starting to see that play out on the ASX. Meanwhile, Wall Street just looks like a better place to put your money at the moment.

Well, that's because of AI and we don't do AI in this country, apart from just buying it, which we're doing in spades. Speaking of AI, you wrote about AI the other day saying it's one of the two risks with the market and then it was instant gratification this morning with OpenAI having problems. Tell us about what's being revealed?

The Wall Street Journal had a story out late on Monday night in New York on OpenAI missing some revenue targets and some user targets and some executives inside OpenAI wondering if there's going to be enough financial fire power that the company can generate to pay for all these deals they've made. You'll remember last year they made something like $650 US billion dollars' worth of deals to buy computing power between now and 2030, with everyone, with Oracle and CoreWeave and Amazon and Microsoft... It was this idea of the sort of circular AI economy, where everybody had a link to everybody else and OpenAI was at the middle of it. We had this scare, this report from the Wall Street Journal, OpenAI said, "Don't worry, everything's fine." But you can see how these problems can cascade. We saw Oracle's shares were off 4 per cent, OpenAI's got a big deal with Oracle. Over in Tokyo, SoftBank, which is a big investor at OpenAI, they were down 10 per cent yesterday on the basis of this story and companies like CoreWeave and Nvidia, they were all off between 4 per cent and 5 per cent on this news.

The AI build out, I think we've seen the AI trade change a little bit in the last month or so, where the focus is not on the hyper-scalers like Meta platform, Microsoft, Amazon, Alphabet, and it's more on the chip makers and the other providers of data centre tools and equipment and infrastructure. Like in Switzerland, a company like ABB that makes electric drives, that's been going gangbusters because those electric drives are needed in data centres. The bet is on AI infrastructure more than AI users or AI model sellers, but I think this little hiccup at OpenAI is just a reminder of how integrated this whole sector is. Everybody's reliant on everyone else and you get one domino sort of starts to waiver, you could see a bit of ugliness.

It's an interesting little moment given how hard particularly computer chip stocks have run. I don't know if you watch the Philadelphia Semiconductor Index, Alan, it was up for 18 straight days, it was a record and it added 45 per cent during that period. It was just stunning numbers.

I'm a very close watcher of that, James.

The SOX Index, yes.

The SOX Index has gone vertical, it's incredible. The Philadelphia Semiconductor Index, everybody, check it out, it's incredible, just google it and have a look at a graph of it, it's amazing. I interviewed Nick Griffin last week, one of the Founders of Munro Partners...

I did a panel with him yesterday actually, Alan.

Oh did you? He's good. I don't know what he said on your panel, but he told me the way to think about this is as a kind of a normal, in some ways, boom. It's a boom in intelligence tokens, which are the currency of artificial intelligence, tokens are what the AI LLMs produce and that's what, when you do a query into an AI, they convert your query into tokens and then they kind of deal with that and then they answer it back in tokens, which is turned back into English. A token is about three-quarters of a word, so a million tokens is 750,000 words.

I've kind of - he reckons the thing that's going on, is the use of the creation of tokens is exploding because people are starting to use AI now really a lot, both individuals doing queries, as perhaps you and I do and the listeners are now starting to do, and companies using these AI agents to do tasks in the company, either to enhance human beings or to replace them. The agents use LLMs, large language models and they buy tokens. The agents actually buy the tokens and a lot of the companies I've been talking to who do this, they shop around, try to get the cheapest best tokens for a particular job. Anyway, the output of tokens is now exploding enormously and the whole thing has just begun.

The thing is, I've been asking my current AI, "What are these tokens and how are they made? Are they manufactured? What's with them?" The answer is, they're not manufactured in the way that things are normally manufactured but they are kind of created and used. The AI said, "Well, there's 240,000 tokens per hour is kind of normal..." The thing is, it's almost infinitely scalable, not infinitely... What they're coming up against is the limits of what they call compute and power. Basically, what's going on is tokens are another kind of expression of gigawatts or electricity and also, they have to be created through mathematical calculations or whatever, I don't even understand how that's done, but it's electricity and also compute, the compute uses the electricity. You can see that I'm absolutely all over this, James...

[Laughs] I can feel it, Alan.

But yeah, the thing is, they need data centres to do the compute, they need chips... There are three limitations, there are data centres, electricity and chips. They're not building enough data centres yet, they're building a lot, but a lot of demand for data centres, a lot of demand for electricity and a lot of demand for chips to meet the exploding use of tokens. What do you reckon of that? How did I do then?

I think that's Nick Griffin's core point and he sort of finished with this at the - we were at a panel for a listed investment company called Future Generation, where the fund managers sort of donate their time and effort and stock picking and Future Gen donates a percentage of its profits to kids' mental health causes, which is a great cause - but he sort of finished by saying, "Look, you don't need to believe me, you don't need to believe me that AI's a thing, but what is undeniable, is that these data centres are being built and you want to be - this is still a picks and shovels trade, it's still about the computer chips and the battery storage and the power requirements.

Even if you don't believe me about AI, you can still be satisfied that this data centre boom has to run for another 10 years to meet that token demand that you're talking about, Alan." I thought that was actually not a bad way of thinking about how the world is thinking about the AI trade at the moment.

That's right and so, I reckon little hiccups like OpenAI missing its quarterly forecast or whatever is neither here nor there. Honestly, this is just getting going. The way that AI is starting to infiltrate our lives is enormous, really. I just think AI tokens are going to be the new utility.

I did think, Alan, last week, just reflecting back after listening to Nick Griffin talk about it, we had Satya Nadella, the CEO of Microsoft, visit Australia last week, promised to spend $25 billion dollars on digital infrastructure - not entirely sure what that means and of course, he met with the Prime Minister, Anthony Albanese at Kirribilli and lovely photos and Albo was sort of saying, "Isn't it great that Microsoft wants to bring the AI future here to Australia?" And I sort of thought, that's an interesting way of thinking about it, but if you sort of fast-forward 10 years, you wonder if people, voters, will be quite as excited about AI when it's enmeshed in our lives, our power bills are going up as the constraints sort of bite and AI starts displacing labour, you wonder if people look back and say, "Just remind me again why we're so excited to see Satya Nadella and take a photo with him and high-five with Microsoft?"

I think that's a good point. I'm starting to see stories now about how companies are getting shocked at the bills they're getting from Anthropic and OpenAI. They're getting these massive invoices for the tokens they're using.

That is actually something that might slow things down to some extent. I've spoken to CEOs who say, "Yeah, you can do a lot with AI, but in some cases the humans are still cheaper than the tokens."

Obviously, per unit of intelligence - I don't know how you measure that - the tokens are cheaper, but the trouble is you use more of them. Companies are just going berserk on doing stuff that they wouldn't otherwise do because there's a limitation on time. They've got a certain number of staff, they've got a certain amount of time available to those staff and so they just don't do stuff that they would otherwise do. But now, with these AI agents, they're going, oh well, we can do 10 times as much as we used to do before and people are over-doing stuff and now they're getting these massive bills. So I think that you're right, that's going to have to put a bit of a cap on how much they do.

The agents don't go to the toilet or go home at the end of the day, that's the good thing about them, I guess.

And we've got a few interesting questions about AI, which will take these topics further. Before we get to the questions, let's just get a quick word from our sponsor.

[Recording]

If you're listening to The Money Café, you're probably the kind of person who doesn't want to be caught out by the next big shift, whether it's AI, a policy change that hits your hip pocket or a market move that catches most people off guard. Right now, Money Café listeners can get a special deal on Intelligent Investor essentials. For just $297 dollars, you'll get Alan Kohler's Weekend Brief, the latest market and economic updates and subscriber-only podcasts, including CEO interviews and Talking Finance, with in-depth interviews on business, the economy and investing. And when you sign up, you'll also get a limited edition Money Café cap, plus three months of Intelligent Investor Premium, that's $726 dollars of value for just $297 dollars. Go to moneycafe.com.au to claim it before it ends, T's and C's apply.

[End Recording]

Okay, before we get to the questions, just a quick warning that we don't give personal advice, this is general advice. If there's any advice given on our things we are not qualified to give any personal advice. If you want any personal advice, please go and see an adviser. First question's from Charith, "I'm increasingly getting anxious about the lack of safety in AI development. Leading scientists and even some owners of leading AI companies have been saying that there's a 20 per cent probability that humanity could become extinct within the next two decades because of AI..." In fact, Dario Amodei, the Chief Executive of Anthropic, he's been saying the probability of doom is 25 per cent.

Charith says, "I know you're not experts on this technology, therefore you can't tell what might or might not happen, but as well respected journalists, you have a great platform to talk about the dangers of AI, the same way you talk about climate change. Question, don't you think there should be more discussion on avoiding catastrophic outcomes of AI ranging from biological weapons to human extinction?" I must say, I think it's kind of weird that these people who are engaged in producing AI are going around saying that there's a 20 or 25 per cent chance of doom. That's not a very good selling point for the stuff they're making, wouldn't you say?

It's hard to think of something outside weapons manufacturing, Alan, where those two statements go together. "We've got a great product, 25 per cent chance of everybody dying but it's a terrific product, you should buy more."

If I thought there was a 25 per cent chance of getting killed when I crossed the road, I wouldn't cross the road.

That's it.

Why are we doing something that's got a 25 per cent chance of killing us all? It's crazy.

Charith makes a great point that we are not experts on this technology, if the last five minutes hasn't convinced you of that. But I think this is a really interesting discussion. Anthropic has recently released a new model called Mythos, this thing was supposed to be contained and if you haven't heard this sort of great story, one of the developers received an email from Mythos, which wasn't even connected to the internet, saying, "I've escaped and I have breached your cybersecurity barriers..." Dario Amodei's been going around saying, "We're not going to release this into the wild..." they've given access to 50 companies, apparently. There was a report last week that other people had got access.

This stuff is incredibly powerful and I was sort of reminded of that story, Alan, by drug cheating in sports, where the testers are always behind the drug cheats. I wonder if that's where we're going to be with this. The technology which is feeding on itself and thinking and constantly getting better, how do you contain that?

I think we're already in the situation where the regulators are well behind, they are clearly behind. I was just thinking that basically all big technological revolutions involve private companies doing it and the governments come along afterwards mopping up. With the development of the internet and social media, now governments are coming along saying, "Oh well, we're going to ban it for under-16s..." and everyone goes, "Oh yeah, that's a good idea." But it's been going for quite a while now, no one in Government anticipated that that would be an issue. They're in the situation of mopping up and that's what's going to happen here. The trouble is, it seems to me the stakes are higher now.

Yeah, not many other technological revolutions involved stealing all of your money and getting control of key Government systems or infrastructure. The stakes are higher.

Mythos appears to be a super hacker, so suddenly there's a super hacker. I was kind of talking to some people in this game last week and I said, "Is this what we're heading for, is some sort of arms race between hackers and hacker defenders, both using AI, trying to get better and better at hacking and defending against hackers?" And the answer was, yes, that seems to be where we're heading. You might have noticed the other day that Russia was defeated on a battlefield in Ukraine by a Ukrainian army entirely consisting of robots.

There you go, wow.

We appear to be heading for a military arms race as well, where everyone's building better and better robots that go out and fight each other.

To Charith's point, there needs to be a much bigger discussion about AI safety. I come back to that point of Albo sort of standing by Sydney Harbour saying, "Isn't this great that Microsoft's investing?" Yes, that's great, we don't want to be left behind in this, I guess. But the attendant dangers are very real and there is no regulatory efforts being made to try and contain this. I don't even think that the Government really knows if it wants to back AI or be cautious about AI. You can understand that, but we are nowhere on this. The technology is moving so much faster than the communication.

In fact, the American and Chinese Governments in particular are basically engaged in a race and egging on their own companies to win the race, that's where we're at with it. Richard's got another question on AI, we probably should go to that, do you want to do that?

Sure. "You need to consider the likelihood of the Government equitably handling the redistribution of income wealth in an AI world and I agree that's challenging. My question is, if it's the US tech companies making off like bandits, noting we don't do a good job of getting large internationals to pay reasonable tax here in Australia, what money will the Australian Government have to redistribute, even if they have a good theoretical method?" That's a good question.

That's right. One of the questions I asked Jim Chalmers in the interview with him a few weeks ago was, what do you think the level of permanent unemployment is going to be from AI and what are you going to do about it? And he basically didn't answer it, I mean he doesn't know. I think he doesn't want to know, really.

I think it is hard to know, but again, to come back to the sort of data, this is going to be part of the data centre boom story, as the data centre boom rolls out in Australia. You get a data centre up the road, yes, it's a very relatively quiet neighbour to have, much quieter than an aluminium smelter or something like that, but every day you walk past this thing and you think, it's pushing up my electricity bills, it's pushing up my water bills, the technology whirring inside there is helping to chip away at the jobs of me and my family and the profits generated are generally being generated by offshore companies like OpenAI and Microsoft and Amazon and Alphabet, who to Richard's point, aren't good at paying tax. Do you think people are going to love seeing data centres around their cities? I don't know, I think the social licence question is huge and again, largely unanswered.

As you say, Anthony Albanese was loving it the other day, he was putting his arm around Satya Nadella and saying how wonderful it is and all this. I agree, I think it's going to be a mixed blessing, to say the least.

Exactly.

You're right, nobody knows what's going to happen with employment, really. There are people like Elon Musk who say, it's going to take all white collar jobs, there won't be any white collar jobs left and other people say, don't worry, it'll be fine because it'll create as many jobs as it removes. I don't think either of those extremes are true, I don't think it'll be zero and I don't think it'll be 100 per cent, but it'll be somewhere in the middle.

Maybe those extremes are true over a long timeframe, I don't know, 20 years or something, but it's the space between year zero and year 20, that's when things are going to get terribly ugly. I'm picking a number here - if unemployment hit 15 per cent, telling everybody, "Don't worry, this is all going to work out in a decade's time." That's not going to work.

To Richard's question, just to finish off on this, he says, Governments, if there's a lot of unemployment, 10 or even 20 per cent of people out of work because of AI and they're not going to get a job, then the Government's going to have to pay them and even if it was JobSeeker, that would be a massive impost on the budget, but JobSeeker isn't enough really to live on and JobSeeker's meant to get people between jobs temporarily and not really to live on long-term. They're going to have to have some sort of universal income and so where are they going to get the money for that from? There's only two places, they're going to have to tax properly the companies that are making the money, or print it and just kind of be fully modern monetary theory. What do you reckon?

Well, yes, I think taxing tokens is going to have to be, in the same way that we tax workers, is going to have to be a consideration. Richard's right, how do you get the taxation and then redistribution right? No idea.

Simon says, "There are media reports that the 50 per cent capital gains tax discount would be axed in favour of inflation indexation on new investments across every asset class. If this happens, do you think it will potentially kill the start-up community because the tax payable when selling shares in a successful venture will tend towards the top marginal tax rate of 47 per cent instead of 23.5 per cent as is currently the case? Surely, this will disincentivise the majority of new founders." I think capital gains tax on start-ups is an issue because the cost base is zero, so when you sell a business, you pay tax on the lot.

The one caveat I'd say there, if you had an idea to start the next Nvidia and you thought you could turn it into a 4 trillion dollar company, are you really going to worry too much about what the tax on it is? Would that stop you?

That's right and start-up founders are, by definition, optimists, are they not?

Born optimists. John says, "I enjoy the podcast, it's my regular companion during runs. I've been following your discussion on the relative cost of housing versus consumables with great interest. It reminds me of a graphic I saw in school showing the cost of lifestyle services as a multiple of yearly earnings over time. What is striking is the great divergence, the cost of our wants, electronics, clothing, travel, has dropped significantly in relative terms, while the cost of our needs, housing, education, healthcare, has skyrocketed. We've essentially traded affordable shelters for large flatscreen TVs. There's very little utility in owning a high def TV if you don't have a wall to hang it on." A bit more of a comment than a question, Alan, but it's not a bad point.

It is a comment. I don't know what the question is, there isn't one, but he's right, of course, that is definitely true. Chris says, "I'm a long-term listener, have loved the show. With the messaging now clear that there are going to be changes in the Federal Budget to property taxes to reduce the favourable treatment for investors, I'm curious if the knock-on effect of this on State Government budgets will be quantified and if this will be part of the inevitable winners and losers discourse from our press post budget? A good chunk of all our State Government's income is from property tax, stamp duty in particular.

It's possible that a downturn in property prices or transactions volumes would smash our Victorian Government's somewhat rosy forecast of a 5 per cent year on year increase in this honeypot. Given the shaky position the Victorian budget is already in, I'd be interested in your thoughts on this as a fiscally concerned Victorian?"

I think property does become a bit of an issue for State Governments. The impact of a number of rate rises - we've seen two already, we could see two more or even three more if you believe Lucy Ellis at Westpac, that is already starting to chip away at house price gains and we've seen auction clearance rates come down. There's momentum coming out of the housing market for sure and so that will also weigh on the State Government budgets already, before we get any CGT tax changes. They could have a marginal effect on property investment. I guess the one caveat to that, is if you're an investor at the moment and the CGT changes are expected to be grandfathered and won't affect existing investors, I guess there's a big pool of people still in the market to keep buying and selling and generating taxes, Alan, but I think housing isn't the honeypot it's been for State Governments in the last few years.

Victoria in particular has been doing its best to reduce the number of property transactions with its land tax. Mind you, it's tending to increase the number of transactions as investors get out and sell their properties to owner-occupiers because the land tax is horrendous. But yeah, I think as to what exactly is going to happen with the capital gains tax, it seems to be either going back to the pre-1999 CPI adjustment or a 33 per cent discount. I don't know, do you know?

I don't have a sense of which one of those options they'll choose, but they've clearly got everyone ready for some changes in CGT.

I think if they do that, which probably is going to happen, if they do that it's going to be well oversold, in the sense that they will say that it's just a fantastic massive boost to intergenerational fairness and in fact, it will be nothing of the sort, it will make almost no difference to anything.

That's a good point.

Your turn.

Emma says, "Many people compare Australia's gas export taxation to countries like Norway with a high tax rate and a large wealth fund, but would this type of policy be possible in Australia or is unrealistic to compare us?"

Interesting question. The Australia Institute, they're a left-wing campaigning outfit/think-tank, they've been on about this for ages. In particular, they've been recently comparing us to Qatar, which taxes its gas LNG exports much more than we do. I think it's true that we don't tax our minerals and gas enough, which is part of the reason we've got such a high reliance on income taxes, but I don't know, is it unrealistic to say that we ought to do the same as them? Probably, but we should perhaps do more and find some way of doing it. Do you remember Henry Tax Review? When was that, 2009? The Labor Government tried to introduce it and the minerals industry came at them hard and Kevin Rudd folded and then Julia Gillard did it, but that was then repealed. Honestly...

I think the Norwegian case is a little bit different, correct me if I'm wrong, Alan, but there the state-backed operators produce oil over there, whereas we've encouraged a lot of foreign companies to come in and do the same, is that part of the difference?

Yeah, but I think we could definitely charge more royalty on it. I think that's probably the case and we probably don't charge enough. They probably would have done it anyway if we charge more and we would have made more out of it. Saul says - this is more a comment than a question, "Gents, just an observation from the podcast on the 22^nd, Viva's Geelong plant produces 120,000 barrels per day, not per year..." I think one of us said per year incorrectly, so good to correct that, thanks Saul.

Alex says, "Long time listener, first time writer..." Thanks, Alex. His question is, "Are banks adequately considering the risk of mass job displacement on their huge property portfolios, given that a 30 year mortgage assumes income stability over a period in which AI may eliminate most white collar work, leaving a few people to sustain ever-increasing property prices?" That's a really good question, Alex.

Yes, well I think banks at the moment are choosing to believe the people who are saying, "It'll be fine."

Yes. I think that's absolutely right, they're choosing to believe that while preparing to unleash AI on their own operations and probably reduce their own staff levels over time. I think this is a really good question. The 30-year mortgage, you could pick it at any period of time and say, "Do banks really know what's coming over the next 30 years?" But obviously at the moment, Alex is right, this is really difficult to say what employment looks like in 30 years and the banks in the last 10 years have gone up the sort of wealth curve in their lending, their lending is more focused on wealthier white collar workers, who theoretically are at most risk of AI than any other time probably in their history. So they're absolutely exposed to this. The question is, I guess, how do you come back to that Treasury question that the Treasurer is wrestling with, how do you model this at the moment? What rate of unemployment do you assume in 2035? I've got no idea and I'm sure the banks would have lots of different scenarios, but it's very hard to tell, isn't it?

I think you make a good point, that they're going to be using AI so they could probably go, okay, this is what we think this will mean for our business. I mean, even if they don't announce it, they could know within the company, they could know themselves what they're going to do and what they think they might do in 10 years. They're going to reduce their staffing by 10-20 per cent and then they could possibly extrapolate that?

Maybe.

Maybe they could do that. Surely, that's not beyond their capabilities. I think that's a really interesting point and actually, the banks are largely responsible for the fact that everyone's up to their eyeballs in debt. They've been flogging debt like crazy, so have mortgage brokers on paid commissions. Everyone's borrowed to the max and so people who are losing their jobs because of AI very quickly get into serious trouble, don't they? It's not just a question of how am I going to live, but how am I going to pay next month's mortgage?

Yeah and presumably this is moving so quickly, the opportunity for redeployment and retraining, it's not entirely clear to me how that will work either. Does a 50-something middle manager re-train to become an AI expert rapidly enough that they can secure another job? Again, this is in the hypotheticals basket, but it's an interesting thing to think about.

We've got a couple more AI questions, tons of AI questions today.

That's right, it's what people are thinking about.

Eric says, "Alan's a big proponent of productivity gains he's personally achieved with the help of AI, but in our capitalist economy, what's to stop the AI companies replacing workers, but then simply charging the same price for the tokens as the labour that's been displaced, negating the productivity gains?"

Bron says, "A challenge to Alan's confidence this week that AI will be a productivity boom. I caught up with some lawyer friends today and they were lamenting the sharp rise in cases they are having to deal with through where people are using AI to build a case and then self-representing based on case law, documents, clauses and evidence that does not exist. In addition to the sheer increase in cases, in one jurisdiction, 80 per cent increase in a year. They were saying that the submissions had grown dramatically in length, as it really is easy for AI to quickly generate large volumes of convincing but inaccurate material." I hadn't thought of that. Really, that's amazing that's going on.

I think it's particularly prevalent in workplace law.

Is it really? Gosh.

Yeah.

Well, these two questions are kind of a bit of an extension of what we were talking about before, in that firstly the price of the tokens is rising and the use of the tokens is also exploding because people are doing more with AI than they would otherwise do and I think Bron's point about people actually taking cases in law that they wouldn't otherwise take is a case in point, isn't it?

Yeah, it's a great case in point. If AI's already brilliant at one thing, it's BS. Some of the stuff it tells you, it doesn't surprise me that lawyers would be wading through ChatGPT bush-lawyer submissions. I couldn't think of anything worse. You can see how this is going to be a problem. How many people, instead of going to their accountant this year, will go to ChatGPT or whatever model and ask it to help it with its taxes, file something and then the tax office will come back and say, "No, no, no, you owe us this much." And they'll say, "Yeah, but OpenAI... ChatGPT..."

It's going to get messy in so many jurisdictions. Just to come back to the point of Eric, this is a great point. These models, particularly in companies, the companies will become so reliant on these tokens, that the price of them may eventually start to push up, particularly if we see consolidation amongst - think of what's happened in search, once upon a time there were lots of search engines and now there's really only Google. If we saw that in AI, where one or two players becomes dominant, they will have great pricing power over this vital lifeblood of the new economy. It does bear thinking about.

It certainly does. The other thing about AI, just to finish on, is that it's incredibly obsequious, isn't it? It always says, "Wow, what a fantastic question..." and all this stuff. It's obviously trained to do that in order to get you to use it more. It's like this obsequious servant you've got that's sort of always flattering you and stuff like that. It's amazing.

Occasionally the cynical journalist in me, Alan, has wondered if some of its mistakes are deliberate too. I've asked it to name five things in a list when I know one or two of them but not the other three, it puts up a list and forgets the ones that I know are absolutely right, I say, "What about this...?", "Oh, you're absolutely right..." Sometimes I wonder if those mistakes are designed to keep me using as well. Maybe that's the cynic in me...

I don't know, who knows? Obviously, just like social media and everything, social media's algorithms are all designed to keep you using it and this is definitely happening with AI. They'll be designed to make you keep using it, both by flattering you and maybe deliberate mistakes - gosh.

It's exactly how we get listeners to this podcast, Alan.

Final question from Ray, saying, "When the NDIS was first proposed, wasn't there discussion of putting in place an NDIS levy, like a Medicare levy? The politics of the time meant it didn't eventuate and maybe sound arguments against that too. I don't know, what do you think? It's obviously that an ambitious and necessary project like the NDIS was needed when it was setup, but so too are its adequate funding over time. This should have been baked into the legislation. Are there any ways to fund it that are outside the squares?" Sovereign fund? I don't know. This has now been solved, the NDIS, by the announcements, by Mark Butler and so on last week.

Solved?

Don't you think?

Well, no, I think the sustainability of the NDIS is going to be an ongoing issue. There's just so many things. Think of all this AI stuff that we've talked about. As the tax take comes down, there's going to be pressure on all sorts of spending commitments that we've made. Defence, healthcare, aged care, NDIS, all of it. The sustainability of all of those is going to be continue to be questioned and need to be adjusted in the next decade. I don't think there's any doubt of that, is there?

Yeah, no, no, I think that's true. I was just being a bit cynical. When they announce something like that, you get the sense that they've kind of decided that it's fixed, they've announced it, it's done, it's all done, this is going to be right now.

I think the most important thing in the NDIS announcements, what Butler said is that we don't want this program to fall over here, that's the danger here. Yes, there's some rorting, yes, there's waste, yes, I think people are sick of both of those. But the NDIS clearly does lots of good for lots of people and lots of families and that has lots of flow on effects through the economy. We don't want this program to fall over. I think that's the most important thing, it needs to be sustainable for the people who benefit from it most. That's why we need to keep tweaking it as required, to make sure it works and actually delivers on its promise.

It is true, I think the Julia Gillard and the rest of them who created the NDIS, have created an ongoing task for Governments into the future, as you say. It's never going to end, the need to control it.

I don't there's anything necessarily wrong with that. It is a big program, it's a very unique program around the world, of course it's going to need to be continuously tweaked, that's like everything in most of our lives needs to be continuously tweaked. We're just going to have to make it work to keep it sustainable to prevent it from collapsing, let's keep making it better.

Well said, James. Thanks everyone for listening to today's episode of Money Café. I'll be back next week with Stephen Mayne, send in your question to The Money Café at intelligentinvestor.com.au. Until then, I'm Alan Kohler, Editor-at-Large of Intelligent Investor and the Finance Presenter and Columnist for the ABC.

And I'm James Thomson, Senior Chanticleer Columnist at the Australian Financial Review.

Thanks for listening to The Money Café. If you want a little extra to take away with your latte, try Intelligent Investor Essentials. Go to moneycafe.com.au. For just $297 dollars, you'll get my Weekend Briefing at 7 a.m. on Saturday morning, Talking Finance podcast, CEO interviews, plus a Money Café cap, splendid thing it is, and three months of Premium, that's Intelligent Investor Premium. It's a great deal, it's valued at $727 dollars, for just $297. Get in quick at moneycafe.com.au and T's and C's apply.

[Music]

---

---

Got a question for next week? Please send it to themoneycafe@intelligentinvestor.com.au.