In an era where media consumption is often dictated by political affiliations, Coleman Hughes’ insights on maintaining an independent perspective stand out. In a recent discussion featuring Megan McArdle on YouTube, Hughes emphasizes the importance of critical thinking and the need to rise above the polarized narratives that dominate outlets like The Washington Post. This blog post aims to dissect the key points made in their conversation, addressing common misconceptions and providing a fact-check on the assertions made. Join us as we explore how to navigate today’s complex media terrain and encourage a culture of informed dialogue that transcends partisan divides.

Find a fact check of this transcript on CheckForFacts

Transcript:

[00:00:00,000]: Welcome to another episode of Conversations with Coleman

[00:00:03,259]: My guest today is Megan McArdle

[00:00:06,659]: Megan is a columnist at the Washington Post where she covers economics finance and government policy

[00:00:12,420]: She originally gained prominence through her early 2000s blog Asymmetrical Information later writing for outlets like The Atlantic Bloomberg and Newsweek

[00:00:21,940]: She’s also the author of The Upside of Down Why Failing Well is the Key to Success

[00:00:27,649]: In this episode we talk about what Megan’s Washington Post readers get wrong about her

[00:00:32,880]: We talk about what exactly is wrong with the American healthcare system touching on insurance pharmaceutical companies and the role of bad incentives

[00:00:40,639]: We also talk about the legacy of Obamacare

[00:00:44,279]: Next we talk about the influence of AI on education including how to use AI in the classroom

[00:00:50,299]: So without further ado Megan McArdle

[00:00:53,080]: ♪ Go go go go go go go go go go go ♪ Okay Megan McArdle thanks so much for coming on my show

[00:01:02,919]: Thanks for having me

[00:01:04,300]: So I’ve been reading you for a long time

[00:01:06,500]: You’re a longtime columnist at the Washington Post and in my view one of the columnists with the highest batting average in terms of just getting issues right in terms of not being influenced by the partisan crazes on the right and the left on any particular moment

[00:01:26,019]: And so you’ve been a pleasure to read for that reason

[00:01:29,620]: That is extremely flattering thank you

[00:01:32,099]: That’s what I’m aiming for

[00:01:33,239]: Yeah

[00:01:33,660]: Whether I always hit it or not

[00:01:35,660]: Yeah

[00:01:36,819]: So let’s start briefly with your background

[00:01:40,099]: I know you’ve had sort of various odd jobs in the past before becoming a journalist and I know that your father was also a big influence on your kind of intellectual curiosity and so forth

[00:01:54,639]: So how did you end up becoming a columnist

[00:01:58,599]: Entirely by accident

[00:02:01,699]: I graduated from college in the nineties because I am an old lady and I ended up working in I worked for a variety of companies that showed an astonishing ability to go out of business like two months after I joined to the point where a friend of mine who’d become like an equity research assistant at I believe Morgan Stanley it’s a long time now said you know I want you to let me know here’s a list of our companies that we cover and if any of these people makes you a job offer tell me because I’m going to short the stock

[00:02:32,020]: And then I eventually landed actually a job that I held for five years in IT consulting through completely I was an English major

[00:02:39,740]: This was not my destiny

[00:02:44,000]: And I decided I want to be a writer when I was eight years old

[00:02:46,820]: And then I just ended up doing network administration tasks for various companies

[00:02:55,039]: And I realized at some point that this was not my calling

[00:03:01,339]: It wasn’t that I was terrible at the job although I guess you’d really have to ask my clients whether I was terrible at the job

[00:03:07,360]: But I remember coming in one day and people have said you know how was your weekend

[00:03:13,339]: And I said oh man I saw this great band and it was amazing

[00:03:15,960]: It actually was not amazing

[00:03:17,580]: It was Gogo Bordello but they did do a very good live show

[00:03:22,300]: And everyone stared at me silently through this monologue

[00:03:27,119]: And then there was a pause and someone said I built a fiber channel network in my basement this weekend

[00:03:32,639]: And that was the moment when I realized these are not my people

[00:03:36,139]: I liked them but I didn’t care enough to do this on the side

[00:03:40,880]: And so I decided to go to business school the last refuge of everyone who doesn’t know what they want to be when they grow up

[00:03:48,559]: And while I was there I wrote the gossip column for my business school newspaper

[00:03:54,440]: It’s the only journalism I’d ever done

[00:03:56,240]: I didn’t write for the DP when I was at Penn

[00:03:58,240]: I never wanted to be a journalist

[00:04:00,080]: I wanted to be a fiction writer

[00:04:01,759]: I realized that that was impractical

[00:04:03,619]: So I was going to get a real job

[00:04:05,320]: Unfortunately the dot com bubble burst

[00:04:07,660]: The consulting firm that had hired me laid off the entire associate class without any of us ever starting

[00:04:14,699]: And 9 11 happened

[00:04:16,920]: And so I ended up down at ground zero where I started a blog

[00:04:21,459]: That took off way more than I expected

[00:04:24,739]: Eventually that job ended

[00:04:27,059]: And I was looking around

[00:04:29,000]: I spent a good two years from 2001 to 2003 trying to find a full time career

[00:04:35,540]: And eventually I met someone at a blogger meetup who worked for The Economist on the web team

[00:04:43,260]: And I said to her well if you guys ever have any jobs let me know because I love The Economist

[00:04:47,660]: And I don’t think I was very serious about that

[00:04:50,380]: I was blogging a lot but I didn’t think of myself as a journalist

[00:04:53,119]: But there was a job

[00:04:54,559]: I applied for it

[00:04:56,579]: I somehow got they had like 150 applicants and somehow they selected me

[00:05:02,179]: Terrible mistake

[00:05:03,820]: And then I ran my blog on the side

[00:05:06,679]: I edited for them

[00:05:08,160]: I wrote some pieces for them

[00:05:09,500]: And in 2007 I had I was still reeling from a terrible breakup

[00:05:17,160]: And two people in my office had been dating

[00:05:20,100]: And when he proposed she said no

[00:05:23,119]: And then the company sent her to Paris for four months

[00:05:26,799]: And so I went to my boss and was like can I go somewhere

[00:05:29,799]: He’s like you can’t go to Paris

[00:05:30,959]: He doesn’t work here

[00:05:32,339]: But if you can find somewhere we have an office and you can shift your own housing you can go somewhere else for a while to get out of the city

[00:05:40,000]: And so I the only place I could figure out to go was Washington D C which is where my sister lived

[00:05:46,220]: While I was there I went to a dinner for The Atlantic where there were a bunch of bloggers explaining how blogging worked to The Atlantic

[00:05:54,980]: Hilariously they made job offers to everyone at that dinner except Ezra Klein which was a terrible mistake

[00:06:06,700]: And I ended up blogging for The Atlantic for five years did a brief stint at Newsweek in its death throes then went to Bloomberg

[00:06:15,619]: And that’s where I really started columnizing rather than blogging

[00:06:18,739]: At The Atlantic I was really blogging

[00:06:20,040]: I was also writing long form pieces for the magazine on business

[00:06:25,019]: And then in 2018 The Washington Post hired me

[00:06:27,339]: And so that is my the capsule history of Megan McArdle

[00:06:30,299]: All right

[00:06:32,119]: So you’ve been at The Washington Post then for seven years by my count

[00:06:36,260]: I have yeah

[00:06:37,720]: It’s like forever in Megan years

[00:06:39,880]: Yeah

[00:06:40,859]: It’s like dog years or reverse or something

[00:06:44,119]: And I wasn’t like a terrible job hopper but around five years I would tend to start getting restless and feel like I had either figured out the place and figured out what to do for the audiences

[00:06:56,839]: And of course once you’ve done that it’s hard to keep doing something different

[00:07:00,040]: Or I would feel like I had figured out that what the place did I wasn’t good at

[00:07:06,959]: Or it wasn’t what I wanted to do

[00:07:09,299]: So I would sort of start hopping again

[00:07:12,140]: Yeah

[00:07:12,339]: So speaking of your audience The Washington Post and perhaps other newspapers now have this great AI function where they summarize all the comments

[00:07:21,959]: They summarize the vibe of the comments

[00:07:24,100]: Oh you’ve been reading my comments have you

[00:07:25,980]: I’ve been reading the vibe summary the vibe summary of the comments

[00:07:30,359]: And that’s something I would never normally have done on a column like yours three years ago but I wouldn’t have gone through all the comments and see what do randos on the internet think of Megan McCardle’s views

[00:07:42,700]: Because I usually get to the end of your columns and think yeah that’s a really that seems like a very right down the middle balls and strikes analysis of what’s going on here

[00:07:52,779]: But it looks like the vibe of the comments often there’s quite a bit of pushback

[00:07:59,179]: What do you think that like what does your Washington Post audience what do you feel they most get wrong about you your views

[00:08:12,179]: How do you what’s the pattern of their mistakes in your view

[00:08:18,299]: I think that it’s very nice of you to say that I try to call balls and strikes and I do but we all have our own biases right

[00:08:23,440]: This is humanity you never get beyond them

[00:08:26,739]: But I think that what my audience misses is that they will ask me why didn’t you write about this

[00:08:34,440]: Why didn’t you write about that

[00:08:35,440]: Why didn’t you write about this other thing

[00:08:37,000]: Because I do tend to write more about the errors of the left

[00:08:40,679]: It’s not the only thing I write about

[00:08:42,539]: I’ve certainly I just wrote a blistering column about RFK Jr

[00:08:47,520]: as well as some of the mistakes that were made during the pandemic which I think helped give us RFK Jr

[00:08:55,239]: But they will and I get this on Twitter too is that they will ask me how come you always write about this thing the left did that thing the left did and you didn’t write about all these terrible things that Trump did

[00:09:06,340]: And one answer to that and an answer that I actually try to fight against is that I just don’t have that much left to say about why I think Donald Trump is not a good president

[00:09:16,919]: I said it in 2015 when a lot of liberals still thought this was funny ha ha and were hoping he would win so that he’d be easy to beat oops

[00:09:26,880]: I said it in 2016 and 2017 and 2018 and I am running out of ways to say he is unstable he has a short attention span and his theory of domestic policy is often wrong

[00:09:41,700]: Not always there are areas where I agree with him but I think he is not an effective leader and he’s not an effective leader at delivering what his audience wants either and I think Doge is a good example of where I think there was a constituency there are multiple constituencies for reforming government and making it work better but he didn’t take a serious approach to that

[00:10:05,080]: Doge goes in for three months the government is a behemoth you’re not gonna get it in three months

[00:10:10,440]: But the thing is I’m part of a team at The Post

[00:10:15,440]: If I were writing my own blog I would write a lot more about my worries about Trump my disagreements with Trump et cetera

[00:10:22,340]: We have a lot of columnists who cover that territory really really well

[00:10:26,640]: They’re often saying what I would have said about why this or that is bad

[00:10:30,859]: What we don’t have as many of is columnists saying where the left is making mistakes and so that is where I tend to focus because we’re trying to do a broad page

[00:10:42,599]: We’re not trying to have you read the same column eight times with different bylines on it

[00:10:49,239]: And so people interpret this as like I have a secret agenda to I’m gonna pretend that I don’t like Trump but really I love Trump and this is all part of my secret agenda to advance Trump is that saying that I’m against Trump and then I’m actually just pointing out all the ways the left is wrong

[00:11:08,880]: And I understand to some extent why it looks that way to them but in fact the real answer is I have to do things that other people on my team are less likely to do

[00:11:20,900]: I’m not just building my personal brand

[00:11:23,140]: I’m working in an institution

[00:11:25,299]: And I think one of the reasons that people don’t think of that is that we are getting worse at working in institutions and thinking about institutions as institutions right

[00:11:36,580]: Media social media especially has turned a lot of people including you and me into our own brands

[00:11:45,739]: And I think that’s great

[00:11:47,559]: I probably spend too much time on Twitter honing my brand although that’s not really how I think of it

[00:11:52,140]: I spend too much time on Twitter because I like to argue

[00:11:55,760]: But that in fact there’s a lot of deep work that goes on in institutions and to be successful as an institution you have to think holistically

[00:12:07,260]: You are not a collection of 20 brands because then why do the brands need you right

[00:12:15,919]: It’s one thing when you owned a printing press and that was the only way for them to get their voice out

[00:12:19,840]: It’s not now

[00:12:20,580]: To work in an institution is to think about doing maybe some less of what you want to do and trying to do things that create collective value for everyone yourself included

[00:12:33,599]: And so that’s the thing where I understand why they have that perspective but I think it’s skewed

[00:12:39,179]: Yeah

[00:12:41,099]: So your Twitter handle is asymmetric info as in asymmetric information

[00:12:47,479]: Can you explain what that concept means and why did you choose that as your handle

[00:12:52,380]: Yeah so by the way I was told a social media guru once asked me did you deliberately choose a bad Twitter handle so you’d be hard to find

[00:13:02,280]: And I said no I thought that was very clever

[00:13:06,320]: So back in the when I was originally started blogging I was blogging as live from the WTC under a pseudonym which didn’t really take

[00:13:17,140]: It was Jane Galt

[00:13:17,960]: There’s a long story there

[00:13:18,960]: I’m not actually an objectivist but I thought it was a very funny pseudonym

[00:13:24,020]: And my oldest readers still know me as Jane Galt and we’ll talk you know so some of my readers now have

[00:13:29,960]: That’s a reference to John Galt

[00:13:30,719]: It’s a reference to John Galt

[00:13:32,299]: And the way this actually got started was that in 1996 I think or 1997 I was reading the comments at the New York Times

[00:13:39,559]: Unlike you I actually love comment sections

[00:13:41,520]: I go into my comment section which always shocks people

[00:13:44,520]: I spent a lot of time curating my comment section right up until the Washington Post where it just overwhelmed me

[00:13:49,200]: It’s too many people I can’t

[00:13:51,179]: But I had a pretty vibrant community which is now actually got a Facebook group

[00:13:55,280]: They named after Jane Galt

[00:13:57,820]: Although they’re really the community is not about me

[00:14:00,280]: It’s about them

[00:14:00,940]: They developed a lot of relationships between each other and it’s lovely to see that I made yeah I didn’t make it they made it but I created the space

[00:14:12,280]: I helped curate this group that’s still like they visit each other they chat

[00:14:17,400]: It’s really lovely

[00:14:20,299]: So when I left the World Trade Center I had to come up with a new name for my blog

[00:14:25,119]: And I should say I am terrible at writing headlines

[00:14:27,280]: I’m terrible at naming things

[00:14:28,580]: I’m bad at graphic design and all of that stuff

[00:14:32,039]: So I had a pretty I made a probably not a very good little header and I thought what am I going to put in it

[00:14:40,580]: And I said asymmetric information

[00:14:43,840]: And then I decided for some reason to call it asymmetrical information

[00:14:47,880]: So asymmetric information is the idea that when you are for example doing a transaction with someone

[00:14:55,159]: Insurance is a good example of this

[00:14:57,659]: When you are going to get insurance you know more about your health status for example if you’re going to get life insurance

[00:15:05,700]: You know more about whether you are say extremely depressed and contemplating suicide or whether you like to drive your car at 100 miles an hour or whether you have a heart condition than the insurance company does

[00:15:17,940]: And that’s asymmetric information

[00:15:19,760]: And when asymmetric information gets too bad it can prevent transactions from happening

[00:15:25,200]: And so for example this was something people worried a lot about with Obamacare and why they ended up using a mandate to force people to buy insurance

[00:15:32,580]: What they were worried about is that people would be like oh I can buy insurance anytime

[00:15:36,799]: I guess I won’t buy insurance until I need it

[00:15:39,239]: And when that happens you get what’s called a death spiral where the only people in the pool are super sick which drives the price of the insurance up because of course that price is related to the average cost of caring for people

[00:15:50,059]: And then the wellest people left in the pool drop out

[00:15:54,020]: And this actually happened in New York State when they before Obamacare because they had passed what’s called a community rating

[00:16:03,280]: You can’t underwrite these policies

[00:16:04,700]: You can’t exclude people

[00:16:05,760]: But they hadn’t added a mandate and the price just kept going up and up and up because the wellest people kept dropping out of the pool

[00:16:13,580]: So solving asymmetric information problems creating markets when they’re present and when they’re really distortionary is one of the big tasks of economics

[00:16:25,979]: And I thought it was just a fun and I feel like every you know all the time out there we’re dealing with asymmetric information problems

[00:16:36,080]: Journalists are you know we are constantly worried about what are the people telling me or is what they’re telling me true is what they’re telling me true but they have left out the crucial fact that I need to know to understand that

[00:16:49,000]: It’s everywhere

[00:16:50,580]: And so that was the genesis of it

[00:16:53,739]: And then when I went to Twitter which is I think like 2008 I don’t even remember it’s been so long I just took my by then I was at the Atlantic I was blogging and I just took my blog name which eventually by the way was like stripped off of my blog and it was just my blog

[00:17:12,920]: It was just Megan McArdle at the Atlantic but I took that and I made it my Twitter handle

[00:17:17,000]: And that is the long ago origins of my name

[00:17:23,939]: Okay so speaking of health insurance I want to talk a little bit about Luigi Mangione the heartthrob healthcare CEO murderer

[00:17:37,280]: So to I mean you’ve been covering healthcare for a long time

[00:17:40,819]: You wrote a lot about Obamacare

[00:17:43,060]: I think this has been from my perception kind of one of the top five areas of focus for you over the years I would say

[00:17:54,359]: So I think it goes without saying and I’m gonna having read you a long time I’m gonna predict you are anti murder

[00:18:01,579]: So we don’t have to litigate that

[00:18:04,540]: Don’t try to like put me in a box

[00:18:06,959]: I contain multitudes

[00:18:08,520]: Yes I’m anti murder

[00:18:10,280]: But if we were to just take Luigi Mangione’s diagnosis such as it is of the problems with the American healthcare system as explained in his manifesto his choice to target an insurance CEO as opposed to say a big pharma CEO or the leader of a network of hospitals or any number of potential targets he quite specifically chose a health insurance CEO and diagnosed the deepest problem with the cost of healthcare as lying with the bean counting insurance companies that are okay denying you coverage and making the whole system confusing enough that people can really get hit with surprising bills in difficult situations and so forth

[00:19:05,719]: To what extent is that diagnosis analytically correct

[00:19:15,160]: Well let me say this right

[00:19:17,300]: There’s another concept in economics called principal agent problems which is where you do not have the same incentives as people working for you

[00:19:28,219]: And what you have in the healthcare system because the way we pay for it is so weird and diffuse is that you don’t have the same incentives as the other people in the system right

[00:19:39,780]: For me I want unlimited care the best possible

[00:19:45,640]: I want to have the best possible doctor

[00:19:47,199]: I want to have any possible treatment that could conceivably help me

[00:19:52,420]: That’s not no one in you know Warren Buffett can afford to have that but that’s not practical

[00:19:57,540]: You can’t have everyone getting every single thing they might conceivably want

[00:20:02,280]: And that is true to be clear in every system in the world

[00:20:07,000]: In fact the United States is bad at rationing

[00:20:09,459]: It’s not good at rationing

[00:20:10,540]: We are not good at denying people treatments

[00:20:13,060]: We probably over treat people and I don’t mean this in terms of cost

[00:20:16,699]: We’re a rich country

[00:20:17,560]: We can basically afford what we’re doing

[00:20:21,319]: But you know For example end of life is something somewhere where you see this a lot and I think the statistic that we spend too much at the end of life is a little you know people will say well like 25 or 50 of spending is in the last year of life and that’s true but you don’t know it’s the last year of life

[00:20:41,061]: Like my mother went from being I mean she had COPD so she was not healthy but there was no reason to think that she was going to die imminently and six weeks later she was dead after a series of incredibly expensive hospital stays and that’s just something that it happens

[00:20:57,281]: People destabilize you spend a tremendous amount of money but we do a lot of treatments that are probably marginal that it’s not clear that there’s a huge benefit to them

[00:21:06,601]: And of course there’s a cost and I think you see this with cancer patients where doctors actually will often choose different things than what most patients do because doctors don’t want to tell people like it’s time to give up they’re reluctant to say that and the doctors when they get cancer it’s not that they don’t keep trying they do but they recognize earlier that it’s time to stop

[00:21:29,781]: Because not because they’re trying to save the system money because they see that at some margin doing these highly speculative I don’t not this probably won’t work but why not give it a go for 500 000 or whatever is I mean these treatments have side effects and they’re quite severe and they just decide not to pursue every last possibility but that’s we want the option to pursue every last possibility

[00:21:55,401]: No system can provide that

[00:21:57,781]: And the insurer in America is the place that makes that clear that you can’t you don’t it’s not just between you and your doctor because you’re not paying for it

[00:22:07,621]: If you’re paying for it in plastic surgery it’s also actually not just between you and your doctor

[00:22:13,701]: Plastic surgeons worry a lot about the problem of people who have like body dysmorphic disorder and will just keep getting surgeries that in the judgment of everyone who is not that person are not helping them and are in fact harming them because they will never get over their belief that they’re not right and they will keep they will destroy their body trying to fiddle with it

[00:22:32,301]: But in it is more true in plastic surgery that it’s between you and your doctor you’re playing you’re paying for it

[00:22:38,241]: And so you can decide whether you want to have enormous breasts or liposuction or whatever and it’s no one else’s business but when someone else is paying for it it’s their business

[00:22:47,741]: And what I think is you know what you see in the healthcare system interestingly is everyone everyone everyone who’s not in the who’s not a wonk who doesn’t study this issue they’re all convinced that the two evil actors in this are the insurance companies and pharma

[00:23:05,461]: And in fact pharma is producing is probably saving us money on that

[00:23:11,161]: They are saving us from having surgeries which are usually worse than drugs

[00:23:14,861]: Not always

[00:23:15,781]: Sometimes you decide it’s like get it like take care of the problem for good but drugs can prevent a lot of things

[00:23:22,601]: They prevent strokes they prevent cardiovascular right

[00:23:24,881]: It’s much cheaper and better for everyone if you take your hypertension meds rather than having a stroke

[00:23:31,061]: But unlike a lot of other things in the system that’s the place where consumers are most consistently reminded that they are paying for this right

[00:23:52,481]: And in fact that’s not what’s driving the cost

[00:23:55,801]: The insurers are holding down the cost right

[00:23:58,161]: That’s that’s their job in the system

[00:23:59,901]: And in fact I think a lot of people aren’t aware they’re medical they’re regulated they can’t when they deny claims that’s saving money for the claims holders they can’t make more profit that way

[00:24:09,541]: In fact they make less profit because they have something called a medical loss ratio really you know wonky term

[00:24:15,621]: Basically what it means is they have to pay out a certain amount of their premiums and claims

[00:24:21,781]: And it is somewhere between 80 and 85 depending on the size of the company and the size of the insurer

[00:24:29,341]: But basically and then the rest goes to overhead and then there’s actually a quite small profit

[00:24:33,701]: They’re not it’s not a very profitable industry

[00:24:36,081]: And so they like they don’t make money by denying you claims what they do is hold premiums down by denying you claims

[00:24:42,081]: And the that has to happen

[00:24:46,801]: Britain does it you know every system does it

[00:24:49,261]: But what’s different is that in the US it’s transparent

[00:24:51,941]: And this is because our system is so fragmented

[00:24:54,421]: There are so many different insurers they each negotiate different deals with doctors and hospitals and all the rest of it

[00:24:59,941]: So your doctor when he’s talking about treatment with you does not know whether it’s covered or not

[00:25:33,081]: Right

[00:25:35,521]: So I think where the error is though is that in fact that fragmented system doesn’t ration care more than other systems

[00:25:43,061]: Russians care less than other systems for a whole bunch of reasons including the fact that rate the incentives of the government and the and you know the national system or the insurer because some of them have private insurers who are working in a highly government regulated market

[00:25:58,361]: The incentive in those cases is aligned

[00:26:02,201]: They all have decided we’re going to cover this and we’re not going to cover that

[00:26:06,461]: And that’s the schedule

[00:26:08,921]: But in the US you have the government’s incentive is just to tell insurers they should cover everything right

[00:26:15,901]: Any individual legislator their incentive is to just and this is why in fact HMOs became a bad name because they were doing a really good job of cost control and they held down cost growth in the late 90s

[00:26:28,361]: And then legislators made them stop and then cost shot up which is part of the reason Obamacare was so popular because costs were really rising very quickly after legislators made it harder and harder to deny treatments

[00:26:43,061]: And then of course patients want treatments employers have cross cutting incentives where on the one hand they want to hold their premiums down

[00:26:51,381]: And on the other hand when Bob in shipping his kid has a rare pediatric cancer no one wants Bob in shipping’s kid not to get the best possible coverage right

[00:27:00,981]: And so there’s all of these cross cutting incentives that actually make it much harder for us to control costs in other countries

[00:27:07,041]: Although I will say this we often think about ourselves as having a cost growth problem

[00:27:11,721]: And we actually don’t anymore

[00:27:13,341]: 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

[00:27:24,751]: And now costs have roughly stabilized as a percentage of GDP with the exception of the pandemic

[00:27:31,061]: It’s a big exception but that was an emergency event

[00:27:33,341]: That’s not normal

[00:27:34,941]: Our costs aren’t growing faster than they are in other countries

[00:27:38,201]: We just started from a higher level

[00:27:40,121]: We let our costs get out of control like 30 years ago and we never brought the level of cost down

[00:27:46,741]: But we did stabilize the growth and we’re now at a level that’s like basically sustainable

[00:27:53,181]: But it is still going to be frustrating to people that their doctor will say well let’s do this

[00:27:58,461]: And then the insurer will come back and say oh no you have to have the generic instead of the brand name

[00:28:04,101]: You have to have this procedure instead of that

[00:28:06,481]: We’re going to make you wait or get pre authorization or all the rest of it

[00:28:10,561]: Try physical therapy before you get surgery

[00:28:12,241]: That frustrates people and they blame the insurers for just being greedy and hurting people for profit

[00:28:22,241]: The insurers are not doing anything appreciably different from what a government insurer does elsewhere

[00:28:27,401]: The government makes exactly the same kinds of calculations and they’re in general less generous than what American insurers do

[00:28:35,461]: I remember looking at the actual profit margins which it’s you know you can Google 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 like a like a you know so I read the manifesto and I thought this guy understands absolutely nothing about the healthcare system

[00:29:04,461]: Not that I’m any kind of expert but I already understand it better than he does

[00:29:10,181]: 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

[00:29:25,221]: 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

[00:29:33,321]: 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

[00:29:41,721]: There’s obviously Medicare and Medicaid

[00:29:43,341]: It’s also employer sponsored health insurance which exists only because during World War Two when there were price controls companies were looking for ways to lure workers to work for them

[00:29:58,501]: And health insurance was exempt from the price controls

[00:30:01,781]: And so they would they they used this as a fringe benefit to attract people to their companies

[00:30:08,961]: And then they made that more and more generous because there are tax benefits for doing this is all an entirely we’ve also got the VA we’ve got the Indian health system right

[00:30:18,101]: We’ve got we’ve got TRICARE and the military

[00:30:21,521]: We’ve got all of these different systems

[00:30:24,741]: And people it’s not seamless moving between them right

[00:30:29,101]: People get into the cracks

[00:30:30,601]: We’ve then got all of these individual insurers who are who are competing less than in other countries on a really tight regulated package

[00:30:40,281]: But we also have a legacy of bad policy design in other ways

[00:30:46,941]: So for example doctors and nurses and basically everyone in the US health care system makes more than their counterparts abroad

[00:30:55,661]: And that’s because when Medicaid was when Medicare was started in the 1960s they didn’t have a fee schedule

[00:31:06,881]: They just said usual and customary fees

[00:31:10,141]: Well doctors had a lot of latitude to decide what was usual and customary and so their income started going up really fast

[00:31:17,561]: And then you know being a doctor in 1950 was not a really good high income job in the way that it is now

[00:31:27,441]: That’s a legacy of Medicare

[00:31:30,001]: And then they started trying to control this

[00:31:34,441]: And then the costs moved to hospitals

[00:31:38,181]: 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

[00:31:48,841]: 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

[00:31:58,501]: And we’ve been trying for decades to get a handle on that

[00:32:02,461]: And we’re never going to I think is the actual answer is that we’re just going to kind of muddle along

[00:32:08,421]: Because the other piece of this is we’ve done a lot of bad policy design

[00:32:11,641]: But each piece of bad policy design has created stakeholders

[00:32:16,161]: And those stakeholders will freak out if you try to take anything from them

[00:32:19,601]: So obviously the doctors and nurses they are extremely effective at not having their incomes cut

[00:32:26,121]: But you also have you know about what Obamacare wanted to do really wanted to do with the kind of vision was if they had had their druthers

[00:32:34,001]: If you talk to people you can it would it was basically with some details changed

[00:32:42,001]: 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

[00:32:57,041]: And but they’re they’re basically all offering the same stuff

[00:33:01,041]: And that everyone is in that and the government just picks up the tab for the really poor people

[00:33:07,481]: And 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

[00:33:18,721]: And they couldn’t do it

[00:33:20,001]: And why couldn’t they do it

[00:33:21,181]: Because people who had employer sponsored health insurance went ballistic

[00:33:28,521]: People say like Medicare for all pulls well

[00:33:32,861]: And this is kind of true and kind of not

[00:33:36,041]: The trick there is that you pull pieces of it

[00:33:38,921]: Should people be able to buy into Medicare at any age that will pull well

[00:33:44,661]: 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

[00:33:56,381]: And similarly people are fine with add ons what they’re not fine with is losing what they have

[00:34:01,861]: And what they have actually is part of the problem in some ways

[00:34:06,561]: Now look I support a really free market system

[00:34:09,621]: My preferred health care plan would be that there is no health care plan for under say 10 to 15 of adjusted gross income is that you sort that out yourself

[00:34:20,181]: If you have catastrophic expenses the government will pick up the tab 100 of the tab above whatever that level is

[00:34:29,981]: But almost anything would be better than the mess we have

[00:34:34,001]: But you can’t touch it

[00:34:35,101]: It’s like look I love Kaiser I don’t want to I don’t want to leave Kaiser I don’t want you to like take it and give me Medicare for all and promise me it’s gonna be just as good as Kaiser because I don’t believe you

[00:34:43,401]: And you probably shouldn’t I’m probably correct not to believe them

[00:34:48,121]: And so that 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

[00:34:59,661]: And what people end up doing what reformers end up doing is that they then like okay well here’s this gap I’m going to add I’m going to do another program here

[00:35:09,261]: And that just makes everything worse right

[00:35:11,301]: Because now you’ve got yet another complicated piece of the system in which there’s new gaps that that arise in between there’s complex interactions

[00:35:21,621]: And 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

[00:35:30,961]: 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

[00:35:40,181]: And it’s amazing

[00:35:41,461]: Right

[00:35:41,741]: And that’s that’s not real right

[00:35:43,461]: So that is not a real system that exists anywhere

[00:35:45,281]: So 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 1 000 in an ambulance bill

[00:36:02,861]: And you know this is a this is the kind of thing that can and does happen in America

[00:36:08,241]: In America you can pay with money in the extreme case

[00:36:11,801]: In the UK you can pay with time in the extreme case

[00:36:16,341]: 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

[00:36:27,441]: Right

[00:36:28,401]: So is it the case that you know Bernie Sanders and probably Luigi Mangione they want a single payer system

[00:36:36,941]: 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

[00:36:47,761]: Um yes with a qualification right

[00:36:50,701]: Single payer there’s lots of different kinds of national health insurance programs single payer which is what they have in Canada

[00:36:57,381]: And like there’s 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 there’s no real private competition

[00:37:10,561]: You have systems like Australia where there’s a basic government system

[00:37:14,741]: And then so there’s lots of different ways to do this

[00:37:17,381]: 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

[00:37:28,741]: 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 is you stand in line

[00:37:39,841]: And if you look at the wait times in the UK and Canada they are to my eye insane

[00:37:47,381]: 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

[00:37:59,601]: 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

[00:38:07,201]: And waiting a year or two for that surgery is a big deal

[00:38:11,081]: Waiting a year or two for your imaging

[00:38:14,081]: I mean like when people tell me the wait times I’m sometimes shocked when they’ll say yeah I waited six months

[00:38:21,061]: Now this is not true for everything right

[00:38:23,201]: 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

[00:38:34,821]: But more and more stuff goes to the back of the line

[00:38:37,741]: And why does that happen

[00:38:39,281]: 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

[00:38:48,581]: When something’s growing on the government budget they can just say no right

[00:38:53,441]: And when they’re under budget pressure that’s what they do

[00:38:56,241]: And so the reason that the UK and Canada have the biggest problems is just that they spend less money on their system

[00:39:04,661]: They’re not I mean Ireland spends very little on their system and should really spend more

[00:39:10,461]: But that you know they could provide faster service if they spent more money and built more capacity

[00:39:17,801]: But the political incentive is not to build capacity because it’s so expensive to maintain

[00:39:26,441]: 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

[00:39:39,881]: And wait times just tend to go to the back of the queue

[00:39:42,881]: 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

[00:39:49,741]: Because if you think about the there’s a lot of things in life where the last mile gets progressively more expensive right

[00:40:00,561]: So if you’re building a fiber optic network laying the big trunk lines is not cheap by any means

[00:40:09,681]: But it’s actually even more expensive just to get the last mile to someone’s house

[00:40:14,461]: And the reason is there’s a bunch of houses in between you and them and it’s going to take a lot

[00:40:18,421]: Whereas they tend to lay big lines out where there’s no people

[00:40:23,701]: And so there’s just a lot of things that have that characteristic getting the last bit of reliability in a system

[00:40:31,121]: And the way I think about this is when 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

[00:40:39,666]: and just take Ubers and Lyfts because they were being subsidized by people who were trying to gain market share

[00:40:45,386]: I never hear that now that the prices have gotten semi rational

[00:40:49,946]: 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 and then the battery dies

[00:41:07,806]: But we have two very large dogs

[00:41:11,606]: 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

[00:41:17,946]: And so that one use case I’m maintaining a car for basically that use case and a couple of other small use cases

[00:41:25,606]: And otherwise I could totally get around on public transit my feet etc

[00:41:29,066]: But I have to have this extremely expensive vehicle that I then have to pay thousands of dollars to insure and find a parking space for and gas and maintenance for this one use case

[00:41:40,126]: That use case is more expensive than every other thing I do

[00:41:44,006]: Those last things are often just very hard to get

[00:41:48,246]: 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

[00:42:01,186]: 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

[00:42:14,186]: Your first 80 is just like people get colds they do this they do that

[00:42:18,406]: It’s a pretty predictable flow and then you can just keep

[00:42:22,446]: But that last 5 to 10 is the stuff that doesn’t happen actuarially in neat little bundles

[00:42:32,406]: It happens

[00:42:34,646]: You go to the ER you have the ambulance demand and suddenly you’ve got 87 cases

[00:42:38,966]: Well if I had all of these spare ambulances I could just send it

[00:42:41,626]: But having all those spare ambulances means I have a lot of ambulances with crews that are often underutilized

[00:42:48,206]: America makes the decision to just have a lot of redundancy

[00:42:51,966]: Other systems because the government’s paying for it very directly and is making a budget not like the US where the government’s paying for it and they find out what they’re paying after

[00:43:03,646]: The government is paying for it in advance making a budget deciding how many ambulances to buy

[00:43:08,986]: And they decide to buy fewer than would cover all of that peak capacity

[00:43:14,226]: And that’s how those calculations

[00:43:17,906]: That was a little shaggy dog

[00:43:20,166]: Apologies

[00:43:21,106]: That’s okay

[00:43:22,366]: Let’s pivot topics now and talk about artificial intelligence the large language model revolution that’s happening now

[00:43:31,726]: You’ve written quite a bit about this

[00:43:34,186]: And so there’s a few different buckets we could talk about

[00:43:37,386]: The first I want to touch is like what’s going on with education and AI in the classroom

[00:43:43,486]: How has chat GPT been sort of affecting how teachers teach how students learn and so forth

[00:43:52,526]: So to give you my window on this I recently taught for three months at the University of Austin a small seminar course just this past January through March

[00:44:04,586]: And my view as someone who loves chat GPT personally I use it every day

[00:44:11,786]: 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

[00:44:19,686]: So it’s like a massively fun tool for people like me

[00:44:25,066]: And so I have like a generally positive outlook on it

[00:44:29,026]: I think people should use it

[00:44:31,306]: I think students should use it

[00:44:32,786]: And regardless of whether they should they’re going to

[00:44:36,666]: I’m also

[00:44:37,766]: I don’t believe that professors in general can detect when their students have used it

[00:44:45,186]: I’m not saying you can never detect it

[00:44:47,806]: Obviously many students have been caught using it on essays and so forth

[00:44:51,686]: But I think in general students have a greater incentive to cheat really well than professors have to catch them

[00:45:00,786]: And obviously the dirty little secret is that similar to how you were saying doctors don’t like to tell their patients that there’s nothing more I can do for you

[00:45:10,006]: You’re a few months from dying

[00:45:12,546]: Professors don’t want to catch their students cheating because it’s a hassle for us

[00:45:18,106]: Like it’s you know God forbid

[00:45:20,506]: If I’m right about your cheating that creates paperwork and meetings for me

[00:45:26,466]: Like it literally creates extra work that I’m not getting paid for

[00:45:30,586]: You know and even let’s say I have a really good bullshit detector I’m right 95 of the time

[00:45:35,826]: Well 1 in 20 I’m going to accuse a student of cheating who in fact did not cheat

[00:45:40,386]: That creates a much much bigger headache for me because now I’ve actually harmed the student in that

[00:45:46,606]: Now I could potentially be in trouble right

[00:45:50,606]: So professors have almost no incentive to catch no personal incentive to catch cheating

[00:45:56,166]: Students individually have a pretty strong incentive to cheat successfully because your grade goes up

[00:46:01,566]: So I just yeah I don’t

[00:46:03,546]: From an incentives point of view I don’t buy that professors as a class are going to win the war on cheating

[00:46:11,866]: And so my approach as a professor was to say fine to use GPT as your personal research thinking buddy tutor at home

[00:46:21,766]: There are going to be no assessments that you could even possibly cheat on in this class

[00:46:27,406]: It’s all going to be blue books

[00:46:29,426]: And if I could do it again I would also go for the oral exam for variety

[00:46:33,886]: Like maybe yeah in class quizzes class participation blue books

[00:46:39,466]: It’s a cheat proof class

[00:46:40,626]: And that’s to me that’s got to be the future of teaching

[00:46:45,506]: So what do you make of this

[00:46:48,546]: Yeah I have

[00:46:51,546]: So I taught a class

[00:46:52,726]: I’ve taught a few journalism classes

[00:46:56,106]: And the thing I always taught them pass fail

[00:46:58,146]: And the thing I always said was I’m not going to try to catch you cheating

[00:47:01,686]: You are adults

[00:47:03,346]: You have a choice

[00:47:04,366]: 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

[00:47:13,506]: If you want that this is pass fail

[00:47:15,586]: If you show up and you hand in something you will pass but you can get

[00:47:21,306]: I’m not going to try to make you get for your 250 000 the value that’s here

[00:47:27,806]: Whatever that is maybe there’s no value

[00:47:29,386]: Maybe they should just come and sleep through my class

[00:47:32,046]: But I can’t

[00:47:33,906]: You know you’re now

[00:47:35,466]: You guys are in your 20s

[00:47:36,586]: You were old enough to go get killed in a war

[00:47:39,606]: And you were old enough to

[00:47:41,506]: Like not old enough to drink because somewhat irrational policy laws

[00:47:46,786]: But like my job is not to make you get something out of this

[00:47:51,886]: My job is to make it possible for you to get something out of this

[00:47:55,666]: But that was easier for me to do because it was pass fail right

[00:48:01,386]: I think it’s a really

[00:48:02,746]: I think you’re right that the incentives are bad

[00:48:04,666]: But I also think there’s a bigger question which is that

[00:48:11,146]: You know what I was saying is this is a pass fail class

[00:48:13,886]: It’s like a half credit

[00:48:15,066]: It’s not the difference between you going to grad school and not

[00:48:18,926]: So you should enjoy this class

[00:48:20,466]: You should only be here if you want to be here and learn something

[00:48:22,806]: And you should get what you can out of it

[00:48:24,886]: Be in the discussion

[00:48:25,866]: Do this right

[00:48:26,566]: Like don’t sit here and passively just try to you know put the time in and grub the grade

[00:48:34,686]: Get what you can

[00:48:36,046]: But that’s not really realistically how most people think about an education

[00:48:42,846]: Right

[00:48:44,266]: 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 like 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

[00:49:06,986]: And 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

[00:49:28,846]: Not that much in my case

[00:49:30,846]: I was a really terrible student

[00:49:35,726]: But that signals that to employers

[00:49:38,146]: 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

[00:49:58,986]: But 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

[00:50:15,366]: You are signaling a lot of stuff about yourself

[00:50:18,766]: This is Brian Kaplan’s argument in the case against education

[00:50:20,906]: Yes this is Brian Kaplan’s argument which I basically buy

[00:50:23,526]: Yeah thank you

[00:50:26,666]: But one of the things that it signals 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

[00:50:50,706]: They may not be the most glistening beautiful prose in the entire world but they can be read and comprehended

[00:50:58,066]: And the value of that skill is itself declining right

[00:51:06,026]: AI 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

[00:51:20,386]: And colleges I think have not grappled with that more fundamental problem

[00:51:24,766]: I mean first of all they haven’t grappled with the cheating problem

[00:51:26,646]: You’re right

[00:51:27,546]: When I talk to academics they’re all just kind of you know home baking solutions or not

[00:51:33,686]: Some of them have just given up

[00:51:35,266]: 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

[00:51:42,706]: That’s like Mike Munger at Duke is doing this

[00:51:46,406]: But a lot of them are terrified

[00:51:49,186]: 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 society loaded all of these different tasks onto universities

[00:52:01,686]: It loaded dating finding a spouse or at least finding some friends who can find you your spouse later

[00:52:11,426]: That’s less relevant to your generation than to mine

[00:52:14,946]: But to my generation that was a big thing

[00:52:17,026]: You would meet your boyfriend in college and hopefully you would marry that person right

[00:52:21,906]: Not in the way that my mother’s generation would joke about getting her MRS degree

[00:52:26,606]: We were also there to get a job and hopefully maybe even learn something and become more interesting people

[00:52:33,166]: But dating was a big feature of it

[00:52:35,466]: We loaded finding sorting people for employers onto it

[00:52:39,606]: 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

[00:52:53,846]: They’re also really not selected for their ability to use AI well right

[00:53:00,566]: 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

[00:53:15,176]: That would not be your first choice right

[00:53:17,756]: Who have spent their entire lives in archives reading and writing critical theory about their period or political scientists or go through the disciplines

[00:53:28,396]: And to be clear I’m not denigrating what they do

[00:53:30,896]: I’m saying that if you were starting from scratch you definitely would not be like yes what we need is a bunch of PhDs in sociology to handle this task

[00:53:43,036]: And that thing and also so the university is really unoptimized to handle the problems of preparing the workforce for AI

[00:53:53,976]: 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 but that didn’t matter because the value of the signal was itself valuable right

[00:54:08,456]: 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

[00:54:16,456]: So hold on can I stop you right there

[00:54:18,816]: So presumably though today it’s still signaling a mix of IQ conscientiousness maybe to some degree competency at a certain upper class way of holding yourself

[00:54:37,856]: 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

[00:54:54,336]: Sort of

[00:54:55,476]: 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

[00:55:03,396]: High school kids are using it too

[00:55:06,476]: 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 but if you went to employers and were like do you want a bunch of people who are smart enough to cheat without getting caught

[00:55:22,316]: They might have questions about for example should you put those people in your accounting department

[00:55:30,776]: And so I think I do think it’s a problem like the value of the signal is going to degrade so strongly and the real value of the skill is also going to degrade

[00:55:45,276]: And the even bigger problem now that I’ve copied is that the governance of universities is completely ill equipped to handle any of this

[00:55:56,296]: Everything’s run by committee

[00:55:57,396]: No one’s in charge

[00:56:00,736]: And so no one has the ability even they’re all forming committees on how to do AI

[00:56:06,356]: 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

[00:56:15,236]: No 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

[00:56:24,076]: And the problem is a university is a bundle

[00:56:28,296]: So you can be Mike Munger and teaching a great AI optimized class

[00:56:33,076]: I’ve never been in his class

[00:56:34,136]: I just I know him personally and I assume he’s doing a pretty good job

[00:56:38,696]: But if your colleagues all over the university are bad at this and they are not 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

[00:56:55,756]: Right

[00:56:56,276]: 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

[00:57:18,436]: I think universities have a massive problem that they have not really begun to confront

[00:57:25,096]: There are exceptions to this

[00:57:26,656]: People like Hollis Robbins are writing great stuff about this but in general I don’t think they understand the scope of their problem

[00:57:34,496]: I think they understand the narrow problem of like we have a lot of kids who are cheating and we got to fix that

[00:57:39,596]: But you can fix it within class assessments right

[00:57:42,516]: You can fix it that way

[00:57:45,256]: 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

[00:58:11,356]: How do you take that institution and transform it into an institution that is preparing kids for the AI job market

[00:58:21,096]: So if you were the czar and had complete control over a top college what changes might you institute

[00:58:31,356]: Well 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

[00:58:43,596]: I think there are like really deep problems with the way that academia hires and promotes

[00:58:52,456]: And I understand the institutional reasons that happened but this obsessive focus on who has the best publication record right

[00:59:00,836]: That is what has status in academia

[00:59:03,596]: Not teaching well not any of the rest of it

[00:59:05,436]: It’s nice if you teach well

[00:59:07,036]: They’re happy if you teach well

[00:59:08,516]: They admire people who teach well but that’s not what has status in their world

[00:59:16,056]: And 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

[00:59:33,936]: I don’t know the answer to that

[00:59:37,236]: But I just think this obsessive focus and then like the humanities got science envy and they all started doing it too

[00:59:42,656]: 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

[00:59:52,696]: We have all of the books

[00:59:56,056]: I’m gonna get really mad English professors writing to me and say I’m not saying that it’s not valuable exactly but that 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 wanna bring people into the classroom you don’t need to communicate your incredibly narrow theory

[01:00:20,536]: What you need to communicate is a love of books

[01:00:23,136]: I just finished Middlemarch which I guess has gone a little viral on Substack and everyone’s reading Middlemarch and I was also reading it but not because I saw it on Substack

[01:00:35,296]: It was so good

[01:00:37,028]: I was sad

[01:00:40,028]: I was both elated that I had read it and I was sad because it was over

[01:00:45,188]: I don’t have that many that experience with that many books these days

[01:00:50,908]: But that’s to me this is the purpose of bringing people into English classes right

[01:00:58,588]: It’s not to teach them to identify the socially problematic behavior in people who have been dead for 300 years

[01:01:09,068]: It’s to teach it’s to show them the human experience to open their minds to another time to let them see right

[01:01:16,528]: And that’s not what they’re selecting on

[01:01:19,548]: They’re selecting on have you written articles that five other professors have read in your subfield

[01:01:25,468]: And that I would just stop that

[01:01:28,908]: I would stop the 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

[01:01:38,528]: I wouldn’t say research is bad but I would just say it’s okay if you don’t publish

[01:01:43,808]: 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

[01:01:56,248]: And I eventually asked someone why that happened

[01:01:59,188]: He ended up at the University of the Pacific

[01:02:02,588]: And I went back to Penn and I was doing a thing at the writer’s house and I asked why that happened

[01:02:07,668]: And they said well his research wasn’t really that impressive

[01:02:09,868]: And I was like okay

[01:02:12,488]: But he had all of these students who came out of his classes on fire for Mark Twain

[01:02:19,558]: And he also and he said well students aren’t necessarily a good judge of what they learned

[01:02:23,668]: And I was like I remember what he taught me

[01:02:26,408]: And let me tell you the people who are more prestigious and you’re a little I don’t remember anything

[01:02:31,248]: I remember

[01:02:32,608]: Yeah

[01:02:33,128]: 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 lifelong passion for the topic

[01:02:47,708]: And so 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 dreck

[01:02:59,708]: We didn’t just read Connecticut Yankee and Huckleberry Finn and Tom right

[01:03:04,608]: We read putting head Wilson

[01:03:07,008]: We read the stuff he was churning out by the fistful when he was on election

[01:03:12,088]: He went bankrupt in the 1890s and then spent the next five years on a constantly obsessively writing and lecturing trying to make enough money back make enough money to pay his creditors

[01:03:22,888]: We read all that

[01:03:24,088]: And that was actually really good because I think the way we usually teach literature is the kind of the highlight reel right

[01:03:30,988]: You see all of the you see the best thing every writer wrote

[01:03:36,508]: And instead of that we just saw everything this writer wrote

[01:03:40,368]: And it was a very different experience

[01:03:42,728]: And and just it was just a wonderful class

[01:03:45,418]: I remember it 30 years now 35 years

[01:03:48,598]: When did I graduate from college

[01:03:49,658]: No not quite 35 years

[01:03:51,918]: But 30 years on I still remember it

[01:03:54,238]: And that’s what you can give to students

[01:03:56,218]: They’re not going to come away

[01:03:58,838]: You know they’re like I had three professors there who were really memorable and none of them was the famous prestigious research

[01:04:08,158]: I took you know a class with Elisa New who ended up at Harvard

[01:04:15,318]: 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

[01:04:28,618]: And I juxtaposed them and I in sort of typical precocious I guess I was a sophomore or a junior

[01:04:36,878]: I was a junior

[01:04:38,558]: I decided that the interesting thing about these books was that actually like the incidents of a life of a slave girl which is a real slave narrative is in some ways kinder to her owners than Uncle Tom’s cabin

[01:04:52,738]: And that really struck me

[01:04:54,298]: And so I wrote I wrote about Stockholm syndrome and this thing where you identify with your captors

[01:05:03,958]: Only I didn’t define it

[01:05:07,058]: I never I was just like such an arrogant little like obviously everyone knows what Stockholm syndrome was

[01:05:12,238]: So I got it back from the TA and I got it back like a month late because I turned it in a month late

[01:05:21,998]: And he said this was one of the best papers I’ve ever read

[01:05:26,658]: I didn’t understand all that stuff about Helsinki but really well done

[01:05:29,678]: And I couldn’t remember what I’d written about

[01:05:32,198]: I was like Helsinki when did where did Helsinki come into my yeah

[01:05:36,578]: And then I read it and I realized that I hadn’t defined it

[01:05:38,838]: He had no idea what I was talking about

[01:05:40,258]: And this is pre Google which is important for you youngins to understand is there was no way to like just go and find out what a term you didn’t understand meant

[01:05:49,618]: And that he could not possibly have understood this paper

[01:05:53,578]: 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

[01:06:02,378]: That’s the only thing I remember from her class

[01:06:04,338]: I don’t remember anything she said

[01:06:06,438]: I’m sure she’s a fine teacher and that many people have had very I don’t want to sort of slam but it’s just that like that thing would to me for undergraduate English which is not going to produce a large number of researchers and professors right

[01:06:23,078]: 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

[01:06:36,898]: In which case what the hell are you doing

[01:06:39,338]: Yeah

[01:06:39,878]: I think you know 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

[01:06:55,738]: But I think of it a little bit like being the warm up comedian in a comedy show

[01:07:03,138]: Because if you go down to the comedy cellar they don’t just hit you with the first comic of the night

[01:07:08,458]: There’s a warm up comic that it’s really essential to the show because it sets a tone and sets a mood

[01:07:15,858]: And only once that mood is set is the audience really capable of letting go of themselves and fully surrendering to the comic like mode of being

[01:07:28,918]: 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 generally hard to snap into

[01:07:52,838]: And once that mood is set you almost don’t have to do that much because suddenly they’re so into the topic for 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

[01:08:13,198]: That’s really what it’s about

[01:08:14,458]: But I think for a lot of professors you know 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

[01:08:31,618]: 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

[01:08:43,038]: Yeah I mean I think that’s a really big piece of it

[01:08:45,758]: But it’s also just that teaching is hard not everyone’s good at it right

[01:08:50,558]: There’s lots of things I’m never going to be good at

[01:08:54,118]: 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

[01:09:07,558]: And there’s some overlap obviously being familiar with your matters et cetera right

[01:09:13,218]: But I think there’s a threshold beyond which and I think this varies by discipline again

[01:09:19,038]: 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

[01:09:28,398]: Like students probably benefit more from being around that

[01:09:33,038]: But I don’t think there’s a cutting edge of Renaissance literature right

[01:09:38,718]: We have the canon

[01:09:41,078]: And lots of people can have really interesting thoughts that will engage kids

[01:09:45,418]: But it’s also just true that the best person to teach something is not always the person who’s best at it

[01:09:52,798]: Oh yeah sure

[01:09:53,818]: That’s Right

[01:09:54,538]: Because they are not the people who are so far beyond they kind of can’t get back

[01:10:03,718]: It’s what mistakes you made

[01:10:07,878]: And this is like when I was teaching I was teaching a class on how to write not bad

[01:10:13,258]: And I really had to think about what are the building blocks right

[01:10:17,418]: Not what do I do on a daily basis

[01:10:20,018]: Like I love the sound of my own voice

[01:10:21,878]: I could talk about what I do forever

[01:10:24,098]: 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

[01:10:34,598]: What can they access in an easy way

[01:10:36,638]: 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

[01:10:49,078]: I actually made them read a book called Save the Cat which is this incredibly rigid Yeah I’ve read it

[01:10:53,118]: Yeah

[01:10:53,538]: So you know Save the Cat

[01:10:54,578]: For the audience it’s this book about screenwriting which is incredibly prescriptive

[01:11:00,518]: Like this happens on page seven this happens on page nine

[01:11:05,518]: And I made them read that to understand

[01:11:07,938]: 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

[01:11:16,758]: Yeah sure

[01:11:17,678]: Because it’s too much freedom when you’re starting out

[01:11:21,178]: 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

[01:11:33,778]: But that’s hard for people to do

[01:11:36,078]: I got a lot of help from my husband who actually developed my opening slide deck how an op ed is like Star Wars

[01:11:42,958]: It was originally his slide deck I stole it and then expanded it

[01:11:46,998]: Because he’s smarter he’s much smarter about structure than I am

[01:11:50,118]: But all of that stuff that’s you know 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

[01:12:01,058]: But if I were starting now I don’t know I think and I think this is how university presidents feel too

[01:12:09,278]: It’s so overwhelming

[01:12:10,638]: Like what would I do

[01:12:11,578]: I don’t know get rid of tenure assess everyone on their AI abilities and their ability to teach and then you know stack rank those people and the people who can’t teach go

[01:12:25,518]: 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

[01:12:35,218]: And to be clear this is not me saying like academia bad

[01:12:40,058]: Journalism is going through the same thing

[01:12:42,018]: And we don’t like it either

[01:12:45,398]: 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

[01:12:57,958]: It’s not this is not a personal like it’s just a really hard problem

[01:13:01,798]: We got a lot of really hard problems

[01:13:03,318]: I think there’s a huge amount of stuff that is super exciting about AI

[01:13:06,218]: Yeah

[01:13:06,818]: I think there’s

[01:13:08,198]: But the problems are hard

[01:13:08,878]: As for journalism you know and taking jobs my instinct is that there’s two kinds of quote unquote journalistic writing

[01:13:18,838]: One is like you know the article about you know 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

[01:13:34,458]: Or let’s say you know an AI could downsize that breaking news department by 50

[01:13:41,458]: Those jobs seem to me very much under threat as would any job that is you know you’re writing WM you know WebMD’s top 10 ways to cure a cold or something you know

[01:13:55,778]: You know I briefly had a job writing like summarizing which strains of weed were the best even though I didn’t like smoking weed

[01:14:05,398]: I was just trying to break into the writing world

[01:14:07,658]: So I took anything I could get

[01:14:10,058]: I transcribed earnings calls was my first writing job for 85 bucks a call

[01:14:14,818]: Yeah

[01:14:15,378]: Well you got to get in somehow

[01:14:16,898]: So those jobs seem to me maximally under threat

[01:14:19,678]: 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

[01:14:27,918]: Why do I listen to a Coleman Hughes podcast or read a Coleman Hughes essay

[01:14:31,478]: It’s because partly I care that they wrote it because I know something about them

[01:14:37,638]: They have a brand that I care about

[01:14:39,398]: It’s a bit more like and this will seem a pretentious analogy

[01:14:43,778]: I don’t mean it to be

[01:14:44,718]: It’s just like the clearest case which is like why is an atom for atom copy of a Vermeer worthless

[01:14:51,498]: But the original Vermeer is worth millions

[01:14:53,278]: It’s like it’s not about the product itself

[01:14:55,678]: It’s about that Vermeer touched it

[01:14:58,598]: Right

[01:14:58,818]: And there’s something about that to listening to your favorite podcasters the columnists that you follow

[01:15:05,518]: So I don’t think you specifically or me or as recline 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

[01:15:22,018]: They read us partly because of the knowledge of who is behind it and whatever their association

[01:15:27,298]: Even if they hate us they read us because of the person behind it

[01:15:32,018]: Well knock wood you’re correct

[01:15:34,918]: Yeah

[01:15:35,098]: My assessment is this is that journalism has two problems and one is on the supply side and one’s on the demand side

[01:15:43,158]: And the supply side problem is this can be done by a machine

[01:15:47,618]: The demand side problem is we now have yet another thing that we’re competing for people’s attention with right

[01:15:54,678]: It’s you know social media that was true too

[01:15:57,358]: Every new thing that that competes with us for attention leaves less attention for us

[01:16:03,658]: The other thing is that on the supply side and yes I think you’re completely correct

[01:16:08,238]: The stuff where you’re just writing up public information that is going to be less and less done

[01:16:14,238]: There’s no reason to do it

[01:16:15,858]: No one not even 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 right

[01:16:29,318]: Because they can they can just ask ChatGPT much faster and get a customized and then unlike when we read it up when they have a question they can just say oh but what about this

[01:16:40,958]: And ChatGPT will tell them right

[01:16:42,998]: So that I think is just going to go away

[01:16:47,018]: Where do I think the value in journalism is going to be

[01:16:51,058]: It’s going to be in people and I think that in relationships and I think that also manifests in two ways

[01:16:57,558]: One is the value of your relationship with the reader

[01:17:01,978]: So people who have a brand people who have a following those people are and I am leaning in more to this

[01:17:13,378]: I’m leaning into the I am not writing up the economic data

[01:17:16,258]: I’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

[01:17:29,218]: Because people don’t need that from me as much anymore

[01:17:31,958]: And having being a personality having people feel like they know you

[01:17:38,198]: I joke that podcasts are your imaginary friends

[01:17:42,378]: And it’s a little weird because I actually know a lot of people on the podcast that I listen to

[01:17:47,598]: So sometimes I’m walking along and it’s like I’m having an imaginary friendship with my real friend and that’s weird

[01:17:53,318]: Right

[01:17:53,318]: Yeah yeah I’ve had that too

[01:17:55,058]: Yes right

[01:17:55,938]: It’s such a very strange Camille Foster is a good friend of mine and I haven’t seen him for nine months but I feel like I’ve been hanging out with him but he hasn’t been hanging out with me

[01:18:05,898]: So it’s like a it’s a problem

[01:18:09,058]: I have this experience quite frequently

[01:18:11,858]: But then the other thing is relationships with sources right

[01:18:14,758]: What is one thing Chat GPT can’t do

[01:18:16,598]: It cannot get a congressional staffer drunk and get it to tell them things about the tax bill right

[01:18:22,798]: Only a physical reporter can do that

[01:18:25,238]: And so that kind of information is also going to be valuable

[01:18:29,498]: And those are the kinds those are the two areas where and so if I think about how this is manifested in my own life partly it’s changing what I’m writing about

[01:18:40,498]: I’m writing a lot more about how to think about AI and how to because I think that’s where I can add value for my readers

[01:18:46,538]: But I also I spent this spring on a basically nonstop conference

[01:18:52,278]: There were months where it was literally like I came home usually on a Friday night sometimes on a Saturday to close out of the suitcase put the clothes in the washer put them back in the suitcase after they got out of the dryer

[01:19:04,538]: And then I just got on the next plane and it was grueling

[01:19:08,178]: And there was but the thing that you can do at conferences that you can’t is get new ideas talking to people be in a group listen to people right

[01:19:16,898]: And I love remote work

[01:19:18,418]: I’ve been working remote since 2006 but you got to get it’s not asking GPT isn’t the same because GPT only tells you what you asked

[01:19:27,278]: Right

[01:19:28,418]: 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

[01:19:40,638]: And so those are the two areas I think this is my analysis for what it’s worth

[01:19:45,498]: That’s where I am attempting to focus what I do which is really intensive

[01:19:51,918]: And then I have to like I’m taking the summer off

[01:19:54,718]: I’m old

[01:19:55,458]: I can’t do this anymore

[01:19:56,378]: But I’ll start up in the fall

[01:19:59,578]: 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

[01:20:11,278]: Or I hope there’s going to be maybe I should say I hope there’s going to be some value for journalists

[01:20:16,018]: But I do think you know who do I read

[01:20:17,858]: I read people

[01:20:19,038]: When I think about my own industry I read Dylan Byers and Oliver Darcy because those guys are on the phone all the time

[01:20:24,598]: And they’re talking to people and they’re giving me information that I can’t get anywhere else

[01:20:28,418]: Right

[01:20:28,818]: All right

[01:20:29,218]: Megan McArdle thank you so much for doing my show

[01:20:32,358]: Thanks for having me

---

---

Transcribe your media with TRNSCRB.
Transcribe multiple languages, effortlessly. Chat with your projects. Enhance your workflow.