Latest / Uncomfortable Conversations with Josh Szeps / Josh vs the News: Why Everyone is Wrong About A.I.

Transcript

Josh vs the News: Why Everyone is Wrong About A.I. · Uncomfortable Conversations with Josh Szeps

• 0:03
  G'day humans. Welcome to the safe space for dangerous ideas. There's a lot going on this week. If you're watching in Australia or if you're listening in Australia, then the big news is the trial of one of Australia's most, I guess, polarising and also influential media personalities, Alan Jones. Many people have been asking me for my opinion on this. Many people think that I already wrote a column about this.
• 0:29
  Because back in November of 2024, when this gentleman was arrested, I did a very well thought out, I think, supremely carefully manicured editorial about my thoughts about it.
• 0:46
  which I encourage anyone who's interested in my opinion to go back and listen to. November 2024, it was called Me and Mr. Jones. I used to work for Alan Jones. I tried to thread the needle in that editorial. I'm not going to republish it now because this is the first week of the trial, and I've already said it. However, what happened at the end of the publishing of that is, of course...
• 1:09
  Some journalists listened to it and then wrote it up in the newspaper with the, you know, out of how many tens of thousands of words do you say in 40 minutes? I mean, a lot, maybe hundreds of thousands of words. And they take it and then they cut out the most salacious bits and then try to represent it as being me having an opinion that I, you know, did not have unless you understand it. It's so wrapped in so many qualifiers and conditions that it renders their interpretation basically a misrepresentation.
• 1:39
  And now when I talk to people about it, they go, oh, I remember reading a column at the time. And I say, I didn't write a column at the time. And they say, yeah, yeah, yeah, I remember. I'm a no, no, no. Someone wrote kind of a column for me and sort of made it look like it was my opinion boiled down to 400 words instead of 40,000 words.
• 1:58
  Such is the liability of working in the media. If you want the entire thing in its glorious, unfiltered, you know, full of all of the necessary qualifications and equivocations, glory, go back. It's in the feed, November of, I think it was November. Yeah, November of 2024.
• 2:15
  um throw in some chats into the chat on sub stack i can see that people are already asking me questions which i'll get to in just a bit uh also we have a survey have you seen this we're doing i really want to know we all on the show really want to know what you like what you don't like what you want more of what you want less of the show is expanding we're doing well we're growing a huge thank you to everyone who bothers to support the show financially
• 2:38
  Um, uh, and so I want to hear especially from you about your opinion. So there's a survey, it takes two minutes to fill out, uh, maybe three if you're slow, maybe, you know, four if you're, I don't know, otherwise handicapped. Uh, and you can go to uncomfortableconversations.substack.com. There's a tab at the top of the screen called survey. It's not hard to find. Just go to uncomfortableconversations.substack.com. Look for survey at the top.
• 3:01
  Click on it, and Bob's your uncle. Or your aunt in these times. Bob could be a non-gender, gender non-binary, fluid they. So either way. Bobina or Bob, we'd love to hear from you. We've got 450 and counting so far. It's only been up about a day. I want to get to 1,000. We can do 1,000, right? If you haven't filled it out, just fill it out now. Well, not right now. Keep watching me or listening to me. Then when I'm done, fill it out. That would be greatly appreciated.
• 3:31
  I've been thinking about AI. Don't know if you've heard about it. It's big. It's a big new thing. The kids are crazy for it these days. Artificial intelligence. And so much of what I read about AI at the moment is based around either the minutiae of what model is doing what, or a scary story about some model leaping out of its box and going and killing a whole bunch of people or breaking into something or other, or
• 3:58
  or it's some tech bro who's talking about how no one's going to need to work in the future, and we're all going to be millionaires, and we're going to live lives of pleasure and plenty, and it's going to cure cancer, and it's going to solve climate change. And a lot of people don't know what to think about this, including myself. Like, should we be terrified? Should we be incredibly hopeful? Is it going to kill us all? Is it going to liberate us all? And to some extent, I've been starting to think that these conversations are the wrong conversations.
• 4:28
  because by definition they just have no answer at the moment. I mean, it is completely uncertain. I mean, I'm pretty certain that if someone expresses to you certainty about whether AI will be on net good or bad...
• 4:47
  What you can be certain of is that that person should not be listened to, because they can't be certain. It's a bit like, reminds me a little bit of like religious people telling you that they know what happens after you die. I mean, you know that they don't know that. And you know that on some level they know that they don't know that. We all know that that's one of the great mysteries of...
• 5:10
  human existence, of existence full stop, of consciousness, of the cosmos, stop telling me you know, you can't know. We know you can't know, you know you can't know. So instead of thinking about all that stuff, I've been very persuaded by a previous guest on the show who's a multi-time guest, very smart gentleman, good friend of mine.
• 5:30
  Well, he's not a good friend of mine. I mean, you know, we're friendly. It's not like I call him up and we go to watch The Odyssey together and share a popcorn. I mean, he's a good person who happens to be a friend of mine. Joel Pearson, he's an academic and he is most interested. I mean, he's essentially in the neuroscience field, in the psychology field. And he is most interested in the way that humans adapt to stressful or tumultuous situations and how...
• 5:59
  all the kind of psychologically fallible ways in which we make predictable errors, the ways in which we can predict that we're going to go awry, how we can preempt them and how we can overcome them. And now he's really interested in the sociological side of artificial intelligence.
• 6:15
  And to me, this gives me a grounding in which I can think about the revolution that we're on the cusp of, not in a way that is full of airy-fairy what-ifs about how the technology might evolve, but is grounded in a certain certainty about the way that cultures and societies and individuals react to widespread change.
• 6:41
  I wrote about this after I had Joel on the show. It's an episode well worth listening to. Also, you should go back and listen to the Joel Pearson episode, which was in, when was that, Jono? That was probably March? March, yes. The article was March. So it must have been March. Terrific. Very fascinating. And certainly if you're a policymaker or if you're someone at the top of a big organization,
• 7:02
  This is, I think, where we need to be thinking about rather than, oh, can I play around with the latest model of fable and what can fable do that Sonnet can't and how is, you know, Gemini compared? Like that all sort of ultimately in three years time, no one's going to give a shit about any of that, about what was going on with a particular model in a particular month of 2026. What they're going to care about is how we as human beings wrestled with the unleashing of a completely new way of living and working.
• 7:31
  So I'm going to read this little piece. Firstly, I'm just noticing before the question disappears in a thicket of other questions from Substackers, Ryan says, off topic, I'm really struggling to form an actual stance on the Israel-Palestine conflict. Join the club, Ryan.
• 8:01
  says Ryan I agree this is also one of those things like artificial intelligence and religion where I mean basically anyone who says anything about the Israel Palestine conflict is a bit right you know they're like probably a bit right you know Israel mistreating Palestinians yeah Palestinians being run roughshod over in the Nakba
• 8:31
  Yeah. Palestinians having a legitimate claim to the entire territory? Yeah. Jews having a biblical claim to the entire territory? Yeah.
• 8:41
  pro-Palestinian protesters dabbling in fairly anti-Semitic, anti-Zionist tropes. Yeah. Right-wing Israeli settlers and Netanyahu's ilk baiting the world and being overly obstinate in refusing Palestinian overtures. Yeah. Like, you know, it's like, where do you end up?
• 9:01
  So I find that I am most sanified by a few voices who try to be maximally generous towards the anti-Israel position whilst remaining tentatively Zionist. I think that's probably the center of gravity that is the least...
• 9:21
  Um, partisan or the least crazy making. So have you read it? Go, who I had on the show last month. That's a great chat. I mean, his, his asked me, any things are fabulous. He has a podcast as well. Uh, Russ Roberts of econ talk is good. Um, Dan Senor is good. Uh, he has, uh, I don't remember the name of Dan's podcast, but
• 9:48
  Basically, kind of like... I mean, maybe this sounds biased if you're super pro-Palestinian, but I do think that the pro-peace Israeli left, especially the people on the Israeli left who have a deep understanding of Arab nationalism, Islamism, jihadism, Jewish extremism, far-right settlers and so on, are the best people who I have heard who can straddle both perspectives and put it in context. Yeah.
• 10:16
  I'm sure I'm missing some very learned pro-Palestinian voices. I mean, I've spoken to some of them. I don't mean to sound kind of inside my bubble. Maybe I am. But in general, my skepticism towards the pro-Palestinian voices who sound super conciliatory and too statey and sort of pacifistic
• 10:43
  is that they have no support in Palestinian society or culture or leadership, with the possible exception of whatever fatuous statements come out of the moribund and corrupt Palestinian authority that nobody pays attention to, likes, or listens to.
• 10:58
  So I'm somewhat more skeptical, whereas I know when I speak to Haviv Retigur or Dan Sinor that they represent a significant portion of public opinion in Israel, that if they could have the peaceful resolution that they want, that there would be a quorum for it.
• 11:18
  And so I have somewhat more faith that they're not just like spinning yarns in the sky and building castles of their own imagination in the some way that the most westernized Palestinians are. I think the truer case of Palestinian liberationism is...
• 11:35
  is the articulate case that gets made by the more, I don't want to say more reasonable, but certainly less rabidly Jew-hating, Israel-hating, Palestinian maximalists, the people who will say there shouldn't be an Israel, at least there shouldn't be a Jewish state, the whole thing should be an Arab state, and the Jews can live alongside Arabs in peace.
• 11:57
  I think that is the basic point of view of the vast majority of Palestinian people. It's becoming the point of view of the vast, well, I wouldn't say vast majority, but it's becoming the point of view of a significant minority of left-wing young progressives in the West.
• 12:27
  voices who could ally with Israeli lefties and have a two-state solution is somewhat more, yeah, I just, an alarm bell goes off in my head because I think like there is no universe in which that person is actually able to gain traction on the ground, and increasingly, sadly, even the Israeli moderates are increasingly losing traction on the ground in Israel as well,
• 12:53
  But at least you can point to long stretches of Israeli history where a significant minority, if not a majority of people, were fully on board with the idea of a two-state solution, and the same can't be said of the other side.
• 13:23
  the situation and the nature of the foe that they are up against in terms of understanding the origins of... I mean, what's so fascinating about Javier Redigur and why you should go back and listen to that episode is he really understands and empathizes with the conditions that gave rise in the 1960s and 70s to a maximalist, Arab nationalist, almost Maoist, so kind of completist, revolutionary anti-colonialism as the dominant strain of Palestinian resistance. Right.
• 13:53
  And for modern-day young lefties to sort of just look at the current snapshot of Israel brutalizing Palestinians and saying that's terrible and it's obviously terrible without understanding the pickle, the way that the Israeli left, not Netanyahu, but the Israeli left, the pacifists, have been snookered.
• 14:16
  by an adversary that essentially makes it impossible for them to pursue any peace overtures, I'm not talking about now, I'm talking about over the past 25 years, 30 years since Oslo, that if you ignore that, then you're actually, you're not being sympathetic to Palestinians because you're not giving them any agency. You're basically just regarding them as props in your morality play because you come from a country where the...
• 14:43
  the fashionable norm at the moment, for good reasons, is anti-colonialism and, you know, uplifting traditionally marginalized communities and instinctively having empathy for the poor brown-skinned person who is having bombs dropped on them over the smart-looking white guy in a suit who's the one in control of the bombs. Like, these are all laudable instincts amongst young pacifistic lefties in the West. But
• 15:09
  They also do deprive Palestinians themselves and Palestinian leadership of the respect of believing that they actually had any role in the long, long, long, long lead up to all of this. It kind of just makes them almost infants or like the pets of Israel somehow.
• 15:29
  When in actual fact, you've got half a billion Arab Muslims almost who have not been uninvolved in the Palestinian issue. You've also got Iran, non-Arab, but also, you know, a meddling Islamist regime. Yeah.
• 15:44
  Funding Hezbollah and Hamas and getting pretty good at manipulating social media and manipulating the global narrative. And so there's just a, there are a lot of threads going through the entire conflict that I find it really eye opening to.
• 16:00
  understand, and that I think are actually more respectful to Arabs and to Muslims and to Palestinians, because it acknowledges that they're actors in this, that they have been participants in this, that they have chosen particular strategies, that they have learned from anti-colonial movements in North Africa against European colonial powers. They have learned from revolutionary movements in places like China that they're not idiots. There have
• 16:27
  And they've chosen, they chose a strategy and Israel sometimes responded the way that they wanted to, which is violently and sometimes responded the way they didn't want to. And to understand that sometimes what they actually wanted to do was for Israel to kill lots of Palestinians.
• 16:43
  just as Hamas obviously did on October 7th, is also part of understanding the whole strategy. And so that's all a long-winded way of saying that I think you need someone who has some understanding both of the Zionist cause, but also of the historical Palestinian cause and Arab nationalism as it fed into Palestinianism in order to get a grasp on it. And I think the best people for that are people like Haviv and Dan Sinor.
• 17:10
  Chickenfog on Substack says, I read today that AI models are hoovering up rare books, scanning and then destroying them. Thoughts? Well, excuse me, a controversial opinion here, but I think that's bad. I think destroying books is bad.
• 17:29
  But also, where did you get that information? I would want to see the source. It sounds like the kind of thing that a grand aunt sends on Facebook alongside a Pope endorses Donald Trump meme. And did I get on the wrong bus? Isn't this about AI? Yes. Sorry. It was an aside. I'm an easily distractible person. Also, I haven't had all my coffee yet. Hang on one second. Look at this. We got my extra strong plant white.
• 18:00
  delicious. And without further ado, the show proper begins.
• 18:07
  with everyone's favorite Josh versus the News trope, the reading of a column, the reading of one of my columns. By the way, I hope that this weekend I'll be able to write about the trial that I alluded to earlier. It's going to have to go through a lot of legal. And so if you have ever thought about subscribing to the Australian Financial Review, now would be your time.
• 18:34
  If you don't happen to subscribe to the Australian Financial Review, good for you. You're watching this or listening because I'm about to read you this piece from March 20th after my interview with Joel Pearson. It's entitled, A Four-Point National Plan to Save Us from the AI Apocalypse.
• 18:54
  and it's subtitled, Rome didn't fall because its engineers miscalculated the tensile strength of aqueducts. Countries fail when they become incapable of managing social, political, or economic change. By Josh Zeps, columnist. And then there's a little photo. I'm not going to read anything else apart from the actual text, okay? It begins, if you've ever had the misfortune of attending an AI conference,
• 19:23
  and I use misfortune loosely since the canapes are usually excellent, the conversation runs on a predictable track. Experts discuss all the things we don't know about AI. From the humdrum, will it put radiologists out of a job, to the fantastical, will AI become sentient and eat us with a side of fava beans and a nice Chianti. What you don't hear discussed is the one thing we actually do know for certain.
• 19:53
  We know, as surely as we know that Barnaby Joyce wears a hat, that when highly evolved primates like Homo sapiens experience prolonged periods of uncertainty, they become anxious, manipulable, and foolish. We need to change the way we're talking about AI. We're on the precipice of a revolution in which the only certainty is the certainty of uncertainty.
• 20:21
  We can't predict the details, but we know there will be sustained, broad, deep, superfast disruption.
• 20:31
  And psychologists, historians, and sociologists know a lot about the human consequences of such disruption. Details be damned. It's the difference between asking which street fighter will win the particular knife fight and asking what would happen if your city was full of knife fights all the time. You don't need to predict each fight to predict the consequences.
• 20:55
  Countries fail when they become incapable of managing social, political, or economic change. They rarely fail for technocratic reasons. Rome didn't fall because its engineers miscalculated the tensile strength of aqueducts. It's time to expand our conversations about AI, to include fortifying ourselves for the greatest social and psychological upheaval since World War II.
• 21:22
  If that sounds like hyperbole, if your picture of AI disruption is some corporate layoffs, deepfakes, and dodgy unfair dismissal claims, you're looking at the first few raindrops and missing the coming cyclone. Already, some of the world's best software engineers have handed the bulk of their work to AI agents.
• 21:47
  These systems are not large language models like ChatGPT. They are virtual creatures, for want of a better word, that work autonomously, unsupervised, for days at a time, testing their output and fixing their errors before they get back to you with a completed, multi-step, real-world project.
• 22:10
  Software is the canary in the coal mine because it's where AI companies invested most of their early efforts. They did this in order to train AI models to produce yet better versions of themselves.
• 22:24
  Now the wave is about to breach the floodgates of Silicon Valley and gush into every nook of our lives. The computer nerds have replaced themselves. The canary is kaput. It has met its maker. It has expired. It is bereft of life. It is a late canary.
• 22:43
  Bah, the skeptics' tut. We've always had economic and technological change. We didn't complain about the whale blubber industry going bye-bye, did we? Chicken littles? Excellent point. Please enjoy your brandy and your pipe. But there are three variables that make any disruption harder to manage. It's speed, breadth, and intensity.
• 23:12
  The Industrial Revolution was broad and intense. That is, it affected nearly everybody, and those affected felt it in a big way. But it unfolded over a century. There are faster examples of economic change, like the industrial collapse of Appalachia or Northern England, which were also intense, but they were geographically contained.
• 23:38
  And as for changes that are ubiquitous and fast, but not traumatic, we've lived through those, like the internet and smartphones. The only recent historical example of a change that satisfied all three conditions, that is sudden, across the board, and intense, is COVID.
• 24:02
  But it didn't last long. And importantly, we all knew that it wouldn't last long. You can endure anything for a while. We now are at the earliest stages of an AI revolution that will be rapid, widespread, disorienting, and open-ended. It comes with no promise of a vaccine, no date on which our leaders will reinstate our inalienable right to picnic with more than five friends.
• 24:29
  It'll hit nearly every profession in every postcode in Fast Forward. It's the everything, everywhere, all at once disruptor. This is not an argument about how AI could go wrong, by the way. This is what happens if AI goes right.
• 24:47
  If we emerge in 2050 onto sunlit plains of pleasure and plenty where AI has cured cancer, solved climate change, invented cold fusion, and explained an IKEA instruction manual, the path from here to there will still be staggeringly disruptive. Luckily, we're not bad at managing change when we take it seriously.
• 25:12
  Professor Joel Pearson is a neuroscientist who works on AI human readiness at the UNSWAI Institute. As he recently explained on my podcast, change management is not a mystery. Lottery companies hire psychologists, financial planners, and life coaches to help winners navigate rapid change. When the NBA hands college kids $100 million contracts, it gives them the same support.
• 25:42
  Australia has had success with public health campaigns like the AIDS ads, Slip Slop Slap and Life Be In It.
• 25:50
  If we're going to get to our AI utopia, we need a plan. I'll take an aside here to explain the slip-slop-slap reference to non-Australians. Now that I look at it through the eyes of a foreigner, or perhaps a non-Australian, I bring a fresh pair of eyes to it, it is a very Australian thing. The slip-slop-slap campaign was a public health campaign to get the rate of skin cancer down. Australia, as you may or may not be aware, has by far the highest rates of skin cancer and melanoma.
• 26:19
  in the world, like, you know, light years ahead of the next neighboring country. And so there was a big policy of like, you know, let's make sure that people slip on a hat, slop on some sunscreen,
• 26:36
  and slap, what was it, Jono, a shirt? You're supposed to slap on a shirt, a t-shirt, slop on sunscreen, and slip on a hat. So slip, slop, slap. Everyone in Australia knows what slip, slop, slap.
• 26:52
  means uh and now i say it the more i say it the more it sounds like some a horrible uh you know uh 1970s german porn movie uh but you know enough of that so australia's had success with public health campaigns like the aids ads we had very graphic very terrifying aids ads and we had one of the best public health responses to aids slip slap and life be in it was a an ad campaign that
• 27:16
  Uh, incidentally, I think that was coined by Philip Adams, if people are familiar with Philip, a long-time ABC radio broadcaster and columnist. Uh, life being it was about getting moving again and not sitting in front of the TV turning into a couch potato. So, my article continues and concludes, the National AI Management Plan might have four components. First, an ad campaign explaining what AI is, what it isn't, what's coming, and what you can do.
• 27:47
  Second, rules around how companies undertake AI-driven redundancies. Workers need to be fired in a structured way with plenty of time and support, not in a Friday afternoon email. Third, tax reform to redistribute the productivity gains from corporate balance sheets and, if possible, from offshore tech companies to the communities bearing the cost of the transition.
• 28:13
  Fourth, a public health campaign about psychological wellness, resilience, connectedness, and mental fitness that gives people tools to manage anxiety. It wouldn't be perfect, but it would be a start. Pulling off a wholesale national strategy like this would be hard to do. The alternative, unfortunately, will be harder.
• 28:37
  So that was my little thought, beginning thought about artificial intelligence, which is, as I said at the beginning, let's stop worrying so much about the, I mean, someone needs to worry about the alignment problem and about whether or not these models are being kept on track, but it strikes me that so much of the lay person's energy at the moment
• 28:57
  is either in fretting about things that they can't control, like the trajectory of the technology that's in the hands of Silicon Valley multi-billionaires, or they're putting the efforts that they can control into what strike me as pointless nonsense, populist nonsense, like opposing data centers. You know, you listen to people talk about, I don't get the whole data center thing. Well, they use a lot of water.
• 29:23
  Yeah, you know, almond farms use a lot of water. I mean, if it were the case that AI was going to completely revolutionize life for human beings on this planet and liberate us from the drudgery of work and usher in a revolution of prosperity like the Industrial Revolution did...
• 29:43
  Do you really want to be the person before the Industrial Revolution going, well, factories, I mean, they really do take a lot of water to, you'd be like, well, yeah, so what? I mean, we're going to take everyone from working 16 hours a day of backbreaking labor to sitting around doing podcasts. I'll take the, just use the water. It's fine.
• 30:00
  Um, so the things like data centers, I think these are just proxies for people's anxiety. And as I said, at the beginning of the article, when people are anxious and they feel destabilized and they don't know what's going on, they are reliably anxious and manipulable and foolish. And so they lash out in certain ways. It don't want the data center here because it's making a strange hum and it's, it's got lights on at night, uh, and it uses up water and it's on agricultural land.
• 30:28
  I don't know. All right, put it somewhere else. I mean, I don't get this. Focus on things that we can control. And what we can control are the things that we know. The things that we know are the way that human beings individually and collectively respond to times of total change. I mean, just think about what,
• 30:47
  has happened to liberal democracies as a result of, I suppose, the twin stressors of the global financial crisis in 2008 and COVID in 2020.
• 31:03
  you could argue that basically the loss of, and perhaps China's entry into the World Trade Organization and globalization and so on, but you could argue that the loss of manufacturing in Appalachia in America and in Northern England in the UK, coupled with the global financial crisis and the way that it put young people back a few decades...
• 31:27
  Those two things essentially gave us Donald Trump and Nigel Farage and Brexit and Make America Great Again and populism and the far right. And of course, you can throw into that however much you want to attribute immigration to being part of that. But the failure to address...
• 31:44
  Illegal immigration and the global financial crisis and then COVID and, yes, perhaps the decline of manufacturing. Those things alone were able to nudge liberal democracy into a precarious situation, which we're having to now proactively try to scramble our way back from.
• 32:04
  If you multiply that by 10 or 100x and you have 40% of the white-collar workforce losing their jobs, you have long-haul truckers losing their jobs, you have no more accountants, you have no more conveyances, you have...
• 32:22
  It's going to be significant. I'm not saying it's necessarily bad. I mean, I think I'm just... My general setting on life is mildly optimistic, so I am excited by it. I mean, I think it's... I can't wait to not have to wait on hold to talk to somebody because the person I can talk to is an ultra-knowledgeable, perfect AI. I can't wait to not have to...
• 32:41
  go to accountants and all of these sort of gatekeepers of society and bureaucracies. And I can't wait to not have to deal with the IRS the way that we currently deal with the IRS, but just to have a personalized tax advisor who can liaise directly with the government via AI. I think there'll be lots of great things. I can't wait to be just running a whole bunch of different projects and leveraging my capabilities by deploying agents to go and do them for me.
• 33:06
  That'd be fabulous. But even in the best case scenario, it's going to be really, really disruptive. And people get stupid when they're anxious and when they're afraid of disruption. So that kind of AI management plan, which I sort of just pulled out of my ass after talking to Joel Pearson, who has thought it through much better than I,
• 33:26
  Um, makes a lot of sense to me. An ad campaign, a government ad campaign, I know Americans will hate that, shouldn't have the government talking to you, but it's a pretty normal thing in other democracies for governments to try to deal with impending crises by explaining to people what they should do, and you don't have the government do it in a partisan way, of course that's inconceivable with Donald Trump in the White House, but ideally you would have...
• 33:48
  public health officials and experts get together and say, all right, here are the things that people need to understand. What is AI? What is not AI? What is likely to come and level with people? I mean, level with people about what's coming. And here's how we're going to have to manage it together.
• 34:05
  And secondly, I do think that industrial relations has to be updated, meaning the relationship between employers and workers. In Australia, we have currently a fairly sort of retrograde, old-fashioned, we thought it was a centre-left government, but it's proving its colours to be considerably more left-wing than a lot of people who voted for it thought it was going to be.
• 34:24
  which has a very pro-union approach to industrial relations, which is fine, that's great, but it doesn't necessarily help the workforce to be as nimble and agile as it might be if you've got gigantic unions that are sort of vaguely and sometimes not so vaguely affiliated with organized crime and coercion running the workforce in large sectors like construction.
• 34:54
  and mining. So, you know, maybe you would want some relationship between employers and employees where in exchange for the employees being a little bit more agile and nimble and not insisting that even the employees who are the worst of the worst can't be fired, the employers would give a little and say, all right, here is our AI plan.
• 35:16
  on the basis of how much better the AI models have been getting in recent history. This is what we anticipate we're going to be able to reduce the workforce by over the coming 18 months. And here's our plan for how to do so. Here's how those losses will be managed. Here's how much notice those people will be given. And if necessary, then the government should force companies to do that and should say, okay, every layoff starting now is going to have to be taking place in the context of...
• 35:42
  If it's an AI-caused layoff, then there have to be this set of checklists that you go through in order for us all to manage the coming wave of layoffs in as emotionally un-traumatizing a way as possible and therefore societally undisruptive way as possible.
• 36:00
  The third point that I am arguing for after talking about an ad campaign explaining to people what's coming and how AI is going to change things, and after workplace reform so that companies have to manage their AI layoffs rather than just spontaneously having a knee-jerk decision to lay off a quarter of their workforce. The third thing would be tax reform. This is probably the trickiest, right?
• 36:25
  Because in all of these visions of a future of pleasure and plenty, where we're all cavorting around lakes of milk and honey, and the lion is sleeping with the lamb, and we don't have to work anymore, there is also always the question of, well, how do we get money exactly? What are we using to buy things if we're not working? Some people say, well, okay, a UBI, universal basic income. Everyone's just going to get money. Fabulous.
• 36:54
  I think there's going to be a lot of impediments to that. I think there is a kind of Protestant work ethic in America. There's a puritanical sense that people should be working for something. There's a resentment towards the idea of welfare and the dole and people just bludging while other people are being productive. So as long as there are some workers in the economy, I think it's going to be very tricky. If 90% of the workforce is out of work and 10% are making 100% of the income and they're all massively rich,
• 37:24
  I don't think that's sustainable. I mean, I think you're going to get a revolution. In which case, what exactly does the tax regime look like to take the profits from the companies? Because remember, when a company lays people off because of AI and it doesn't need those people working for it anymore...
• 37:43
  the value that those workers were producing is not lost from the economy. I mean, the company isn't laying them off because the company wants to get poorer. The company is laying them off because it's going to be able to create the productivity that the worker was creating in other ways, that is, by using AI.
• 38:02
  So, if a company is spending 10% of its revenue on paying workers to do something, and it can then fire that, that 10% of its revenue, that value of that, is not vanishing out of the economy, it's just going onto the profit margins of that company. Right.
• 38:21
  So you do somehow have to devise a tax system which captures some significant portion of that productivity gain that the companies that are employing AI are enjoying in order to compensate the people who are being fired. Now, whether that means welfare or UBI, which I think is not the best option, or whether it means some form of retraining or some form of AI-adjacent industry that can be funded by profits, who knows? Yeah.
• 38:48
  The other trick here, of course, is that unless you're the United States or China, it doesn't make much sense to do such a tax reform within your own borders, because then companies are just going to try to go offshore as much as they can, or they're going to try to create shell companies in Ireland or some other place where the tax regime is more advantageous to them. It's just going to encourage creative accounting. So in some way, you sort of need to have some kind of a global alliance, right?
• 39:15
  I mean, this is where it gets really harebrained because it's hard to imagine it actually happening. But ideally, if we're going to manage this thing properly, we're going to need some kind of an agreement where either countries like Australia, medium-sized countries, are able to capture the profits of offshore tech companies in addition to the profits of the companies, the non-tech companies that are deploying AI here and that are enjoying far higher profits as a result of having laid people off. Yeah.
• 39:41
  Or you're going to need an agreement between the US, the EU, and the sort of scattering of other non-US, non-EU countries like Canada, Australia, New Zealand, Japan, South Korea, whatever, Brazil, to say, okay, everyone, we're all going to agree that there's like a minimum amount that tech companies are going to have to contribute back into compensating for the disruption that the technology is causing.
• 40:08
  Some tech people in Silicon Valley are even talking about nationalizing AI, that AI will get to a point at which it will be so powerful, so potentially hazardous, so capable of super empowering individuals to create bioweapons and nukes and whatever, that it's going to need to have power.
• 40:28
  government oversight involvement and if that's the case why shouldn't the government see the upside of the profit as well which means that basically the people that all of us become shareholders in ai and everyone all 350 million americans or whatever get one 350 millionth of some ai company i don't think it's i don't think it's crazy even though it sounds crazy now or put it this way
• 40:55
  I hope that it's not crazy because if we can't do something like that, and if AI is as disruptive as we are almost certain it's going to be, I don't think there's another way out of massive societal and cultural disruption and economic dislocation and possibly revolution or the decline of liberalism unless we can figure out an economic way to share the spoils of the increased productivity. Yeah.
• 41:20
  And the fourth thing is what Joel Pearson is really interested in, which is psychological wellness and resilience and connectedness with other people and mental fitness, because the next few decades are going to require something of us. They're going to require mental resilience, agility, fitness, wellness, a kind of probably contemplative practice, like some way of grounding yourself amid the maelstrom,
• 41:50
  Because old certainties are not going to apply. And everything you try to grab onto that's external to you to provide you with solace and support is likely to slip out of your hands. So you can't, it's like sort of squeezing the soap so hard that it slips out of your hand. You can't keep going around trying to find clarity and certainty and comfort in...
• 42:11
  in the traditionally stable institutions of the workplace and, you know, the government and political parties and so on. It's going to be disorienting and therefore we have to sort of train ourselves to find a way to ground ourselves both internally and in our communities differently.
• 42:28
  that uh involves a bit more touching grass and a bit more uh a bit more breathing maybe we all need to do the like the wim hof breathing method we don't yeah we just just or ice baths ice baths get an ice bath in a sauna you'll be fine um and uh are there any other comments that we should talk about before um uh
• 42:49
  One person says, nationalizing AI in a country that can't even get everyone healthcare. Yeah, I know. It's tricky. That brings me to another thing that I will talk about on another date, and I'm glad you reminded me of it. American lefties seem to think that socialized medicine, like the good kind, Medicare for all.
• 43:11
  would just mean that Medicare as it exists in the United States, which is an incredibly generous, opulent, expensive, and overblown program, could just apply to everybody.
• 43:25
  If you actually crunch the numbers, you do not want to bankrupt the country with that misbegotten adventure. And so something that I try to impress upon my American lefty friends is that in countries with functioning, good, popular Medicare-for-all systems...
• 43:44
  And I exclude from this the United Kingdom, which has a strange system where the healthcare system is actually run by the government, the National Health Service. In most countries that have universal healthcare, the government pays the bill, but private providers provide the service. So that's the way it works in Australia and Canada and so on.
• 44:03
  that we have rationed care. I mean, significantly rationed. The government says, this is what we think this thing is worth. Like, do you remember back, just to harken back, do you remember death panels? Does that ring a bell? Death panels? Remember when Obama was introducing Obamacare and Sarah Palin came out and did a scare campaign about how there's going to be death panels, which are going to be a bunch of
• 44:23
  pointy-headed, like, government bureaucrats who are decreeing that grandma can't get such and such a benefit because the government's not going to pay for it. And all the lefties were like, this is absolutely ridiculous. Sarah Palin is scaremongering. There aren't going to be any death panels. And I was living in the States at the time, and I was thinking, well, I mean, there sort of are going to have to be death panels if by death panel you mean some mechanism by which...
• 44:54
  care is targeted to the neediest people. I mean, death panel isn't a very nice way of putting it, because obviously you're not going to let somebody die if they're an otherwise healthy person who could be cured by a treatment, then the panel is going to approve that.
• 45:11
  But, yeah, if grandma's 92 and she's incredibly expensive and there's some relatively ineffective treatment, say, for Alzheimer's or something, which we know does not enhance life very much...
• 45:27
  And the adjusted life expectancy of giving her an extra four months of totally demented existence is millions of dollars. And those government funds could otherwise be used to vaccinate 10 million young girls against cervical cancer.
• 45:48
  then yes, the government will make that adjudication and call that a death panel, if you will. So the way that healthcare is run in America is this weird sort of schizophrenic bifurcation where the people who have really good health insurance get ridiculously lavish care. And I think that they think, if they're lefties, that everyone could just enjoy that if only the American political system wasn't so dysfunctional and if Republicans weren't so mean. Yeah.
• 46:16
  They couldn't. That's not the way it would work. Every system has upsides and downsides. In Australia, I always pay a copay to go and see my family doctor. Technically, I shouldn't have to because I should be able to find one that doesn't charge a copay, but good luck. And what's my time worth spending 45 minutes waiting in a waiting room instead of my doctor being on time?
• 46:36
  similarly dental isn't covered you have to pay out of pocket for dental there's like you know so not we don't have this lavish american system for everybody and i think the sooner that left-wing americans are able to reconcile themselves to the fact that a universal health system would for the wealthy or rather for the middle class entails some trade-offs uh the healthier the conversation about it will be so thank you for reminding me of that
• 47:01
  Oh, do you still lift? Someone said, how's it going? I don't. I fucked my shoulder. I gotta get back into it, man. Look, I mean, I'm getting squishy. I'm getting squishy around the face. Look at that. Slip on a shirt, slop on sunscreen, slap on a hat. Oh, did I get the slip and the slap back to front? Yeah. Okay. Slip on a...
• 47:21
  Do you slip on... Yes, you slip on a shirt. You don't slap on a shirt. Okay. Read I Am Not a Robot by Joanna Stern, writes Jeff. She spent a year engaged with AI in every aspect of her life. And another person says, I'm probably the only one here that thinks this, but a city full of knife fights sounds amazing. All right. Well, lucky you, because you're about to get one, metaphorically speaking. Okay.
• 47:48
  uh thank you very much for tuning in thank you for listening go and do the survey uncomfortableconversations.substack.com and then find the survey on the tab i want to know what you think we're at like 450 and counting i want to get to thousand or thousands and if you get to 10 000 i'll do a show naked okay we get to 10 000 responses i'll do a show naked you can take that to the bank thanks see you next time