(2024-04-06) ZviM On The 2nd CWT With Jonathan Haidt

Zvi Mowshowitz: On the 2nd CWT with Jonathan Haidt. In my childhood and education roundups, and of course with my own kids, I have been dealing with the issues Haidt talks about in his new book, The Anxious Generation. Ideally I’d also have read the book, but perfect as enemy of the good and all that. (Teen Mental Health)

I will start with my analysis of the podcast, in my now-standard format. Then I will include other related content I was going to put into my next childhood roundup

(0:00:30) What makes someone a better parent? Haidt says right wingers and religious folks make better parents, citing that they are increasingly happier and more rooted in communities, liberal kids are more depressed and more vulnerable to phones

being a good parent is not primarily about whether your kids are happy now.

Haidt will note later he is now a centrist, seeing both extremes as too illiberal, which he largely attributes to social media.

(0:02:00) Tyler Cowen asks who Haidt has met who is most wise, Haidt names two and finds many role models for wisdom. I notice that I find the opposite. I know plenty of very high intelligence (INT) people but find it hard to name very high wisdom (WIS) people I have met.

(07:30) Haidt values the Bible because of the need of every culture to have shared stories and reference points, comparing it to Homer and the Greek myths for ancient Greece and Rome. I agree that we need these shared reference points, and I increasingly worry about the fragmentation there, not only away from the Bible but also away from sharing popular culture stories as well

(09:00) Haidt opposes identitarianism in the sense of putting identity first as an analytical lens, and especially orienting others in this way, often in a mandatory way. He also warns of monomania

Haidt instead responds that identitarian political movements organizing for politics is fine, it just doesn’t belong in a classroom, citing past rights movements.

Haidt strikes me as someone who feels unable to speak their mind on this, who is choosing his words carefully and out of fear.

(13:10) Tyler asks, why won’t AI soon solve the screen time problem? The AI agent will process the information and then talk to you.

AI cuts your screen time by 2X or 3X.

Haidt absolutely nails the response

say that mundane AI doubles Tyler’s productivity in using screens to process information and complete tasks. What will happen to Tyler? Well, obviously, he will follow the supply and demand curves, and respond to decreased cost of information by increasing his consumption

The ability to use LLMs has definitely net increased my screen time so far.

For an average person, or an average child, the picture here looks gloomier still, to me. Time spent on television or watching videos or playing games will be made more addictive and to involve better selection via AI

Haidt says the primary problem with screens is opportunity cost of time, that they are experience blockers, and half his book is about the importance of play

we refuse to let kids be kids, do not permit them to go play unsupervised

I see this as the best counter-argument against ‘the phones did it,’ if the kids wouldn’t be allowed to play anyway then of course they will be on their phones and computers and televisions

Could AI instead be implemented in ways that simulate true play, that involve physical activity, that gives you virtual people to interact with that challenge you and train your social skills and other talents? That is definitely technologically feasible if we want it enough. But will the market give that to us, in practice? Will we choose to consume it?

why can’t we just train an AI friend to be like a real friend and get in fights with you sometimes?” Maybe in theory that’s possible, but that’s not what it’s going to be. It’s going to be market-driven. It’s going to be friends and lovers who are incredibly great for you. You never have to adjust to them. You never have to learn how to deal with difficult people, and it’s going to be a complete disaster for human development.”

“Complete disaster strikes me as too strong a term for something that hasn’t happened yet. I think you’re much too confident about that.” I do actually think Haidt is overconfident here

my answer is, again and again, that sometimes and in some ways you should be highly uncertain about future outcomes, especially when you lack parallels

There are indeed many things that have not happened yet, that I am confident would be ‘a complete disaster’ if they did happen

A large asteroid impact. A widescale global thermonuclear war and many other wars too. A pandemic

Tyler clarifies the real question, which is: If screens are making children so miserable, why won’t they use new AI innovations to fix that? Why are they so ‘failing to maximize’? To which the obvious retort is, it is not like there are no alternatives or innovations available now, yet the kids remain miserable. They are not maximizing now. The ‘market’ here has failed us. Children, even more than adults, do not optimize their consumption basket taking into account all dynamics and long term effects, mostly they (as per our experiments, this is not speculation) end up using apps with Skinner boxes

(20:15) Tyler challenges the importance of face-to-face interaction by noticing that the pandemic didn’t damage well-being for kids too much

time in school is time with people and friends, so the decline in 2020 must have been dramatic, yet well-being problems did not change much.

I note that I would be prepared to defy the data (if I need to do that) that mental well-being did not decline a lot for kids, or for everyone else, in 2020 and 2021? I mean, what?

Alternatively, we actually have an explanation for this, which is that schools are very bad for children’s mental health, as you would expect given what physically takes place there and how they treat children in most schools.

time within school is now largely spent with phones, not interacting with friends or those physically next to you, most students check their texts during class. So to the extent that time used to count, now it mostly doesn’t. After 2012, academic achievement goes down, loneliness in school goes up.

I would say: You can sort of count time when you are forcibly imprisoned next to arbitrary other people as social time, but that stops working if you instead have the option to ignore them and be on your phone.

Also we should totally ban phones in schools

if you don’t think we should ban the phones then at that point we should instead dismantle the schools, what is the point.

(25:30) Tyler asks why around 1900 European culture became more neurotic, depressive, negative and hostile

asks aren’t big shifts in mood often happening for small reasons, why attribute it to the phones? Why not simply say that big mood shifts we can’t explain are the norm?

But this isn’t history. It is now, and we can observe it in real time, and we indeed have a very good explanation of what happened.

Tyler keeps pounding on this later, so I want to say clearly: If there was an ‘exogenous mood shift’ in the 2010s, then all plausible candidates for it, including the rise of both wokeness and Trump and the loss of credibility of elites, are causally heavily intertwined with smartphones and social media

(30:00) Both agree girls are more mimetic, and this is one cause of them being impacted more by whatever is happening. (sociogenic)

everyone in the developed world, even in Eastern Europe, everyone — their kids are on phones, but the penetration, the intensity, was faster in the richest countries, the Anglos and the Scandinavians. That’s where people had the most independence and individualism

It seems important to be precise here. What this is saying is that it is the combination of smartphones and individualism that causes the issue

(31:45) Tyler also notes that phone usage explains only a small part of variance in happiness outcomes. Haidt agrees that the overall correlation coefficient is only something like 0.04, but if you focus on social media and girls the correlation coefficient gets up to something like 0.17, that even the skeptics are at between 0.1 and 0.15 without splitting by gender.

We also have to ask what directions causation goes to what extent. It is plausible that being depressed causes you to spend either more time or less time on social media, I can think of mechanisms for both.

(34:10) Tyler asks, why no talk in your book about the extremely large benefits of social media?

Tyler makes a pretty bold claim here: “At Emergent Ventures, we support many teenagers, young women. Many of them not 13 years old, but very often 16 to 19 years old. They’re doing science

They’re an incredible generation, smarter, more dynamic, probably more productive than any other scientific generation ever, and that’s because of social media.”

I can totally buy that there are a lot of very smart teenagers out there

But… the most dynamic and scientifically productive generation? Oh my is citation needed here, I do not believe this

Tyler gets to see the success stories, the most extraordinary people

(35:00) Haidt says he does have a section on benefits, which (matching Tyler’s statement) is almost always for older teenagers, he can see the collaboration story for them, but for 11-13 year olds they have different needs.

I see Twitter as a hybrid, that can be used in any mix of both, and as much less of the bad thing than other social media, but of course I am biased, it is vital for my work. I would be happy for a compromise that said kids get Twitter at 13 outside of school hours, say, but other ‘purer’ social media only at 18. Or even better

as per later discussions, you can get only the non-algorithmic ‘following’ version of Twitter at 13.

(36:45) Tyler once again: “It could be the case, maybe only 5 percent of teenagers benefit from this Twitter function, but that could, by far, outweigh the costs, right?” This seems to be a common pattern in Tyler’s thinking that is behind many of his weirdest takes, where he finds things he thinks are massively oversized in their benefits because in a small minority they promote the kinds of talent development or inspiration or capital formation (or what not) that he thinks is most important, and he is willing to throw the rest of life under the bus

one easy response is to say that this is the kind of child who should have special technical chops and determination and be impossible to stop

If they’re all that do you think you can keep them off Twitter?

(37:10) Tyler agrees that girls 12-14 are likely worse off because of Instagram. He dodges the question of TikTok, but it seems like his objection at that age is entirely about Twitter? Haidt says that we must talk price, the question is whether the age threshold should be 14 or 16, and he thinks that algorithmic feeds should be gated to age 16.

(38:45) Haidt claims Gen-Z spends a huge portion of their time and attention managing their network connections, it is the first and last thing they do every day. If true (and I think it is) this seems horrible, they are paying very high maintenance costs and not getting much in social benefits in return, in a way that makes it very costly to opt out.

One way to look at this is that we have raised social signaling costs that people can pay and made such payments highly visible

the worst kind of anxiety-infused life-consuming Skinner box.

Haidt refers to Patrick Collison noticing no major person in software is under 30, that Gen-Z aren’t starting companies and doing things. What young people are impacting the world?

(42:45) Tyler once again goes back to, whatever problems there are, why not just think we’ll adjust to them? We adjusted to agriculture

One response would be yes, we adjust, but taking social media away from kids like we took away leaded gasoline is exactly the kind of way in which we adjust.

Later Tyler suggests for example people saying “‘I’m going to form like a little polycule but without sex, and my polycule will be based around not doing so much social media.’ Like my friends and I in high school — we didn’t go to parties. We seceded from that.” This was young Tyler’s solution to the collective action problem. (tribe)

And yes, some people will always (in normal worlds) be able to form close-knit groups that ignore everyone else, and a small group of friends can do very well on all fronts, but that has always been highly limited as a strategy, most kids and people are incapable of it or won’t do it under the pressure.

Yes, as Tyler says, meeting up with your friends is fun, but when he says ‘kids will find ways of doing this,’ they are not currently finding ways of doing this. Time with friends is way down. Most social activities are way down. Relationships, sex and children are way down

(47:30) Flagging the huge agreement by all three of us that there is far too much homework, especially in the early grades. My kids school has them do homework with the justification that they need to learn how to do homework, the generalized version of which I would call the worst argument in the world

(48:00) Haidt frames his book as offering four norms that solve collective action problems and that would help get children time and ability to play as they need to, with number one being no smartphone before high school, let them use flip phones. Second, no social media until sixteen. Third is phone-free schools. Fourth is far more childhood independence, a la Free-Range Kids and Lenore Skenazy.

I am strongly in favor of all four planks as norms to strive for, especially taken together, and for the laws to at least facilitate all this

We need to stress the fourth one most of all, you can only take away the phones if kids can otherwise use that time.

we agree the social media decision needs to be up to the parents, at least at age 15. That you should require very clear opt-in from the parents, but if you have it, then go ahead.

(53:40) Tyler says, Instagram has parental controls but no one uses them. Haidt points out few people are able to use such controls well. I would add, the implementation matters. The defaults matter a ton

(55:00) Haidt says a lot about how the government is not doing anything to enable safeguards. Tyler points out that any version of this is effectively a ban, that it would bankrupt such companies if they could be sued every time a kid got on without permission.

Presumably this is one of those ‘either our legal system has rules for liability that work, or it does not’ situations?

When tech companies say they cannot survive ordinary liability law, that implies strongly that either we should change that underlying law for all cases, or there is something deeply wrong with the business.

Tyler doubles down, says even if Meta was 99% effective, they’d still be sued into oblivion on the other 1%. Whereas Haidt says correctly, that would be incredible, great success, we happily accept a 1% or even 5%-10% failure rate here

As a bonus, here are two sections that would have been in my next childhood roundup:

Ban Phones in Schools

England to give the power to ban mobile phone use on primary and secondary school grounds, students will have to switch them off or risk confiscation

In the US, a teacher (Mary Garza) instructed her students to set their phones to loud mode. Each time a notification was received they’d stand up & tally it under a suitable category. This occurred during ONE class period. Each mark is a learning disruption

Did you know that Snapchat+, the $4/month subscription service, offers friend rankings?

This often goes exactly the way you would think, with both friend and relationship drama ensuing when someone is not ranked high enough.

Katy Potts: I call it the “anxiety app” in online safety training I run - grim - unbelievable they get away with it.

*Haidt also wrote a book, The Anxious Generation. As I noted earlier I haven’t had opportunity to read it. Candice Odgers reviews it here in Nature. Here is the teaser line of the review.

The evidence is equivocal on whether screen time is to blame for rising levels of teen depression and anxiety — and rising hysteria could distract us from tackling the real causes.*

Remember The Law of No Evidence: Any claim there is “no evidence” of something is evidence of bullshit

The studies are asking the wrong questions, this is dumb

I am confused what we are even at risk of being distracted from. What does she offer? Researchers cite access to guns, exposure to violence, structural discrimination and racism, sexism and sexual abuse, the opioid epidemic, economic hardship and social isolation as leading contributors. The idea that kids today have more contact with guns, violence, structural discrimination and racism, sexism and sexual abuse than they did in the past is obviously backwards

*Economic hardship is complicated, as I’ve discussed in the past, but certainly there has not been a dramatic rise in economic hardship starting in the mid-2010s.

That leaves the opioid epidemic and social isolation, which are indeed getting worse.*

The opioid epidemic is bad, but this can’t be primary. The fall in child well-being doesn’t map onto the opioid epidemic.

Yes, there is narrative among the youth that all these things are worse than ever. And that narrative is bad for mental health. But do you know what is a prime driver of that? Social media and everyone constantly being on their phones.

There is an alternative hypothesis that does make sense. One could say that kids are on their phones this much exactly because we do not let kids be kids.

Let Kids be Kids

We used to let kids babysit other kids. I remember having at least one sitter, a neighbor from upstairs, who was only twelve or so. As opposed to now, when someone is terrified their 13-year-old is in a house with a friend, their mom and an uncleared third adult. We still use the term ‘babysitter’ but it means paying an adult at least $25 an hour, rather than letting kids learn some responsibility and earn some cash.