(2026-10-16) Zvim Monthly Roundup47 October2026

Zvi Mowshowitz: Monthly Roundup #47: October 2026. AI might not yet be eating the physical world. AI is still eating the world of news, and I am finding less and less opportunity to post about other things

Table of Contents

• Home Alone.
• The Men Have Never Been Okay.
• Antisocial Media.
• While I Cannot Condone This.
• Zero Marginal Product.
• Good News, Everyone.
• 100 Good Television Shows.
• For Your Entertainment.
• Gamers Gonna Game Game Game Game Game.
• I Was Promised Flying Self-Driving Cars.
• Sports Go Sports.
• Government Working.
• Jones Act Watch.
• Variously Effective Altruism.
• Walk Away.
• The Lighter Side.

Home Alone

Americans are abandoning quite a lot of formerly popular activities, especially hosting at home, as Derek Thompson discussed. Phones and other electronics are in, most things that are not phones are out, with notably rare exceptions.

That last line rules everything around us. And we all spend more time at home, which then exploded during Covid and mostly did not recover afterward.
Derek does discuss the last two lines later on.

You can view exercise classes as an adaptation in four ways: To a lack of other physical activity, to a lack of social activity, to a lack of motivation, and as a wealth effect. Regular people can afford this, and it is a safe way to get things they now lack.

Then there are the concerts. This one is what some call a Narrative Violation. We have this collective story about how in the good old days, like the 1960s, people went to all these great cheap concerts, that was part of the good life

actually they’re making a quiet comeback despite people getting older and all the competition, and also they are great. On the flip side, they are also ‘free’ actions, in that you don’t have to overcome any social barrier. You can buy a ticket.

Whereas hosting is, as Derek says, a lot of work to set up, often not fun work

Parties also require (and help create and maintain) friends, and creating and maintaining those friends takes even more time.

Meanwhile, being alone at home is no longer boring. Being alone at home is great, in the short term.

Derek Thompson: Machines in the last half century have accomplished something stranger. They have made other people less necessary to our leisure

This is not so much an individual failure as a coordination failure.

*There is some coordination failure, but I think this is the wrong central diagnosis. Coordination failure can be overcome, and given the will to do the thing the logistics are far easier today with all our social and communication and logistical tools.

The bigger problem is hyperbolic discounting in the face of the new alternatives*

*In order to get the returns from social activity, you have to invest energy and effort now, and also endure some mild social awkwardness and delay gratification and tolerate some boredom now, and often do things for someone else, in order to build up your friendship and other networks and get into regular habits.

I’m also guilty of this trap. I don’t socialize enough. I don’t host enough, and I don’t visit enough. I realize this*

In my case I like to think I have a good excuse, but it is still a mistake.

The Men Have Never Been Okay

Institute for Family Studies reports at length on America’s demoralized and lost men. Their central presentation is not that young men have, as Senator Hawley put it, ‘no vision for what it is to be a man,’ or that their vision would be bad and captured by toxic influences. It’s that they do have such a positive vision, and it would be good but they despair of living up to it and do not see a path to getting there. We no longer lay out a simple ‘success sequence’ style strategy where you play by the rules and it does not work even for those without problems, and many have problems

But 84% say they are at least somewhat ambitious and 81% say they can ‘do just about anything I really set my mind to,’ 75% say their opportunities in life will be greater than their parents

This seems like a lot of very strong framing effects combined with the Revolution of Rising Expectations.

hard work is valid, and if you want it enough and keep your head on straight and resist temptation and despair you can still succeed. But do you want it enough? Can you keep your head on straight and resist temptation and despair?

And can you keep searching for an unknown, rather than for a known, so long as the ultimate odds have not changed so much?

if you keep pushing, and try to make your own luck, you will probably eventually get lucky.

Over half of young men feel lonely ‘at least some of the time’ but it is kind of weird not to, isn’t it? The right rate of feeling lonely, like many bad things, is not zero.

The report has many other interesting charts. It would be cool to see time lapses.

Remember that throughout history, a huge percentage of men failed to have children. It is not a new situation for men to be tested and have to find a way to prove themselves.

Broadly my current perspective is the men as a group have never been okay. The difference is that we used to sort of think not being okay was fine. Now we don’t.

Antisocial Media

Texting more often is associated with higher life satisfaction. Tyler Cowen asks ‘should you text more?’ and my response would be ‘yes, but only insofar as you should have more friends that you want to text.’

Twitter identifies and suspends a Chinese bot farm of ~200,000 accounts engaged in influence operations. This included 200 working to build American opposition to data centers. Those 200 were shown. The other 199,800 were not shown. I wonder what they were up to.

While I Cannot Condone This

If you don’t want to wrongspeak, you have two strategies. You can suppress thoughts, or you can suppress words.

Paola: I’m realizing that my ratio of “thoughts that explicitly enter my consciousness mind” to “thoughts I share out loud” must be absolutely crazy relative to the general population. My sense is that agreeable people basically suppress their thoughts as opposed to just their words and it’s at least slightly weird that I’m doing basically 0 thought suppression and a large amount of word suppression.

In Western Massachusetts, in the year 2026, there was a lesbian bar still requiring KN-95 masks. They changed Saturday nights to mask optional, and all hell broke loose like it was still 2020: “Just say you’re being complicit in a mass disabling event and get it over with.”

And then it got worse, or better, I’m not sure which.
Instead of defusing the situation, the apology seemed to make things worse. The bar had erred: it posted an Instagram card without providing alt-text for the visually impaired. This set off a fresh round of accusations of ableism.

Zero Marginal Product

Fun hypothetical:
Kurosai: Do you know how fucked up you have to be to reject a job that lets you sleep for 45 and hour? This is a joke.

Good News, Everyone

We have an excellent permitting deal, the Bipartisan American Affordability and Jobs Act (S. 5653), if we can get it over the finish line into law. I don’t have time to properly geek out on all the details, but this is a big deal. Even better would be huge sweeping ‘cut the Gordian knot’ changes, but if you’re going to do incremental stuff this would be pretty great.

Back in July we got a report to the president called Science: A New Golden Age. In a better or quieter world I would have looked into this in detail. I am sad I never got to do that, but I need to admit defeat. By all accounts there were a lot of good ideas here.

I confirm, based on both personal experience and what I see in others, that Justin Skycak is correct here: High performers start out with an edge that motivates more work with increasing returns to scale, then work more and harder than everyone else, and end up acquiring a vast array of related skills as they climb the entire associated tech tree. You can see it in various places. In sports, where the top players are better at actual everything, including all the tiny details. In media, where high quality things are just high quality across the board, with no limit. And so on.

100 Good Television Shows (TV)

A remarkably good list, the product of an industry survey:

True Detective is a weird case, where they’re finally going to make Season 2 soon, and they’re calling it Season 5. Nicolas Cage will star.

The next question would be, how many are conspicuously missing?

*I have 7 Tier 1 TV shows. 5 of the 7 make the list, and the other two are obscure and thus highly forgivable mistake. All the straight dramas on my list make their top 20, including their #1 and #2.

I have 40 Tier 2 TV shows, 30 of which are eligible. Here the list does a lot worse, and only has 10 of them. I’d only call maybe 5 of the remaining 20 objective mistakes, rather than differences of opinion. That illustrates how long is the long tail of television.*

*In any case, yes, Matthew Yglesias, there are not only 100 good TV shows but 100+ great TV shows out there, and yes you do need a canon so people can find the great ones. Matthew’s main argument is that if you have a great TV show, and then before it gets cancelled it gets bad, then the whole TV show is now not good, and this happens often enough that there aren’t that many good TV shows.

I think this is bollocks whenever you can pretend the later seasons do not exist.*

For Your Entertainment

There is talk about rezoning the Kips Bay AMC 15 site, my second-most common theater location, which will close at the end of the year due to termination of its lease, which was bought by NYU Langone, who wants to use it to expand the hospital. Some including the Borough President are insisting that a cinema and a grocery store (currently there is a Fairway next door with a long term lease) be kept. The grocery store part is absurd.

Gamers Gonna Game Game Game Game Game

I Was Promised Flying Self-Driving Cars

Waymo was already absurdly safer than human drivers, and it keeps getting safer, despite reports that the driving is getting less passive.

Comparison w/ human drivers on rate of crashes with serious injury:
Latest data: 270M miles, 20X better
Mar 2026 : 170M miles, 13X better Before that (forget exactly when): 10X better

*The Chinese are very much in the driverless car race.

David Manheim: I just wondered why other countries weren’t deploying autonomous taxis, then checked. China’s Baidu has ~2000 cars under “Apollo Go” in 6 cities. Waymo has ~4000. I’m pretty unsure who wins scaling globally, given the larger addressable domestic market in China, vs Waymo’s lead.*

This is presumably not winner-take-all, rather it is a threshold oligopoly.

The Waymo debate is happening right now in DC. There is a strange emphasis on ‘countless’ ride share drivers, given DC only has 200 full-time ride share drivers among its residents, and as always calls not to ‘rush’ legislation.

Sports Go Sports

Government Working

You can just whiten teeth, except you can’t, because dentists. Many such cases

Airlines will no longer be required to provide compensation for lodging, meals and other assistance when flights are delayed or canceled due to certain causes now classified as ‘uncontrollable.’

The airlines do have some influence on outcomes, especially since the top reclassified cause is ‘unscheduled maintenance.’

I get that maybe you need cannabis to be legal because of the alternative but why the hell are we advertising its legality in the subway with public service announcements?

Mamdani to ban Central Park’s horse-drawn carriages, one of the famous things to do in New York. I’ve never actually done it, but I feel like I should do it now while I can. Sounds like we will have until June 2028. You know it is deeply stupid when they are laying off 200 people and destroying a cultural touchstone purportedly in large part because of one accidental death, of Romanch Mahajan, and also because people think it is inhumane to the horses to have them draw carriages, and various other things that are not true.

Of course then, he’s promising to pay the 200 people to not drive horses around, and also paying for the horses people will not ride in on. You can imagine what I have to say to Mamdani, and also his horse.

Jones Act Watch

Yes, America still massively subsidizes shipbuilding, but so does everyone else.

Variously Effective Altruism

The bad guys have come with bad faith attacks against Effective Altruism, as has the Department of War

I certainly have my critiques of Effective Altruism.

That’s because the EAs are trying hard enough to be worthy of critique. So critiques they shall have, good ones, that hold them to the standards they aspire to fulfill.

This wave is different. This is a deliberate campaign of attack, the true face of politics. No one involved is trying to make anyone better.

That’s not cool. No one picks on my little brother but me, especially in bad faith.

There are plenty of public and acknowledged connections between various EA orgs and people and funders, which causes some to then try and force QAnon styles of connections, such as ‘person [P] who funded or invested in [X] also funded [Y], therefore [X] funded [Y].’ Which of course is not how it works.

A lot of the time, the true argument is of the form ‘there are these weird (and other attributes you or your politics will dislike) people who care about AI safety, but as I’ve shown such people are weird and bad, therefore you shouldn’t care about AI safety.’

Then there are those who accuse the EAs of saying ‘only they can be trusted with AGI’ or that they want to take over the world. Some form of this is narrowly true of the leaders and many employees of the major labs, which includes many EAs at Anthropic, but the same is true of many at OpenAI and Google DeepMind and also Elon Musk, which mostly are not EAs. The active ingredient lies elsewhere.

There is also, yes, a faction that says ‘no one can be trusted, but someone is going to build it and this is our best option,’ the same way that most people are not happy about their choice of Presidential candidate these days. Many are quite upset that Anthropic turned into another race participant, and then there is a split about whether that still makes Anthropic a good enough least-bad option. I believe, in both cases, that those saying ‘you have to pick one’ are making a mistake, but it is similar.

Then there is the larger group of EAs warning that no one can be trusted to build AGI and ASI, period. People are accusing an atheist group of being monotheist because they don’t believe in anyone else’s Gods.

I especially don’t appreciate those who accuse EAs and rationalists of being ‘uniquely isolated’ or ‘ideological’ or attached to their beliefs or anything like that, as in:

Another common one, also available within that thread, is ‘if you follow their logic to its conclusion it implies bad things.’ Yes, if you blindly and endlessly extrapolate from chosen beliefs, in a way that approximately zero people actually do, without pausing to realize your mistake, you eventually get to some rather large mistakes. That is true of everyone who believes something, and also is an important hint about difficulties with AI alignment.

I do think, similarly to Trace Woodgrains, that EAs in particular will often use calculations to imply things that if fully naively extrapolated would go badly, biting bullets in theory, but then in the face of messy reality they wisely pull back in practice.

It got bad enough that Tyler Cowen had to go and write In Defense of Effective Altruism. The defense is measured. EA is endorsed on the margin and as directionally correct. As much as Tyler ‘just says things’ regarding AI existential risk, he agrees that obviously we should pay more attention to AI safety, as well as other EA causes like global poverty. Does Tyler have digs and critiques as well? Why yes he does, but to be directionally right on the margin is high praise from a coauthor of Marginal Revolution.

I think ‘directionally correct’ is important and spot on. Yes there can be overshooting. I do not, for example and as noted above, endorse impartiality, but most people could be less partial. You don’t want to rely on only doing calculations, but most people desperately need to do more calculations and rely on them more. And so on.

I think it is actively a good thing that some people make stupid and stupid looking mistakes like the Shrimp Welfare Project, that then are endlessly mocked, and there is very little call to try and jettison such views or people. You can be wrong.

Any group that collectively takes a handful of big ‘look over here’ shots in the 2010s, where two of them are pandemics and AI, should get you asking what they did right, and yeah maybe let a few of them sound crazy and talk about electron suffering. You don’t have to help fund them.

There are of course various EAs and rationalists and associated persons with money and power and connections. There is no inner ring. There are only other rings.

roon (OpenAI): this is so stupid. there is an inner ring of effective altruists and they are well on their way to being trillionaires, or they’re professors at Oxford, or have a tremendous political lobbying machine

Oliver Habryka: Eh, I think this is overstated.

The top of the community really is a bunch of people who basically never talk to each other. What’s your guess when Will last talked to Dario last talked to Eliezer last talked to Leopold last talked to Alexander Berger?

Walk Away

One of the best rationalist norms is that if you are in a conversation, and that conversation is boring to you, or you otherwise want to leave, you can quietly leave.
Words cannot express how wonderful this norm is. A lot of why Lighthaven is so wonderful is that within its confines you can use this norm.