How to Avoid the Partisan Trap Even at The Washington Post w/ Megan McArdle

26:59And so there's all of these cross-cutting incentives that actually make it much harder for us to control costs than other countries. Although I will say this, we often think about ourselves as having a cost growth problem, and we actually don't anymore. We had a cost growth problem in the 80s in a big way, in the late 90s and early 2000s in a less dramatic way, although still significant. And now costs have roughly stabilized as a percentage of GDP, with the exception of the pandemic. Like, it's a big exception, but that was an emergency event. That's not normal. Our costs aren't growing faster than they are in other countries.

27:36We just started from a higher level. We let our costs get out of control, like, 30 years ago, and we never brought the level of cost down. But we did stabilize the growth, and we're now at a level that's, like, basically sustainable. But it is still going to be frustrating to people that their doctor will say, well, let's do this, and then the insurer will come back and say, oh no, you have to have the generic instead of the brand name. You have to have this procedure instead of that. We're going to make you wait or get pre-authorization or all the rest of it. Try physical therapy before you get surgery.

28:10That frustrates people and they blame the insurers for just being greedy and hurting people for profit. The insurers are not doing anything appreciably different from what a government insurer does elsewhere. the government makes exactly those same kinds of calculations and they're in general less generous

28:30Coleman Hughes:than what uh american insurers do i remember looking at the actual profit margins which it's you know you can google of of like united healthcare and all these other insurance companies and they were like low single digit percentages in recent years which is like you know not that much higher than a grocery store or or like a like a you know so i read the manifesto and i thought this guy understands absolutely nothing about the health care system not that i'm any kind of expert but i already understand it better than he does and and i essentially felt the same thing which is that there's nothing more psychologically painful than getting the rejection from your insurance saying we can't cover this procedure, that you formed an expectation is the best.

29:20Coleman Hughes:So it's the natural object of hatred, but it doesn't actually mean that it's in any way the root cause of the dysfunction of the system. No, the root cause of the dysfunction of the system is that we have like nine different systems, many of which are created by the government one way or the other, right? There's obviously Medicare and Medicaid. It's also employer-sponsored health insurance, which exists only because during World War II, when there were price controls, companies were looking for ways to lure workers to work for them. And health insurance was exempt from the price controls. And so they used this as a fringe benefit to attract people to their companies.

30:00And then they made it more and more generous because there are tax benefits for doing this. This is all entirely weird. We've also got the VA. We've got the Indian health system, right? We've got TRICARE and the military. We've got all of these different systems and people, it's not seamless moving between them, right? People get into the cracks. We've then got all of these individual insurers who are competing less than in other countries on a really tight regulated package. But we also have a legacy of bad policy design in other ways. So, for example, doctors and nurses and basically everyone in the U.S.

30:40healthcare system makes more than their counterparts abroad. And that's because when Medicare was started in the 1960s, they didn't have a fee schedule. They just said usual and customary fees. Well, doctors had a lot of latitude to decide what was usual and customary. And so their income started going up really fast. And then, you know, being a doctor in 1950 was not a really good high-income job in the way that it is now. That's a legacy of Medicare. And then they started trying to control this. And then the costs moved to hospitals. and it's been my husband actually wrote a great article for reason magazine which i commend to all of your listeners and this is now 10 or 15 years old but called medicare whack-a-mole every time they would whack one of the cost centers another cost center would pop up and the money would just flow into that channel instead and we've been trying for decades to get a handle on that and we're never going to i think is the actual answer is that we're just going to kind of muddle along because the other piece of this is we've done a lot of bad policy design, but each piece of bad policy design has created stakeholders.

32:03And those stakeholders will freak out if you try to take anything from them. So obviously the doctors and nurses, they are extremely effective at not having their incomes cut. But you also have, you know, what Obamacare wanted to do, really wanted to do with the kind of vision was if they had had their druthers. if you talk to people you could it would it was basically with some details changed either people just wanted a single payer system or they wanted something like the swiss system or the dutch system where basically there's these tightly regulated insurers they compete on a package of benefits and on like customer service and how nice their offices are and but they're they're basically all offering the same stuff and that everyone is in that and the government just picks up the tab for the really poor people.

32:53And I think that if you asked a lot of the architects of Obamacare, that was really their vision is we're going to do this system where everyone's getting good insurance. It's all the same stuff and so forth. And they couldn't do it. And why couldn't they do it? Because people who had employer sponsored health insurance went ballistic. People say like Medicare for all pulls well. and this is kind of true and kind of not um the trick there is that you pull pieces of it should people be able to buy into medicare at any age that will pull well should people be able to buy into medicare at any age if this costs us like 8 000 a person in the united states and then you know it's trillions of dollars that is not a popular proposition and similarly people are fine with add-ons, what they're not fine with is losing what they have.

33:46And what they have actually is part of the problem in some ways. Now, look, I support a really free market system. My preferred healthcare plan would be that there is no healthcare plan for under, say, 10 to 15 % of adjusted gross income. You sort that out yourself. If you have catastrophic expenses, the government will pick up at the tab, 100 % of the tab above whatever that level is. But almost anything would be better than the mess we have. But you can't touch it. It's like, look, I love Kaiser. I don't want to leave Kaiser. I don't want you to take it and give me Medicare for all and promise me it's going to be just as good as Kaiser because I don't believe you.

34:25And you probably shouldn't. I'm probably correct not to believe them. And so that creates a huge obstacle to reform. If you really can't touch anything that anyone has in a serious way. You are just really, really constrained in what you do and what people end up doing. What reformers end up doing is that they've been like, okay, well, here's this gap. I'm going to add, I'm going to do another program here. And that just makes everything worse, right? Because now you've got yet another complicated piece of the system in which there's new gaps that arise in between, there's complex interactions.

35:00And so we're probably never going to have the dream system that I, partly because the dream system that everyone wants doesn't exist, right? If you talk to liberals, there's like this fantasy European system where you just go to the doctor and your bills are negligible and they cover everything. And it's amazing, right? And that's, that's not real. That is not a real system that exists anywhere.

35:24Coleman Hughes:Tell me what's wrong with this opinion, if anything. In America, the worst thing that happens to that happens still on a regular basis is that you get hit with, you know, several thousand dollars in an ambulance bill. And you know, this is a, this is the kind of thing that can and does happen in America. In America, you can pay with money in the extreme case. In the UK, you can pay with time in the extreme case. So you see these cases, which New York Times and others have reported on where you call the 911 equivalent in the UK and the ambulance gets there eight hours from now, right? So is it the case that, you know, Bernie Sanders and probably Luigi Mangione, they want a single payer system?

36:15Coleman Hughes:Isn't it just true that in an American style system, you can pay in money in the worst case, but in the single payer system, you're likely to pay in time? Yes, with a qualification, right? Single-payer, there's lots of different kinds of national health insurance programs. Single-payer, which is what they have in Canada, like they're single-payer. There's also like what they have in Britain, where the government's running the health system directly. You have Canada, where it's single-payer and there's no real private competition. You have systems like Australia, where there's a basic government system.

36:53So there's lots of different ways to do this. But yes, I would say that the more the government is the loan provider in one way or another, either because they're paying all the bills or because they're actually directly running the system. The more that that's the case, the more you are likely to end up in a situation where you're rationing by queuing, which is the technical, right? You stand in line. And if you look at the wait times in the UK and Canada, they are to my eye insane for things that are elective in the sense that getting a hip replacement is elective but having watched my grandmother go through that it wasn't very elective because she couldn't walk until she got one right so she just had to and she was in terrible pain every time she had to get up and go to the bathroom or anything it was really awful to watch and waiting a year or two for that surgery is a big deal.

37:49Waiting a year or two for your imaging. I mean, like when people tell me the wait times, I'm sometimes shocked when they'll say, yeah, I waited six months. Now, this is not true for everything, right? Like I have definitely talked to people who got pretty timely cancer care. Trauma care tends to be, you know, if you roll into the ER with a gunshot wound, they're not going to like put you on a list. But more and more stuff goes to the back of the line. And why does that happen? Because when something is growing in the private sector, the government doesn't have either the incentive or the ability to stop it from growing.

38:27When something's growing on the government budget, they can just say no, right? And when they're under budget pressure, that's what they do. And so the reason that the UK and Canada have the biggest problems is just that they spend less money on their system. I mean, Ireland spends very little on their system and should really spend more. But they could provide faster service if they spent more money and built more capacity. But the political incentive is not to build capacity because it's so expensive to maintain. And you're always making this trade-off between, well, I could spend this money on nine other things the taxpayers want, or I could reduce wait times.

39:12And wait times just tend to go to the back of the queue. There's just like, I did not mean to make that pun, but they just, they tend, it's very hard to fix wait times because if you think about the, there's a lot of things in life where the last mile gets progressively more expensive, right? So if you're building a fiber optic network, Laying the big trunk lines is not cheap by any means, but it's actually even more expensive just to get the last mile to someone's house. And the reason is there's a bunch of houses in between you and them, and it's going to take a lot, whereas they tend to lay big lines out where there's no people.

39:53And so there's just a lot of things that have that characteristic. Getting the last bit of reliability in a system, and the way I think about this is, in the heyday of Uber and Lyft, when people were like, I'm not even going to own a car. I'm just going to sell a car and just take Ubers and Lyfts because they were being subsidized by people who were trying to gain market share. I never hear that now that the prices have gotten semi-rational. But, you know, I asked the reason we didn't sell our car, even though we drive it so infrequently that we've had to replace the battery some ridiculous number, I think three times in the last three years because we keep forgetting to drive it.

40:31And then the battery dies. but we have two very large dogs. And if one of those dogs needs to go to the emergency vet, he needs to go right now, not when I can get an Uber pet. And so that one use case, I'm maintaining a car for basically that use case and a couple of other small use cases. And otherwise I could totally get around on public transit, my feet, et cetera. But I have to have this extremely expensive vehicle that they then have to pay thousands of dollars to insure and find a parking space for and gas. and maintenance for this one use case, right? That use case is more expensive than every other thing I do.

41:08Those last things are often just very hard to get. And so getting enough capacity in the system to take care of 95 % of your cases can cost less than getting enough capacity to make sure that you have 100 % under two hours or whatever. I mean, it's not literally going to cost less, but that last 5 % of spare capacity that last 10 % of spare capacity is phenomenally expensive because that's where it's lumpy, right? Your first 80%, it's just like people get colds, they do this, they do that. You have, it's pretty, it's a pretty predictable flow. And then you can just keep, but that last, um, that last five to 10 % is the stuff that doesn't happen actuarially in neat, like little bundles.

41:57It happens, you know, the ER come, you go to the ER, you have the ambulance demand, and suddenly you've got 87 cases. Well, if I had all of these spare ambulances, I could just send it. But having all those spare ambulances means I have a lot of ambulances with crews that are often underutilized. America makes the decision to just have a lot of redundancy. Other systems, because the government's paying for it very directly and is making a budget, not like the U.S. where the government's paying for it and they find out what they're paying after. the government is paying for it in advance, making a budget, deciding how many ambulances to buy.

42:33And they decide to buy fewer than would cover all of that peak capacity. Right. And that's like, that's how those calculations, that was a little, that was a little shaggy dog apologies, but

42:45Coleman Hughes:that's okay. This episode is brought to you by what could go right. A podcast that pushes back on the doom and gloom in the news to ask a simple question. What if we're actually making more progress than we think? Each episode, host Zachary Karabell sits down with leading thinkers, policymakers, scientists, and cultural voices to dig into the biggest issues of our time. Episodes cover conversations like, can we achieve super abundance without AI doom? Or why AI and drones won't bring the apocalypse? Listen to What Could Go Right on Apple Podcasts, Spotify, or wherever you get your podcasts. New episodes drop every Wednesday.

43:28Do you ever gaze longingly at your old NPR tote bag and wish you could call up the car talk guys and ask them how to fix the meeting crisis? Do you get AFAB confused with ACAB, but are too embarrassed to admit it? Do you long for a social life that was a little more nuanced? Sounds like you might need the unspeakeasy. The unspeakeasy offers intimate, off-the-record gatherings with incredible guest speakers, amazing food, and life-changing conversations. This October 19th through 21st will be in New York's Hudson Valley with John McWhorter, Jungian analyst Lisa Marciano, and rising theater star PJ Adzema.

44:06Spaces are very limited and going fast, so go to theunspeakeasy.com to learn more. If you're listening to this podcast, use promo code FREEPRESS at checkout. That's one word, FREEPRESS, and get$100 off your retreat fee. That's the unspeakeasy, home of the original free speech vacation.

44:28Coleman Hughes:Let's pivot topics now and talk about artificial intelligence, the large language model revolution that's happening now. You've written quite a bit about this. And so there's a few different buckets we could talk about. the first I want to touch is like what's going on with education and AI in the classroom? How has ChatGPT been sort of affecting how teachers teach, how students learn and so forth? So to give you my window on this, I recently taught for three months at the University of Austin, a small seminar course, and just this past January through March. And, you know, my view as someone who loves chat GPT personally, I use it every day.

45:16Coleman Hughes:I think I use it so much because I'm very intellectually curious and there's just always something that I want to know about randomly. So it's like a massively fun tool for people like me. And so I have like a generally positive outlook on it. I think people should use it. I think students should use it. And regardless of whether they're what they should they're going to. I'm also I don't believe that professors in general can detect when their students have used it. I'm not saying you can never detect it. Obviously, there have been many students have been caught using it on essays and so forth.

45:55Coleman Hughes:But I think in general, students have a greater incentive to cheat really well than professors have to catch them. And obviously the dirty little secret is that similar kind of to how you were saying doctors, doctors don't like to tell their patients that there's nothing more I can do for you. You're, you're a few months from dying. Professors don't want to catch their students cheating because it's a hassle for us. Like it's, you know, God forbid, if I'm right about your cheating, that creates paperwork and meetings for me, like literally creates extra work that I'm not getting paid for. You know, and even let's say I have a really good bullshit detector, I'm right 95 % of the time, well, one in 20, I'm going to accuse a student of cheating, who in fact, did not cheat.

46:44Coleman Hughes:That creates a much, much bigger headache for me, because now I've actually harmed a student and that now I could potentially be in trouble, right? So professors Professors have almost no incentive to catch, no personal incentive to catch cheating. Students individually have a pretty strong incentive to cheat successfully because your grade goes up. So I just, yeah, I don't, from an incentives point of view, I don't buy that professors as a class are going to win the war on cheating. And so my approach as a professor was to say, fine to use GPT as your personal research thinking buddy tutor at home.

47:25Coleman Hughes:there are going to be no, no assessments that you could even possibly cheat on in this class. It's all going to be blue books. Uh, and if I could do it again, I would also go for the oral exam for a variety, like maybe, um, yeah, in class quizzes, class participation, blue books, it's a cheat proof class. And that's, to me, that's gotta be the future of, of, of teaching. So what do you make of this? Yeah, I have. So I taught a class. I've taught a few journalism classes. And the thing I always taught them pass fail. And the thing I always said was, I'm not going to try to catch you cheating. You are adults.

48:07You have a choice. You have a Washington Post columnist here, a Bloomberg columnist, who is going to teach you as well as she can what she knows about her profession. if you want that. This is pass-fail. If you show up and you hand in something, you will pass. But you can get... I'm not going to try to make you get, for your$250 ,000, the value that's here. Whatever that is, maybe there's no value. Maybe they should just come and sleep through my class. But I can't... You know, you're now... You guys are in your 20s. You were old enough to go get killed in a war. And you were old enough to... Like, not old enough to drink because somewhat irrational policy laws.

48:50But my job is not to make you get something out of this. My job is to make it possible for you to get something out of this. But that was easier for me to do because it was pass-fail, right? I think you're right that the incentives are bad, but I also think there's a bigger question, which is that you know what i was saying is this is a pass fail class it's like a half credit it's not the difference between you going to grad school and not so you should enjoy this class you should only be here if you want to be here and learn something and you should get what you can out of it be in the discussion do this right like don't sit here and passively just try to you know put the time in and grub the grade get what you can um but that's not really realistically how most people think about an education right the the way most people think about a college education and especially the incredibly expensive four-year residential college experience that both you and I had, where you are not majoring in engineering, is that the primary reason people are willing to pay six-figure sums for that is to get the kid a good job when they get out.

50:11And the primary reason that that works is that getting into one of those schools, getting into an elite school signals that you have usually a class background, that you have a certain network, that you have a certain amount of cognitive ability, a certain amount of conscientiousness. Not that much in my case. I was a really terrible student. But that signals that to employers. If you are coming from one of many schools, you have signaled that you can put your butt in a seat for four years, show up, turn in your assignments on time. And the value of what you actually learned in that class, it's quite unclear whether they value that at all, whether there's a low value.

51:00But if you look at the sheepskin effect, which is that there is a real discontinuity between finishing all but one semester of college and finishing college in the labor market. And it's not because they wait until the last semester to tell you all the good stuff, right? You are signaling a lot of stuff about yourself.

51:19Coleman Hughes:This is Brian Kaplan's argument Yes, this is Brian Kaplan's argument, which I basically buy. But one of the things that it signals, right, is that you had to do all of these complicated, if you're a humanities major, you had to do a bunch of complicated research tasks, read a lot of stuff, assimilate it into something, some basically coherent idea about the world, and write that down in sentences that could be understood by another human being. They may not be the most glistening, beautiful prose in the entire world, but they can be read and comprehended. And the value of that skill is itself declining.

52:02AI is going to undercut the value of reading and writing, of doing what I do, what you do, which is reading and assimilating vast amounts of text and then translating it into words for other people. and colleges, I think, have not grappled with that more fundamental problem. I mean, first of all, they haven't grappled with the cheating problem, you're right. When I talk to academics, they're all just kind of home-baking solutions or not. Some of them have just given up. Some of them are being really innovative and designing assignments that you cannot cheat on because the object is to use AI to do something cool.

52:39That's like Mike Munger at Duke is doing this. but a lot of them are terrified and I think the even bigger problem than that is that beyond the fact that they have the value of this signal is decreasing is that like society loaded all of these different tasks onto universities it loaded dating finding a you know finding a spouse and or at least finding some friends who can find you your spouse later that's less relevant to your generation than to mine, but to my generation, that was a big thing. You would meet your boyfriend in college and hopefully you would marry that person, right? Not in the way that my mother's generation would joke about getting her MRS degree.

53:23We were also there to get a job, but, and hopefully maybe even learn something and become more interesting people. But there was, you know, dating was a big feature of it. We loaded finding, sorting people for employers onto it. We loaded training onto it, onto this medieval institution that weirdly, the people teaching it have no idea, for example, what's happening in the job market at all. They're also really not selected for their ability to use AI well, right? If you wanted to design something that is going to prepare kids for the world of AI, if you were starting that from scratch, you would not go out and be like, you know what I need is a bunch of experts in 14th century Renaissance literature.

54:11That would not be your first choice, right? Who have spent their entire lives in archives reading and writing critical theory about their period, or political scientists, or go through the disciplines. And to be clear, I'm not denigrating what they do. I'm saying that if you were starting from scratch, you definitely would not be like, Like, yes, what we need is a bunch of PhDs in sociology to handle this task. And that thing, and also, so the university is really unoptimized to handle the problems of preparing the workforce for AI. The signal, the thing, it wasn't great, well-optimized to handle the problem of preparing people for the workforce of 1990 or 2010 either.

54:59but that didn't matter because the value of the signal was itself valuable. It was signaling abilities, some of which had been honed at university and some of which hadn't, but it was signaling a bundle of abilities that were valuable in the job market.

55:11Coleman Hughes:Hold on. Can I stop you right there? Presumably, though, today it's still signaling a mix of IQ, conscientiousness, just maybe to some degree competency at a certain higher, upper-class way of holding yourself. And wouldn't those all still be valuable even in a post-AI world, even if this more narrow ability to basically do our kind of work is being competed away? Sort of. I mean, I think that that's true for people who were in college when chat GPT, but it's turtles all the way down. High school kids are using it too. and so ultimately what it may be signaling is that you are smart enough to cheat without getting caught which is I guess a kind of if you went to employers and were like do you want a bunch of people who are smart enough to cheat without getting caught they might have questions about for example should you put those people in your accounting department and so I think Like, I do think it's a problem.

56:28Like, the value of the signal is going to degrade so strongly. And the real value of the skill is also going to degrade. And the even bigger problem, now that I've completed it, is that the governance of universities is completely ill-equipped to handle any of this. Right? Everything's run by committee. No one's in charge. And so no one has the ability even, you know, They're all forming committees on how to do AI. But what happens when the committee reports is, are people going to go out into the classroom, show people how to do it right, and then check back and make sure we're getting a valuable assessment tool?

57:06No. They're going to put out some guidelines, and the professors will implement them with varying degrees of precision, and then that will be that. And the problem is a university is a bundle. So you can be Mike Munger and teaching a great AI-optimized class. I've never been in his class. I know him personally, and I assume he's doing a pretty good job. But if your colleagues all over the university are bad at this, and people are noticing that graduates are coming out without skills that are useful to them, it doesn't do you any good to be the one guy who's good because employers don't know who you are.

57:46Right. And so there's a real problem of how are you going to reform this when you have a faculty that's selected for their research abilities, not even their teaching abilities, and certainly not their AI abilities, when you cannot kind of reform the institution very effectively, when you can't reform classroom practice very effectively. I think universities have a massive problem that they have not really begun to confront. There are exceptions to this. People like Hollis Robbins are writing great stuff about this. But in general, I don't think they understand the scope of their problem. I think they understand the narrow problem of like, we have a lot of kids who are cheating and we got to fix that.

58:30But you can fix it within class assessments, right? You can fix it that way. But those bigger problems of how do you make this institution that has basically been hiring its faculty for 50 years based on their ability to publish articles that almost no one else ever reads on quite narrow, specialized topics, and which has this crazy governance system that makes it very hard to change anything. How do you take that institution and transform it into an institution that is preparing kids for the AI job market?

59:11Coleman Hughes:So if you were the czar and had complete control over a top college, what changes might you institute?

59:21well i think i would probably at an even more basic level i think that in a lot of the humanities the whole publish or perish model is bad i think there are like really deep problems with the way that academia hires and promotes and i understand the institutional reasons that happened but the this obsessive focus on who has the best publication record, right? That that's what, that is what has status in academia. Not teaching well, not any of the rest of it. It's nice if you teach well. They're happy if you teach well. They admire people who teach well, but that's not what has status in their world.

1:00:04And there is an argument for that in the sciences, although I think even in the sciences, it's not entirely clear to me how much do you really benefit from being taught by a Nobel prize-winning physicist, being taught physics 101 by a Nobel prize-winning physicist rather than a TA who is closer to your age and better at teaching? I don't know the answer to that. But I just think this obsessive focus and then the humanities got science envy and they all started doing it too. Well, it's not that you shouldn't be publishing if you're writing about Renaissance literature or something, but it's not like there's stunning new developments in Renaissance literature.

1:00:41We have all of the books. They're not, I'm, I'm slightly, I'm going to get really mad English professors writing to me. And so I'm not saying that it's not valuable exactly, but that you should not, it should not be like for an English professor. I think if you want butts in seats and English is losing majors incredibly fast, I think in part, as a result of this, if you want to bring people into the classroom, you don't need to communicate your incredibly narrow theory, what you need to communicate is a love of books. You know, I just finished Middlemarch, which I guess has gone a little viral on Substack and everyone's reading Middlemarch.

1:01:18And I was also reading it, but not because I saw it on Substack. It was so good. I was sad. I was both elated that I had read it and I was sad because it was over. I don't have that many, that experience with that many books these days. But that's, to me, this is the purpose of bringing people into English classes. Right? It's not to teach them to identify the socially problematic behavior in people who have been dead for 300 years. is it's to teach, it's to show them the human experience, to open their minds to another time, to let them see, right? And that's not what they're selecting on. They're selecting on, have you written articles that five other professors have read in your subfield?

1:02:13I would just stop that. I would evaluate people for teaching, not research, in a lot of fields where I think, and I wouldn't stop them from doing research. I wouldn't say research is bad, but I would just say, you know, it's okay if you don't publish. If you were a great teacher, my best teacher in my English major, or one of my best two teachers, he taught a class on Twain and it was fantastic and everyone loved him and he didn't get tenure. And I eventually asked someone why that happened. He ended up at the University of the Pacific. And I went back to Penn and I was doing a thing at the writer's house and I asked why that happened.

1:02:54And they said, well, you know, his research wasn't really that impressive. And I was like, okay. But he had all of these students who came out of his classes on fire for Mark Twain. And he also, and he said, well, students aren't necessarily a good judge of what they learned. And I was like, I remember what he taught me. And let me tell you, the people who are more prestigious and you're a little, I don't remember anything. I remember.

1:03:18Coleman Hughes:I would also argue a la Brian Kaplan that just the amount that anyone retains from any college class is just by default very low. But you can retain a kind of lifelong passion for the topic. And also, he did this really interesting thing with his Twain class, which I recommend to people who are interested in reading literature, which was he made us read all of Twain's Drek. We didn't just read Connecticut Yankee and Huckleberry Finn and Tom Sartre. We read Puddin' Head Wilson. We read the stuff he was churning out by the fistful when he was on a lecture. He went bankrupt in the 1890s and then spent the next five years constantly obsessively writing and lecturing, trying to make enough money back, make enough money to pay his creditors.

1:04:09We read all that. and that was actually really good because I think the way we usually teach literature is the kind of highlight reel right you see all of the you see the best thing every writer wrote and instead of that we just saw everything this writer wrote and it was a very different experience and and just it was just a wonderful class I remember it 30 years now 35 years when did I graduate from college no not quite 35 years. But 30 years on, I still remember it. And that's what you can give to students. They're not going to come away. I had three professors there who were really memorable and none of them was the famous, prestigious research.

1:04:54I took a class with Elisa New, who ended up at Harvard. And I don't remember anything from her class except that I wrote something on incidents in the life of a slave girl and Uncle Tom's Cabin. And I juxtaposed them. And in sort of typical precocious, I guess I was a sophomore or a junior. I was a junior. I decided that the interesting thing about these books was that actually like the incidence of a slave girl, which is a real slave narrative, was in some ways kinder to her owners than Uncle Tom's Cabin. And that really struck me. And so I wrote about Stockholm Syndrome and this thing where you identify with your captors.

1:05:48Only I didn't define it. I never, I was just like, I'm such an arrogant little, like obviously everyone knows what Stockholm Syndrome was. so I got it back from the TA and I got it back like a month late because I had turned it in a month late and he said this was one of the best papers I ever read I didn't understand all that stuff about Helsinki but really well done and I couldn't remember what I'd written about I was like Helsinki? where did Helsinki come into my yeah and then I read it and I realized that I hadn't defined it and he had no idea what I was talking about and this is pre-Google, which is important for you youngins to understand, is there was no way to just go and find out what a term you didn't understand meant.

1:06:34And he could not possibly have understood this paper. And that somehow this had made him think it must be very good rather than giving me an F for failing to find my terms. That's the only thing I remember from her class. I don't remember anything she said. I'm sure she's a fine teacher and that many people have had very, I don't want to, you know, sort of slam. But it's just that like that thing would to me, what is for undergraduate English, which is not going to produce a large number of researchers and professors, right? What is it going to produce is a lot of people who have read books. And you can either make that a wonderful, exciting experience for them that inspires a lifelong passion for literature, or you can make it a dull rote experience for them, in which case, what the hell are you doing?

1:07:24Coleman Hughes:Yeah, I think from my experience, limited experience teaching, I think of teaching definitely a seminar course. But I guess I think it would also apply to a lecture in the humanities, to be clear, not in statistics or engineering. But I think of it a little bit like being the warm-up comedian in a comedy show. Because if you go down to the comedy cellar, they don't just hit you with the first comic of the night. There's a warm-up comic that it's really essential to the show because it sets a tone and sets a mood. And only once that mood is set is the audience really capable of letting go of themselves and fully surrendering to the comic mode of being.

1:08:11Coleman Hughes:And, you know, so a lot of it is like my job is to set a tone and a mood in the room such that once that mood is set, the students will start accessing parts of themselves and their minds in relation to the material that just creates a kind of magic conversation that is just like generally hard to snap into. do. And so, and once that mood is set, like you almost don't have to do that much because suddenly they're so into the topic for these, this hour that they end up having some of the most profound and unusual and unique and out of character conversations with their fellow students than they would ever normally have.

1:08:55Coleman Hughes:Like that's really what it's about. But I think for a lot of professors selected on this niche research model, their default mode is like, impress my peers, impress about 50 peers in the world whose opinion I care about. And I think a lot of them who lack self-awareness maybe don't know to take off that hat when they enter a classroom and become like the warm-up comic. Yeah, I mean, I think that's a really big piece of it. But it's also just that, look, teaching is hard. Not everyone's good at it, right? There's lots of things I'm never going to be good at. And that if you are selecting mostly on research ability, if that's what you really care about, you're going to get not as good teachers as if what you really cared about if you were selecting on teaching ability, right?

1:09:49And there's some overlap, obviously. being familiar with your topic matters, et cetera, right? But I think there's a threshold beyond which, and I think this varies by discipline again. If you are in the sciences, it's probably really, after kind of introductory courses, it probably gets progressively more important to be near the cutting edge of your field, right? Like students probably benefit more from being around that. But I don't think there's like a cutting edge of Renaissance literature, right? We have the canon. And lots of people can have really interesting thoughts that will engage kids.

1:10:26But it's also just true that the best person to teach someone something is not always the person who's best at it.

1:10:33Coleman Hughes:Oh, yeah, sure. Right? Because they are not, the people who are so far beyond, they kind of can't get back. It's hard to get back to your level and remember what mistakes you made. And this is like, you know, when I was teaching, I was teaching a class on how to write an op-ed and I really had to think about what are the building blocks, right? Not what do I do on a daily basis? Look, I love the sound of my own voice. I could talk about what I do forever, but what is something that for someone who hasn't been doing this for 20 years, what's a way they can, what do they need to understand about the form?

1:11:12what can they access in an easy way. So I basically taught everyone a kind of a four block structure for, I made everyone write 800 word op-eds quite rigid in a way. I actually made them read a book called Save the Cat, which is this incredibly rigid. Yeah. So, you know, Save the Cat for the audience. It's this book about screenwriting, which is incredibly prescriptive. Like this happens on page seven. This happens on page nine. and I made them read that to understand and I did that in some ways if you teach a kid to write poetry I think it's easier to teach someone to write a sonnet than to write blank verse because it's too much freedom when you're starting out you need that guideline that framework gives you more freedom to then think about the pieces instead of having to start by figuring out what it all looks like but it's that's hard for people to do.

1:12:12I got a lot of help from my husband who actually developed my opening slide deck, how an op-ed is like Star Wars. It was originally his slide deck. I stole it and then expanded it because he's smarter, but he's much smarter about structure than I am. But all of that stuff, that's, you know, sometime maybe it is better to have people who are less good at research doing more of the teaching. But I definitely in lots of disciplines, I would not have that.

1:12:40But And I think this is how university presidents feel too. It's so overwhelming. What would I do? I don't know. Get rid of tenure. Assess everyone on their AI abilities and their ability to teach. And then stack rank those people and the people who can't teach, go. I mean, I'm already imagining the angry tweets and blue sky skeets and all the rest of it that I'm going to get for saying that. And to be clear, this is not me saying academia bad. Journalism is going through the same thing, and we don't like it either.

1:13:20Coleman Hughes:We don't like that AI is going to take a lot of what we do, and a lot of good people are going to get totally hammered by that, possibly including me. This is not a personal, it's just a really hard problem. We got a lot of really hard problems. I think there's a huge amount of stuff that is super exciting about AI. Yeah. I think there's the problems are hard. As for journalism and taking jobs, my instinct is that there's two kinds of quote unquote journalistic writing. One is like the article about any given topic that just happened in the news written by someone whose name I've never heard before and who, you know, an AI could have written that article more quickly, or let's say, you know, an AI could downsize that breaking news department by 50%.

1:14:16Coleman Hughes:Those jobs seem to me very much under threat, as would any job that is, you know, you're writing webmd's top 10 ways to cure a cold or something. I briefly had a job writing, summarizing which strains of weed were the best, even though I didn't like smoking weed. I was just trying to break into the writing world, so I took anything I could get. I transcribed earnings calls was my first writing job for$85 a call. Yeah. Well, you got to get in somehow. So those jobs seem to me maximally under threat. But if I think about someone like me or you, people who have personal brands, it's like, why do I read a Megan McArdle column?

1:15:03Coleman Hughes:Why do I listen to a Coleman Hughes podcast or read a Coleman Hughes essay? It's because partly I care that they wrote it because I know something about them. They have a brand that I care about. It's a bit more like, and this will seem a pretentious analogy. I don't mean it to be. It's just like the clearest case, which is like, why is an Adam for Adam copy of a Vermeer worthless, but the original Vermeer is worth millions. It's like, it's not about the product itself. It's about that Vermeer touched it. Right. And there's something about that to listening to your favorite podcasters, the columnists that you follow.

1:15:40Coleman Hughes:So I don't think Like you specifically or me or Ezra Klein or Tyler Cowen or Ross Douthat or Jamel Bowie, like our jobs, I don't think are under threat because people don't read us merely for the words on the page. They read us partly because of the knowledge of who is behind it and whatever their association, even if they hate us, they read us because of the person behind it. Well, knock wood, you're correct. Yeah, my assessment is this, is that there's, journalism has two problems and one is on the supply side and one's on the demand side. And the supply side problem is this can be done by a machine.

1:16:23The demand side problem is we now have yet another thing that we're competing for people's attention with, right? It's, you know, social media, that was true too. Every new thing that competes with us for attention leaves less attention for us. The other thing is that on the supply side, and yes, I think you are completely correct. The stuff where you're just writing up public information, that is going to be less and less done. There's no reason to do it. And I'm not even saying it's not valuable to my institution or someone. people are not going to go to the Washington Post or the New York Times or anywhere else to get a write-up of the latest economic data.

1:17:03Because they can just ask ChatGPT much faster and get a customized... And then unlike when we write it up, when they have a question, they can just say, oh, but what about this? And ChatGPT will tell them. So that, I think, is just going to go away. Where do I think the value in journalism is going to be? It's going to be in people. And I think that in relationships, and I think that also manifests in two ways. One is the value of your relationship with the reader. So people who have a brand, people who have a following, those people are, and I am leaning in more to this. I'm leaning in more to being, I am not writing up the economic data.

1:17:49I'm leaning into the, and I will occasionally when I feel like I have something unique and interesting to say about it, but I'm not just going to do a, well, it's Tuesday. I got to write a column. There was a release because people don't need that from me as much anymore. And having, being a personality, having people feel like they know you, you know, I joke that podcasts are your imaginary friends. And you know, it's, and it's a little weird because I actually know a lot of people on the podcast that I listen to. So sometimes I'm walking along and it's like, I'm having an imaginary friendship with my real friend.

1:18:25And that's right.

1:18:26Coleman Hughes:Yeah. Yeah. I've had that too. Camille Foster is a good friend of mine and I haven't seen him for like nine months, but I feel like I've been hanging out with him, but he hasn't been hanging out with me. So it's a problem. I have this experience quite frequently. But then the other thing is relationships with sources. What is one thing ChatGPT can't do? It cannot get a congressional staffer drunk and get it to tell them things about the tax bill. Only a physical reporter can do that. And so that kind of information is also going to be valuable. And those are the two areas. And so if I think about how this is manifested in my own life, partly it's changing what I'm writing about.

1:19:11I'm writing a lot more about how to think about AI and how to, you know, because I think that's where I can add value for my readers. But I also, I spent this spring on a basically nonstop conference. There were months where it was literally like, I came home usually on a Friday night, sometimes on a Saturday, took clothes out of the suitcase, but the clothes in the washer, put them back in the suitcase after they got out of the dryer and then I just got on the next plane. And it was grueling and there was, but the thing that you can do at conferences is get new ideas, talking to people, be in a group, listen to people, right?

1:19:47And I love remote work. I've been working remote since 2006, but you've got to get, it's not, asking GPT isn't the same because GPT only tells you what you asked.

1:19:58Coleman Hughes:Right. Right? And so that thing of creating serendipity for people, having relationships, hearing things that you wouldn't have heard if you hadn't been in the room, that's the other place where journalism is really going to add value. And so those are the two areas, I think, this is my analysis for what it's worth. That's where I am attempting to focus what I do, which is really intensive. And then I have to, like, I'm taking the summer off. I'm old. I can't do this anymore, but I'll start up in the fall. but doing that thing of being talking talking to as many people as possible being in a physical room with them and then taking that out to your audience is the other area where i think there's going to be a lot of value for journalists or i hope there's going to be maybe i should say i hope there's going to be some value for journalists but i do think you know who do i read when i think about my own industry i read dylan byers and oliver darcy because those guys are on the phone all the time and they're talking to people and they're giving me information that I can't get anywhere else.

1:20:56Coleman Hughes:Right. All right. Megan McArdle, thank you so much for doing my show. Thanks for having me. Hey, everyone. If you've got thoughts on this episode or just want to keep the conversation going, come find me in the Free Press Forum. The forum is kind of like a Reddit page for free press subscribers. And on Tuesday, August 18th at 3 p.m., I'll be in there live to answer all of your questions. Head to the fp.com slash forum or use the link below to start asking now.