What a month. So much to cover.
What this post does not cover beyond this introduction is the biggest news story of the month, a potential room temperature superconductor.
If this discovery is real, it could be transformative. Think of the potential. We could be so back. Chances are that instead it is over, but even a small chance of something this big is huge. Even if it ends up being all over, it was amazing to see people around the world come together and do science to this to try and actually figure out something physical and real. Even in failure, we are so back.
What I am not going to do is pivot to suddenly becoming an expert on physics or get into the weeds myself. There is no need for that, I work fast and it would be fun but there are limits and I simply haven’t the time. We have a whole world stepping up to replicate the results and soon we will know the answer.
In the meantime, it is easy to know the probability of the answer. All you have to do is check Polymarket, for the largest real money prediction market. Size talks.
Polymarket does not allow Americans and requires crypto. So if you are in American or want to get involved via charity funds instead, you can bet much smaller amounts at Manifold Markets instead. They also have an endless running discussion, whose quality I am choosing not to be qualified to judge.
You can also enjoy my movie retrospective, Barbiehammer: Across the Dead Reckoning. It has been quite a good year for movies.
So what else has been happening?
How bad are car break-ins in San Francisco and Oakland? A CNN reporter shooting a story about crime gets her rental cars broken into so often she is careful not to even leave candy bars behind. Then we get the hard data, which is that of 250 cars returned the previous day, 27 had been broken into. A reasonable guess is that the average rental period is 5 days. So that means that there is a 2% car break-in rate per day.
Portugal is struggling with drug legalization, Washington Post reports, as rates of use go up and the systems for dealing with drug use increasingly aren’t working. Whenever full legalization happens, this effectively means that drug users are violating a whole host of regulations and rules and degrading general quality of life, in ways that would never be tolerated from other ordinary citizens or ordinary business activities, but no one has any leverage for enforcement once you lose the teeth of illegality.
We make many aspects of ordinary living illegal unless large regulation-imposed mandatory costs are paid, but we enforce that be threatening your standing in society, which drug users are willing to lose or have already lost. And allowing more drug use increases the number of people who end up like that over time. What to do?
You now have even better reasons to be completely done with RFK Jr. The list goes on.
US issues travel advisory to consider avoiding China because of arbitrary law enforcement and exit bans and risk of unlawful detentions. What changed? I presume that we wanted to say boo China. The risk is real and does give me pause, but this issue has been there for a long time.
‘CRATER’, starring Mckenna Grace and Kid Cudi, has been removed from Disney+ due to cost-cutting reasons. The film just released on Disney+ on May 12.
Will Feral: gotta say at this point the pirates won the moral argument
Yet another debate debate continues. Previously I looked at claims from The Free Press that debate had been taken over by judges uninterested in debate who instead rewarded whoever successfully deployed the proper left-wing shibboleths. That post failed to provide statistics or base rates. After looking at judge profiles, it was not clear that there was a severe problem.
Maya Bodnick’s Slow Boring post was more convincing. By doing a systematic tabulation of high-level debate tournaments, she finds that the strategy of ignoring the debate topic entirely in favor of advancing an arbitrary and orthogonal critical theory – a strategy known as a ‘kritik’ – is highly popular and successful. As level of competition intensifies, frequency of kritik strategies goes up, composing a majority of policy debates and almost half of Lincoln-Douglas debates. Which means that even when opponents are ready for it, such tactics work. There is a dominant strategy of debate for those with the resources to learn to use it, and it does not involve the topic of the debate.
Instead it requires arguing obvious nonsense, with failure to accept such nonsense as the baseline of discussion, or insisting that it is irrelevant to the topic, treated as a de facto automatic loss. The format is broken when played at the competitive level, so you can perhaps get value playing it casually, but that’s it, and if someone pulls out a bazooka then you lose. A shame.
Are we indeed this lame? It seems that we are. We also don’t value unique identifiers sufficiently, also the people choosing this many white cars seem low-level insane. Product variety is one of the things that make America great and we are letting it slip away. Worst of all is that Uber and Lyft seem to encourage black cars, the hardest ones to distinguish, when strangers have to find your car constantly. Give me the purple or green Uber every time.
A fine explanation of the dangers of all loyalty statements, which have now come to mathematics.
Why are Kosher restaurants so bad? The economics. You can’t be open 30% of the time due to Jewish holidays and Shabbat, including half or more of the Friday and Saturday nights. Your food costs are through the roof since you have to buy from the handful of Kosher venders, and you need to hire extra staff for certification. Taken together, that’s a nightmare, especially with many Jews not keeping strict Kashrut. Perhaps a crazy idea is to combine the restaurant with the Synagogue? If one is open, the other is closed, so you get to reuse space and even employees, and everyone dining there knows the proceeds support the shul and gets to complain to the rabbi.
Tech No Longer or Not Yet Calling Itself Artificial Intelligence
Patrick McKenzie and Ryan Peterson warn and remind us that SMS for two-factor authorization is very much not secure when it matters, that you can get your account taken over by the method of finding a gullible Verizon employee willing to swap SIM cards. It is vastly more secure than relying only on a password, so as a standard thing for most people it makes sense, but do not confuse it with security against a determined attacker. This echoes many other safety issues, including in AI both present and more importantly future. There is safety for practical purposes that breaks down when you need it most, and there is safety that is robust to adversarial attack, and solutions for the first often are useless for the second.
Dustin Mowshowitz (private account) lays out the reasons Twitter’s current owner engages in censorship, in addition to the ones we mostly all agree upon: When it’s about him, when competitors are being promoted, when authoritarians ask him to, and sometimes on left-wing things he disagrees with. I have yet to see substantiation on the left-wing thing happening often enough to matter, the others seem clear and need to stop. In particular, treating Substack as competition and a rival, as opposed to part of the same synergistic ecosystem, is a huge unforced error doing real damage. Yes, Substack Notes technically exists, which is what caused Elon’s fit of pique, but I’ve never seen a link to it or had it be relevant to anything at all. Does anyone actually use it?
An accusation from United States v. Twitter that the FTC flat out told Ernst & Young to produce a negative report, and that E&Y feared retaliation if they didn’t.
It has been explained but I will never stop being confused by the dislike of personalized advertisements, which are universally far superior to the alternative.
Elon Musk: This will increase the spam level of ads. Fortunately, we already suck at ad targeting on this platform.
I get the view that this could be predatory or what not, but in practice what you get are fewer, higher-priced, higher-relevance ads that people actually do not mind seeing. The alternative is a scattershot army of random useless junk. Perhaps we can have a compromise, where we define certain types of advertising as potentially predatory – such as things like gambling or recreational drugs or multi-level marketing schemes – and say that those in particular can’t custom target.
What we actually deserve, and might benefit from requiring, is user control over ad customization. This would benefit everyone. The ads I want to see are the ads you want to show me. For me, the best ad types are movie previews, game advertisements, restaurants and of course delicious Goldbelly-style temptations.
Flashing elements of websites are increasingly ubiquitous. Scott Alexander cannot stand any amount of them however unobtrusive and harmless, like the site saying ‘draft saved,’ and 16% of his users agree. As with many similar design choices, I think the ‘draft saved’ level is a good default, I don’t even notice it, but am baffled that there is no settings option to turn it off.
Car knows some things in some places, not in others.
Cate Hall: When we start driving, car says it has 230 miles left on the charge, 130 miles from home, but says we’ll get back with 11% (~30 miles) remaining. We’ve driven 15 miles since (115 miles from home) but are down to 205 on charge (-25 miles) — any guesses what’s going on?
We recently updated the software & tire pressure is good. We’ve been driving for most of the last 2 days & recharged 8-10 times over that period.
There are quite a lot of time or duration estimation systems that are reliably horribly wrong, with sufficient reliability that you can adjust to find the right estimate. Often you see exactly the above, where you get two answers from two different processes, one of which is lying. It always points to something very wrong on some level, so if you care about fixing problems it is highly useful.
It continuously baffles me that such estimations cannot be done better. Food delivery time estimates consistently make no sense to the point that I ignore them.
San Francisco residents attempt to sabotage driverless taxis by putting orange cones on the car hood. Tyler Cowen says solve for the equilibrium. I believe the equilibrium is a software patch. There is no law of nature that says this trick has to work. Certainly the solution won’t be that authorities enforce the law and stop this, not in San Francisco. Also I cannot express how much I deeply dislike the people doing this, or endorsing it. No, they’re not asking you to be guinea pigs, they are offering an amazing service.
Dan Luu predicts Threads will have more daily active users within a year, Manifold Market put this at 41% earlier, it is now down to 22%. Dan’s logic is that Meta will make better decisions than Twitter or other alternatives. He notes that Threads is exactly what he does not want, but that could be a lot like what most people want. Prove him wrong, kids. You’re off to a great start.
Several studies show that Facebook’s algorithms create echo chambers, far in excess of what would occur with a chronological or balanced feed, yet a short term correction of this did not noticeably alter political views. Tyler Cowen sees this as vindication of social media. One could also view this as reflective of damage already being done, a temporary reduction in the error chamber effect on one website being insufficient to reverse it, or more generally as reflective of an insufficiently powered or specified set of studies. I do think people unfairly blame Facebook for its impact on politics, when we should instead be blaming it for making our lives directly worse in other ways.
A theory that academic research recently has been similar to when you could not compete for the Tour de France without doping. Not all academics cheat, but the rewards for cheating have been high and the chances of being caught low, so among the ‘successful’ academics we should expect to see a lot of cheating. If academia is asking big questions and seeking important truths, this effect will not be as big, because your fortune or skill in other ways can overcome this. If the game is all about bureaucracy and signaling and social games and p<0.05, perhaps not so much.
EU to require online Visa applications by Americans before visiting, including background check and work history, in retaliation to a similar set of rules by the USA. It seems that the response to shooting oneself in the foot is to follow suit. The following is very right, although it won’t fully be half.
Paul Graham: A lot of people replying to this seem to think the problem is the cost. It isn’t, of course. The problem is the extra friction. American tourists aren’t used to that. This will stop half of them in their tracks.
More related bad news you can use:
Ryan Peterson: So insane that the way to get a passport processed is to contact your congressperson because they each have a dedicated person for this. I’m no org design expert but what if instead we put those 538 people in the state department passport office and cranked through the applications?
Patrick McKenzie: So to “Yes and” this observation: the government has many people working for it, and chooses to use specialization among them, including stratification by skill, drive, and status. The reasons are similar in character to the reasons tech companies do these things.
I have some reservations about tech’s implementation of this, but observationally, no serious person doubts there is a wide gap between business decision makers at large tech companies and between their CS teams. Both can touch CS cases, though, and not infrequently do.
Why is that? Well, partly it is an economic optimization. Partly it is very difficult to hire arbitrarily competent driven people and staff them, then keep them staffed, in a department seen as (and treated as) low status.
So you end up with creative arrangements whereby some people working CS have titles that are not Customer Support Representative and their pay is six figures and they do not report to a manager who only cares about tickets per hour.
And why, if we can do that, do we have large CS teams grinding out tickets in call center type environments? Because the economic case for that is *extremely compelling.*
“Put a number on ‘extremely compelling’ for me.” More than a factor of ten, driven both by higher utilization (under call center management style versus working professional style) and lower fully loaded cost of the employees at issue.
People sometimes diagnose this as greed or incompetence, and… the world is complicated.
You can have professionals intervene for you in a variety of fields on a concierge model. They will cater to your every whim and then present you with a bill for services.
One group of such professionals is called “Lawyers” and peoples’ complaints about how much they charge for six minutes of their time are so well-established that they are almost folklore.
The option people are most willing to pay for is a call center.
A further way in which the world is complicated: even accepting the above defense of the status quo, there are frequently large uneaten free lunches in e.g. routine administration of government.
I made the observation at lunch recently that an anecdote from a particular state agency suggested it’s executives were less competent in operations management than a 16 year old who had recently beaten Factorio and, reader, I stand by that statement.
But there are complicated incentives for executives of government agencies and often excellence in swiftly turning paperwork into outcomes would conflict with their more important goals, like personal career advancement or functioning as an effective jobs program.
Will the UK really drive all encrypted messaging services out of the country? (BBC)
Some claim that this is actually on track and going to happen.
Alec Muffett: WOW: – No Signal – No WhatsApp – No iMessage – No Facetime
All of Apple, WhatsApp & Signal will be pulling their products out of the UK market because of proposed Government surveillance obligations. Yes, really.
All that will be left will be SMS, Email, and a few niche/small messenger apps for geeks, with poor international coverage.
This seems like the kind of thing that continuously threatens to happen, then does not actually happen when people realize the consequences. They do on occasion actually happen, though.
Our system of ‘any old judge can stop anything for any little procedural reason no matter how much buy-in it has’ has to end. This is so far beyond ridiculous.
Matthew Yglesias: Judges having a normal one.
James Coleman: Big news as 4th Circuit stops construction of the Mountain Valley Pipeline. Congress & President had mandated approval, forbidden review, and said any further challenges must go to D.C. Circuit, not 4th. (As part of debt ceiling deal.)
In other NEPA news, the CHIPS bill is not going to do much if we don’t address those concerns, as well as other regulatory requirements effectively nullifying our efforts. There is a bipartisan bill to streamline NEPA for this particular purpose, the Building Chips in America Act, creating a limited carve out from review. What we actually need is complete reimagining of the entire NEPA process, which I hope Balsa will be able to outline.
Why is turbulence worse on planes? The headlines blame it on ‘climate change.’ The actual answer is the FAA told airlines to prioritize saving fuel over passenger comfort, despite passengers having a strong revealed preference for spending the extra cost of fuel to have a more pleasant flight. This then became ‘because climate change.’ This kind of thing damages public trust in all such claims, making solving climate change (and everything else) that much harder.
Sugar shortage threatens candy production ahead of Halloween. There is of course no actual sugar shortage, there is a shortage in America because of Big Sugar’s import quotas. I mostly want to say ‘good’ here, we are too generous with Halloween candy and it devalues the whole experience. A supply shortage makes it fun again.
A cute tax trick (direct link):
Jesse Eisinger: NEW: -Buy your house for $26M -Get it valued at $130M -Put it in your “foundation,” meant for the public -Allow the public access 2hr a week -Voila! Massive tax break
There was a 14-year gap in which some appreciation doubtless did occur. Still, not like this, also this was literally higher than any recorded private home sale in the country.
Emmett Shear notes that with FOIA and streaming of Congress we ran the experiment on a more transparent government, and the result is the real work mostly moving away from monitored channels towards unmonitored ones. If you read everything I write down, my response is to not write anything down, and so on. If you went full 24⁄7 surveillance you’d make things even more implicit and also drive actions towards whatever looks and sounds superficially good.
I think this is right. You want to protect against corruption. You want transparency of who does what, and who gets paid what and how, and other such outcomes. You also need to allow the people involved to sometimes speak freely, ideally also to write freely, especially in the age of social media.
Yes, we do need to fix this, requiring ID to buy spray paint is an outrage.
Mike Stabile: I hear this argument A LOT from censorship advocates, usually in regards to liquor or Playboy, not spray paint — so let’s quickly tackle this, shall we?
Lalia Mickelwait: A person has to show ID to buy spray paint. But no ID is required to have a person’s nude body globally distributed as amateur, user- generated, masturbation material to over 100 million users per day on the world’s most popular p*rn sites? Make it make sense!
I would actually be mostly fine with mandating porn sites require ID in order to upload material, for authorization reasons, there are already plenty of similar rules. For viewing it? The issue is that there is no ‘flash you ID badge to prove you are 21 without creating any stored data’ on the internet.
When you buy liquor (or spray paint, who knew?) you flash your ID. It’s quick, it’s easy and it’s free. If this all that was required of adult sites, I don’t think there would be any complaint. But that’s not what’s happening.
I notice I remain confused that no good technological solution exists here. In such a solution, some websites would be classified as ‘adult.’ To access them, you present a token that contains only the information ‘this is an adult’ as provided by, say, Apple or Google or whomever, with no identifying information and no ability to trace it back, and they check your ID one time at some point. If you want your child to not do this, you give them a phone or computer without such verification, and you can easily do a spot check to see if they’ve hacked around it. Yes, it’s not bulletproof but it does not need to be. Why is this impossible or non-viable?
On a podcast, Connor Leahy pointed out that there are larger regulatory barriers to selling a sandwich than there are to deploying a superintelligence. There are two different ways to fix this and I am open to both of them.
FTC attempting to rewrite American merger guidelines via administrative fiat. Together with their actions against Amazon (and one could add OpenAI and others) it seems clear the FTC has gone off the rails in terms of its alignment with consumers. Alex Tabarrok has a follow-up here including links to more analysis. The new guiding principle seems to be that big corporations are presumed to be no good and opposed to consumers, so if mergers are good for big companies then they must be bad for consumers.
There is no more vile industry or lobbying group than the tax preparation firms, whose entire existence seems mostly to be lobbying the government to require mountains of completely unnecessary paperwork so that citizens will pay the firms to make this problem go away. The fight has now shifted to a new file-free electronic tax system the IRS hopes to debut.
Can I say, perhaps, new EA cause area? Since 2006, total lobbying by these companies is on the order of $40 million, if we assume all their lobbying money is opposing filing reform. Public opinion is overwhelming here, and there are approximately zero other people on the side of more paperwork, so presumably you would at most have to match that rate of spending.
The logic invoked by those who take the bribes is so absurd.
Republicans on the House Appropriations Committee in June proposed a budget rider that would prohibit funds to be used for the IRS to create a government-run tax preparation software, unless approved by a group of House and Senate committees.
The move “safeguards the IRS from an obvious conflict of interest where the tax collector becomes the tax preparer,” the bill’s summary states.
The idea that the IRS is going to use this opportunity to cheat the taxpayers is, well, it’s an idea I suppose. We have an existing service, which is designed to intentionally not be used for obvious commercial reasons.
A Government Accountability Report in April 2022 found that 70 percent of taxpayers were eligible to use an existing free-file program but just 3 percent actually used the service. That program consists of a public-private partnership of tax software companies that offers free services to certain taxpayers outside of the IRS website.
No, seriously, we could devastate China if we allowed skilled immigrants to come contribute here instead of there (or from other places as well), until we do this I do not want to hear anyone talk about having to ‘beat China’ some other way, and I will keep talking about this.
Dan Wang: I’ve chatted with a good number of Chinese undergrads in the US, who almost to a person tell me that their parents are urging them not to return to the mainland.
In related news, we put arbitrary limits on H1B visas, our loss is Canada’s gain.
Patrick McKenzie: Interestingly Canada’s recent H1B program (essentially, offering expedited status to U.S. immigrants with H1B status who are negatively impacted by policy environment) a form of policy competition generally seen between competing internal U.S. jurisdictions.
The broad form of that is “Oh you are making an obviously suboptimal choice due to the configuration of your local polity? We will happily and explicitly take advantage of that then.”
(The sticking point in U.S. re: high-skilled immigration, and I am attempting to phrase this as neutrally as possible, is that both parties want it but have other things they want more, and tied those goals (which oppose each others’) to it for essentially structural reasons.)
For social reasons I’ll elide specific identification of places where e.g. U.S. states robustly and explicitly compete against each other, but it is one of the healthiest parts of the American experiment, particularly since coupled with ~unrestricted right to “exit” internally.
(Indeed, “Pass California’s requirements but California loves you not? Come to Texas! We will offer you an expedited path!” is not merely an explicit competitive strategy of California/Texas, but rather the default regime under the full faith and credit clause of Constitution.)
(This is one reason why there is a robust internal market among the states for e.g. incorporating companies, with different states offering different bundles of goods at different prices to entice you to choose their laws. It would be cyberpunk if not hundreds of years old.)
Daniel Di Martino: Dozens of people told me on Twitter that no H-1B visa holder in the U.S. would ever want to move to Canada. We are the best!! Right?? Guess what? It already reached the 10,000 cap in 48 hours. You don’t understand the U.S. legal immigration system. The economy is great but the law is insane. So much that people are willing to leave. Our loss. Their gain. I expect Canada will open more spots.
Our immigration system is utterly broken, and no one seems to care to fix it. Masoud here documents a student who gets turned down for a visa for the very common reason that the person given authority felt like it.
People mostly agree that high-skill immigration helps the economy, as do foreign students, but they are split on low-skill immigration. It is so crazy that both sides continue to hold high-skill immigration hostage to try and extract other concessions.
Your periodic reminder: AML/KYC direct laws cost $300 billion a year and recover about $3 in illicit funds while gumming up the works in far more costly ways. There are are remarkably large classes of transactions and business were AML/KYC is a gigantic effective tax on activity, often prohibitive. What do we get for this? Almost nothing.
Finch downgrades American sovereign debt from AAA to AA+, the second rating agency to downgrade us. As other have noted while speculating on the motives involved, this is obvious nonsense.
Green Engineer (QTing Matthew Lewis below): You know what else is funny? Essentially no VC will pay for third party technical due diligence. They will place multi$M bets but won’t pay $10K to have actual subject matter expert take a deep dive on the tech first.
Joe Devon: They’ll get free advice from people though. I’ve been asked and gave free advice by VCs in the past, but you’re right, for 7 figure investments, it’s worth paying for advice instead of asking for a favor, because you’re more likely to take the advice.
Matthew Lewis: I have a story about the Bay Area that seems appropriate to share. It’s about venture capitalists. I have a bunch of friends who are VCs. They’re brilliant. Most of the time I have no idea what they are working on because it’s a level of tech I don’t grasp.
But sometimes they say silly things..
[tells story of VCs making technically dumb bets, then story of VCs not understanding the basics of the regulatory barriers they would face, at all.]
These seem like highly different failure modes, if they even are failures. For sufficiently large investments a lot of investigation makes sense, but VCs have learned to focus on certain aspects in general rather than others. In a hits-based business, it can be right to mostly not sweat future regulatory hurdles, and focus on other considerations.
The Fed has a lot of work yet to do in order to get inflation back to target, made harder by people’s inability to realize that money is not tight, we are doomed to never listen to Scott Sumner.
Joe Weisenthal: Poll: Do you believe the following phrase is true? “The Fed still has a lot of work yet to do in order to get inflation back to target?
I agree inflation is looking better for now, with supercore down to 4% and headline at 3% earlier this month, and the news since then good as well. That’s still very much not 2%, let alone proper level targeting. I do not think ‘team transitory’ was vindicated, but their position looks less wrong than it did a few months ago. It’s fine to admit you were wrong and then say now you were not as wrong as you thought when you admitted it, the issue is pretending this actually made your previous claims right.
Ed Elson looks back at his old crypto pitch deck. If you invested based on a pitch like this, I have absolutely zero sympathy for what happened next.
SEC has made their opinion official that every cryptocurrency other than Bitcoin is a security, as they requested Coinbase halt trading in every other coin. Coinbase asked the SEC to justify that position, the SEC refused. This is not a reasonable position. As Matt Levine has pointed out, Dogecoin is quite obviously not a security, so now we are talking price, even if you throw out the court ruling on Ripple.
Usually a technical issue is only a technical issues, but in many contexts such as crypto it is also Bayesian evidence of deeper problems even if it is indeed a technical problem. One must balance the reasons to run and not run on a given bank-like thing.
Robin Hanson disproves EMH by finding $20 bill on the sidewalk. Eliezer reports that 1 in 5 young people surveyed had found one and 1 in 20 had found two.
New non-profit rating agency for stable coins, none of which got an A+. As with all such rating systems, there will be mistakes and one cannot fully rely on them in either direction, but it is a very good heuristic, although not investment advice, to avoid trusting anything that has a bad rating from such an agency. Tether has a D.
I am one of many idiots who did not get in early on the ‘buy thing with superconductor in its name’ trade, although not one of the bigger idiots who buy late and lose money. Happens every time. This is about how false the EMH is.
DeSantis plans to make student loans dischargeable in bankruptcy, with the school responsible for payment if that happens. Model this if it somehow actually happened? My first thought was that this would create interesting incentives for schools, including to generate goodwill. My second thought was how this would greatly impact who gets admitted. Perhaps we should allow both types of loans to compete in the marketplace, and learn about which institutions believe in their students?
Alyssa Vance: I think history shows that four-dimensional-chess reasoning, of the kind “smart person X would never do dumb thing Y, so there must be an elaborate hidden plan”, usually falls to the Occam’s Razor of smart people just doing lots of dumb things
“The presumption of access to these documents is greatly outweighed by the need to avoid their public dissemination at this time,” Bankman-Fried’s lawyer wrote to the court.
The need for SBF to avoid this, that is?
Good News, Everyone
Base Rate Times continues to be an excellent source for probabilistic future news, which also makes it an excellent source for knowing what is the important news about the past. All it needs are better and more robust and expansive prediction markets from which to report, especially real money markets – Manifold and Metaculus are remarkably good but there is no substitute for large amounts of cold hard cash changing hands.
There is still time to sign up for the Roots of Progress Blog-Building Initiative, deadline is August 11. If you want to start a progress-relevant blog and can invest 10-15 hours a week for 8 weeks, this seems like a fantastic opportunity. This is not intended for me, but I did consider applying anyway, and if those involved said ‘no actually you should participate at least somewhat’ then I probably would.
Tetraspace: mugger, meet unthreatenable FDT agent.
WATCHAtlanta police searching for a man who made a failed attempt to rob a salon because no one listened.
If you pump your own gas you’ll have to go with Thursday:
Scott Lincicome reports: Good news for almost everyone: “Any restaurant in the U.S. can now legally use the phrase ‘Taco Tuesday,’ except in New Jersey.”
From the legal filing in question, I will be in the future somewhat less mad about all the Taco Bells wasting valuable New York restaurant slots:
D. “Taco Tuesday” is a common phrase. Nobody should have exclusive rights in a common phrase. Can you imagine if we weren’t allowed to say “what’s up” or “brunch”? Chaos.
E. This Petition is brought because Taco Bell believes that tacos, just like the joy they bring, belong to everyone on any day. Ergo, “Taco Tuesday” should belong to everyone.
F. Taco Bell believes “Taco Tuesday” is critical to everyone’s Tuesday. To deprive anyone of saying “Taco Tuesday” be it Taco Bell or anyone who provides tacos to the world is like depriving the world of sunshine itself.
G. Taco Bell supports everyone’s right to celebrate, and say, “Taco Tuesday,” no matter who they are. How can we tell our fans to Live Más if their favorite taco joints aren’t even allowed to freely say “Taco Tuesday”? Anything else is menos.
H. Taco Bell seeks no damages; it simply seeks reason and common sense.
I. If one of us is not free to celebrate “Taco Tuesday,” then none of us are free to celebrate “Taco Tuesday.” A win for Taco Bell here is a win for all. When tacos win, we all win.
Jeremy Munday, a professor of electrical and computer engineering at the University of California, Davis, who researches clean technology, said this redirection would barely affect space. The sun already emits more than a billion times more heat than the Earth, he said, and this method merely reflects heat already generated by the sun. “It’d be like pouring a cup of regular water into the ocean,” Dr. Munday said.
He calculated that if materials such as Purdue’s ultra-white paint were to coat between 1 percent and 2 percent of the Earth’s surface, slightly more than half the size of the Sahara, the planet would no longer absorb more heat than it was emitting, and global temperatures would stop rising.
Someone’s reaction to this proposal tells you a lot about what it is they actually value.
Also works wonders insulating buildings.
The joys of an industry having good ops that trust and work with each other. It is a miracle that things work as often and as safety as they do.
Report on the Civic Futures Summit and unconference. If you are interested in a topic and expect to encounter people you want to talk to, I have yet to hear even one person regret going to an unconference. The format remains highly underrated. Tyler Cowen encouraged Brits to be more optimistic as South England is still one of the only areas in the world ‘where you can start ideas and get things done.’ That is indeed a huge deal, but time seems like it is running out for one of those ideas to be ‘build a house.’
David Perell: Ordering a coffee at a 5-star hotel so you can work in the lobby is one of the world’s great arbitrages. You can either pay $3 for a run-of the-mill Starbucks with rickety seating and janky Internet, or $6 for a soft leather couch, table service, and supersonic Wi-Fi speeds.
Patrick McKenzie: I was probably 30 or so before I realized that many hotel services are open to people who are not guests, such as their bars and restaurants, which are pricey but very conveniently located and frequently underused.
And so if you are in a central business district and just need somewhere to be for two hours, you can just walk in one, sit down, and order a coffee/lemonade/sandwich/whatever and then stay there a socially appropriate amount of time, which can be considerable.
And depending on your comfort level you can also use a hotel concierge without being a present guest of the hotel, though I would likely not. (Tip generously in that case?)
“For what?” Restaurant recommendations and/or reservations, etc. (Because concierges are frequent flyers they sometimes have the ability to get desirable tables which are otherwise less gettable, including on very short notice.)
Incidentally, there are probably suaver ways to do this, but I always approach hotel concierges and doormen holding a palmed bill and, after the request has progressed naturally, offer it to them with the words “Thank you.” (Some people need this written exactly. Trust me.)
Tipping culture is weird and one learns it by osmosis but there is no concierge who has ever been tipped before that does not know what money presented in their direction means.
How does Patrick McKenzie keep getting all these mountains of paperwork that would presumably drive ordinary people to tears? Either way, be in the bug fixing business.
A “culture eats strategy for breakfast” anecdote: So I have to file a small mountain of paperwork incident to leaving Japan. The ward office is very good at paperwork and has a ticketing system. I had slipped between cracks and, after waiting 15 minutes, unstuck myself.
At 16 minutes, a young lady who (using context cues) is clearly part-time employee of the office and clocked out came up and said translated “Sir, is everything OK? I noticed you in line 20 minutes ago and you haven’t left yet, so I just wanted to make sure you are being helped.”
You can write a million procedure workflows but you’ll never, ever write down the culture that makes a clocked out government employee see someone sitting in a chair and think “Ah we potentially have a bug; I am in the bug fixing business.”
Pause to acknowledge that “20 minutes waiting time is so obviously a noteworthy SLA breach at a government office that you’d ask what circumstances caused it” also noteworthy.
Rob Henderson: 20% of the population is responsible for…
•100% of criminal convictions
•99% of govt benefits received
•84% of cigarettes smoked
•83% of credit card debt
•78% of all injuries
•67% of fast food consumption
•53% of all alcohol consumption
•47% of all sex partners
It seems we still have >20% of people smoking and taking on credit card debt. Stop it, everyone. Having credit card debt should be an act of desperation, not something about half of people carry, which is the only way the 83% number is that small.
The most surprising are the last two. The top 20% only have half of the sex partners and alcohol? I would have assumed both were highly fat-tailed distributions. About 20%-25% of American adults don’t drink at all, many of the rest drink only rarely, and alcoholics often drink quite a lot. The sexual partner distribution seems like it should have quite a lot of 0s and 1s in it. A CDC source says the median number of partners is 4.3 for women and 6.3 for men (which is quite the contrast), whereas 28.3% of men have 15 partners or more as do 12.9% of women. It’s still in theory possible (I think) to get to only 47% there, but it is not easy.
While I Cannot Condone This
Eliezer Yudkowsky lays the smackdown of the righteous on UFO claims. Standard betting line is 150:1 against any of this being shown to be real, the odds should if anything be far longer.
List by Anna Gat of things you learn from travel updates me towards feeling less need to travel. Is that the best we can do? Tyler Cowen instead suggests one reason is that travel makes you a better reader, especially for history, geography, (factual) economics and political science. This seems like an excellent reason for Tyler in particular to travel, since his goal is to read everything about everything. My experience however is that this effect is minimal unless you ‘travel well’ in the sense of immersing yourself deeply, in a way that almost no one considering marginal travel does.
In Medical and Health News
Everyone else is right, Boris Johnson can write. He missed his true calling. In this case, he writes about the promise of Ozempic. Worth reading purely for the phrasing.
FDA Commissioner Robert Califf: I’ve been going around saying that misinformation is the most common cause of death in the United States. There is no way to prove that, but I do believe that it is.
He has also called it the ‘leading cause of meaningful life-years lost.’
What does he mean by this absurdity?
Many don’t use statins, vaccines, or covid-19 therapies. Many choose to smoke cigarettes and eat the wrong food.
This is the branch of philosophy that says that anyone who disagrees with you, or makes decisions of which you do not approve, you are right and it is the people who are wrong. Also that there are no trade-offs or preferences, only that which public health thinks is right and that which stupid ugly Americans do instead that causes them to die.
All the crusades Califf actually lists are petty. He’s at least mostly not wrong, but none of the issues in question kill that many people. Instead Califf is essentially saying that ‘misinformation’ is behind every decision to not eat optimally healthy food, or engage in proper exercise, or not seek proper health care, or to take various other risks.
This is such nonsense. People have other preferences and other things to do. They make choices, sometimes badly, sometimes foolishly, sometimes out of laziness or short time horizons or addictions or feeling like it. They get to do that.
Misinformation is not ignorance, nor is it lack of wisdom, or lack of caring, nor is it having different priorities. Misinformation, to the extent it is a thing, is when people are told things and those things are false.
When one attributes smoking to ‘misinformation,’ or blames it for more deaths than heart disease which kills 23.5% of us, one reveals a failure to understand this distinction. And also one reveals… misinformation.
I find this dilemma oddly relevant. What do you value, exactly?
Will Newsome: it’s pretty upsetting to see someone order a 10 piece mcnugget at McDonald’s when paying one more dollar has gotten you a 20 piece instead for at least two decades years now.
Paper claims to find strong causal relationship where higher income causes improved mental health, also claiming to have corrected for reverse causation and various intermediating effects. The correlational effects are clear enough. The causal ones seem so confounded as to almost be a wrong question, I have no faith that the methods here, while creative and effortful, are sufficient.
Suicide rates among young people are rising in America, but falling in many other countries, most of which also have a lot of iPhones and social media.
Did you know you can use a shot blocker to minimize pain and fear of injections, for kids and also for adults? It appears asking for this annoys techs, so I guess we’ll simply have lots of people avoid their shots, then, instead. Or did you also know that you can request a smaller needle called a butterfly for blood draws that reduces pain? Also:
Payant: When you get your eyes dilated at the eye doctor, I’ve been told they have another eye drop that un-dilates the eyes without having to wait, but you have to ask for it.
Claude-2: This claim is partly true. There are eye drops that can help reverse mydriasis (pupil dilation) more quickly after an eye exam, but they are not always offered proactively. Dilation drops used for eye exams contain medications like tropicamide or phenylephrine that can take 4-6 hours to completely wear off. This can potentially decrease the dilation time to 1-2 hours rather than 4-6 hours. However, the effect is variable person to person.
These drops are not harmful, but some eye doctors avoid using them routinely due to the extra cost and work.
Amanda Askell: I once said to my dentist “those numbing injections last so long every time and make me feel awful” and she was like “oh, we could give you this other kind that only lasts as long as the procedure and won’t make you feel bad”. Apparently that was just… an option the whole time?
No doubt the list goes on.
David Sinclair claims chemically induced reprogramming can reverse cellular aging (direct to paper), and that they are planning human trials soon. Huge if true. As with most such things I will assume not true until I see much stronger evidence.
Sarah Constantin notices paper by Eric Vallabh Minikel and others that existing genetic linkage doubles the chance a drug will test as effective, but most successful drugs lack such a link. Is this an inefficiency? Could we look at genetic correlations with disease to identify more good candidate drugs? Perhaps, also perhaps not. My hypothesis is that many of the ‘obvious’ drug candidates have such a link and often succeed, driving the average up without impacting the marginal case.
Arnold Kling points out that marijuana use tends to be net negative. He suggests fines for public use, and generally a regime similar to alcohol where society signals its disapproval. Both are drugs we cannot in practice ban, and that some people use responsibly in ways that leave them better off, but that many others use in ways that leave them worse off, and that on net I believe do great harm and where we would as a society greatly benefit from reducing usage. On a personal level, I think almost everyone would be better off if they cut out both mostly or better yet entirely, especially alcohol.
There exists a theory, says Scott Alexander with links, called the Social Theory of Disability that ‘individual limitations are not the cause of disability.’ There also exists, he says, a Medical Theory of Disability, which says disability is only caused by disease (or other physical limitations, presumably, I wouldn’t call ‘missing your arms’ a disease but it sure is disabling), and also that no one should ever accommodate anyone on any level and presumably it’s also good to stigmatize them. I would like to think everyone instead believes the obviously correct Interactionist Model where disability comes from a lack of ability, which is a combination of your lack of ability and society’s failure (reasonably or otherwise) to address that. Scott then spends a bunch of words explaining in detail how people decided to use rather strange word choices that result in absurd-sounding claims and occasionally believe their implications, rather than making the case most everyone agrees with, that we should strive to make accommodations for people.
We need better research on the cognitive effects of CO2. There are potentially large effects here, or it might have remarkably little impact on its own and mostly be an indicator of other air quality issues.
Paper reviews most common happiness interventions for scientific validity: Expressing gratitude, time in nature, exercise, talking to strangers and mindfulness.
Michael Plant: As you can see, not much top evidence As they point out, this doesn’t necessarily mean recommendations are wrong. Absence of evidence =/= evidence of absence. But it suggests caution. What did they find?
Practising gratitude. Evidence of temporary increase. No evidence of long term increase.
For socializing. ‘Solid evidence’ you should talk to strangers (more? sometimes? only?)
For mindfulness. This one half-surprised me: limited evidence of benefit, and challenge is disengaging the benefit of mindfulness *itself* with participating in a group activity.
Limited evidence for exercise. Tbh, this one annoyed me a bit, and I have concerns about ‘scientism’, that I’ll come back to
Experiencing nature improves wellbeing but, acc. authors, previous research was not best practice.
My view continues to be that different people are different. All five interventions make it easy to run the experiment and cheaply gather data on whether it works for you. Nature and socializing in particular seem like things that benefits some people a lot and others little or not at all, for me socializing is iffy and nature mostly doesn’t work. Exercise I am confident is helpful overall but the people who expect it to make you happy directly rather than as a secondary effect of being healthy are highly annoying, that does not work for me or many others I now. Mindfulness I strongly suspect is mostly a scam.
My experience says that gratitude does work, and that it has a lasting effect when you feel gratitude in a systematic way, or for things that mattered in your life. It can make a big difference. But while you do have to actually do it to get the benefits, I do think you have to mean it, this can’t be forced or it won’t work. And no doubt there are people for whom it won’t work at all.
Lying About the Lab Leak Hypothesis
We likely will never know whether Covid-19 leaked from a lab.
What we do know is that there were those that also did not know the answer, yet misrepresented their beliefs and suppress the lab leak hypothesis, as the result of a collaborative process that did so in part for political reasons and in part to protect gain of function research. If people knew they might have caused the pandemic, how would they be funded to go cause another pandemic?
I have not independently gone into the details, because I do not find them so important, and do not think it makes much difference whether this was deliberate explicit coordination or implicit forces pushing outcomes in a particular direction. That failure to care itself reflects a problem. To me, the biggest scandal of all is our collective failure to take this issue seriously as a scandal. The scandal is that at this point such behaviors are expected, considered standard procedure, do not represent news, are not something I have even bothered diving into so deeply, do not seems likely to make any difference.
Nate Silver is still all-in on this issue. He fleshed out his full argument in this post.
Nate Silver: Here’s the scandal. In March 2020, a group of scientists — in particular, Kristian G. Andersen the of The Scripps Research Institute, Andrew Rambaut of The University of Edinburgh, Edward C. Holmes of the University of Sydney, and Robert F. Garry of Tulane University — published a paper in Nature Medicine that seemingly contradicted their true beliefs about COVID’s origins and which they knew to be misleading. The paper, “The proximal origin of SARS-CoV-2”, has been cited more than 5,900 times and was enormously influential in shaping the debate about the origins of COVID-19.
The messages show that the authors were highly uncertain about COVID’s origins — and if anything, they leaned more toward a lab leak than a spillover from an animal source. But none of that was expressed in the “Proximal Origin” paper, which instead said that “we do not believe that any type of laboratory-based scenario is plausible”.
On Twitter: Let’s start by 1) retracting the “Proximal Origins” paper; 2) having scientists like you (i.e. people sympathetic to the natural origins case) call out Andersen et al for their gross misconduct. Then we might have the semblance of an honest discussion.
I’m not a big petition guy but if Nature isn’t ready to retract this paper on their own that’s a big L for their credibility and about as clear a sign as you can get that they’re elevating politics above science. This paper was not the result of any sort of scientific process.
Stephen Hudson: For what possible scientific reason could you possibly demand that paper be retracted?
Nate Silver: That we have a record of their internal communications, and that they’re extremely explicit that they think a lab leak is plausible if not likely, and also extremely explicit that they’re going to say the opposite in the paper for political reasons, and also lie to reporters.
Other Thread: This is a huge scandal. Scientists like @K_G_Andersen believed a lab leak was extremely plausible, if not likely, they concocted a plan to deceive the public about it, and they’ve been caught red-handed. There’s not really any ambiguity here. They are unethical as it gets.
Third Thread: I mean it’s some of the most incontrovertible evidence one can imagine. To be fair I always thought he was full of it. But the leaked messages are in the 99.9th percentile of what everyone would have agreed was incontrovertible if they’d negotiated evidentiary standards ex ante.
Francois Balloux: Though, surprisingly maybe, and at the risk of annoying of you, I don’t believe either there’s a real textbook conspiracy here. To me It’s primarily misaligned incentives, and personal ambitions, and a bit of intellectual weakness and laziness that led to this total cock-up.
The first preprinted version was actually pretty balanced, and it was a far more interesting paper. It got rejected by Nature largely (I guess) because it didn’t rule out a lab leak scenario.
The follow-up Nature Medicine version was much shorter and more assertive. The language got further strengthened during the review process at Nature Medicine (which looks rather unimpressive, intellectually speaking).
Scientists are human, with all virtues and vices shared by all of us. But, they’re hyper-competitive, and if the only obstacle between a scientist and a high-impact paper is some subtle turn of phrase, then the scientist will likely cave in, and cannot be blamed for it.
Essentially everyone wanted SARS-CoV-2 to have a ‘natural’ zoonotic origin. This bias turned into a strong prior, and in the absence of any real evidence overriding it, the prior became the posterior. And ho and behold, we got served some unsubstantiated claims.
I believe the the Proximal Origins authors screwed up and bottled it, but not in the way – and not for the reasons – most of their detractors tend to view it.
Doing Science To It: Marinara Edition
Terese: Conclusions first because this is general stuff that applies even if you don’t care about which particular brands I tested. I definitely learned more from this experience than just which sauces I did and didn’t like, so I wanted to share that here.
BRANDS: I went into this expecting more of the smaller brands to taste better. Small-batch is usually a positive descriptor, right? As well as artisanal, local, etc… but all of my least favorite sauces were smaller brands.
One of my favorite sauces was a smaller brand too! But I’d say that in addition to being higher variance, the smaller brands generally scored worse with me. Perhaps that big brand money spent on product testers really does matter.
This matches my general experience more generally. The big brands, the standard products, are popular for a reason. You can sometimes do better by moving up the cost curve and finding a bespoke alternative that fits exactly your preferences or is that good, but most of the small brand alternatives will be worse. One of the glories of America is that many of the standards – such as Heinz Ketchup or Oreo Cookies, or famously Coke – are flat out the best version of that product known to exist and the rich man cannot do better.
PRICE: Note that the prices are listed as they appear at my local grocery store, which is on the more expensive side. With that said, there was almost no correlation between price per ounce and my scoring of a sauce. I was very surprised by this!
However this may have been muddled by my trying a lot of small brands, which one might expect to be generally more expensive and have higher variance in quality. I supposed I’d expect there to be more of a correlation when only testing national brands.
I would expect the correlation to not be as strong as most people would expect, but almost none surprises me as well. There are clear price-quality tradeoffs one makes when producing sauce. Among the bigger brands, I would expect price to indeed predict quality quite a bit.
TEXTURE: Before this project I didn’t realize how much variance in texture I’d encounter. Some of the sauces were basically just a smooth puree, and some of them were like chunks sitting in juice. And it turned out that texture really mattered for my enjoyment of a sauce.
TEMPERATURE: I already knew that things can taste different at different temperatures, but this was interesting to pay close attention to. I found that at higher temperatures, tomato flavor was pretty robust. Garlic and herb flavors tended to come through more as a sauce cooled.
I hadn’t thought about temperature. Texture issues are clear.
Her ratings, all with detailed descriptions in the thread:
8/10: Newman’s Own
7/10: Barilla, Michael’s of Brooklyn
6/10: Classico, San Merican
5/10: Mantova, Bove’s
4/10: Victoria, Prego
3/10: Cucina Antica, Pizza Girl, Di Bari
1/10: Billy’s Italian Market
Here’s her note on Rao’s:
It was one of the more difficult sauces to describe, as the flavors were very well blended. It was hard to pick out notes of particular things like garlic and herbs because everything was very well melded. I don’t mean to say that the flavor didn’t have complexity, because it did, but just that I couldn’t easily go, “ah, yes, that’s a bit of garlic flavor there.” So instead while taking notes I resorted to more abstract language.
“Smell of tomatoes, but with an accent.” “Distinguished, sophisticated, vintage.” The flavors were interesting but came together smoothly, with well-balanced acidity and a bit of spice. The texture was good, pulpy with some small chunks of tomato.
What is missing is the sauce I actually use, which is Lidia’s, and also Trader Giotto’s Marinara, and one commentor mentions il Mulino. America, what a country.
To the extent I have tried the sauces, which is not that many of them at this point, I can confirm these ratings make sense. To the extent that I haven’t, based on my general heuristics, they continue to make perfect sense.
While I Cannot Condone This
Philippe Lemoine: I thought that not everybody was equally capable of understanding mathematics, but a bunch of Fields medalists assured me that it was just a matter of pedagogy, so I guess I was wrong.
Misha: You know this gives me an idea for a reality tv show.
GDsimms: “The Tutors” Season 1: We took 4 Fields medalists and assigned them to teach algebraic topology to 4 randomly selected 25 year olds with only high school diplomas
Aaron Sibarium: A lawyer in Big Law told me today that the firm’s corporate clients are all reaching out in the wake of Students For Fair Admissions to ask if their diversity policies are legal. In most cases, even liberal attorneys are telling them no.
Patrick McKenzie: I have had a really surprising number of conversations over the years with people who have hiring authority in the United States and believe racial discrimination in employment is legal if locally popular.
“That… is not how that law works.”
But people were saying in Slack that…
“That’s discoverable. You should call your employment lawyers, immediately.”
IANAL and none of this is legal advice, but as I understand it, this is at least kind of how all laws work. If everyone is doing it and no one is suing or bringing enforcement actions over it, is it illegal? There is a reason Jewish law says no, a law that is not enforced is null and void. American law is not something that can, strictly speaking, be entirely adhered to, and if your action is popular locally such that a jury will likely not find you liable or guilty, and the government is doing similar things and unlikely to sanction you, I do not see the urgency in calling one’s employment lawyers.
Steve Stewart-Williams: Sex differences in work preferences, life values, and personal views among gifted men and women. <0: endorsed more by men than women >0: endorsed more by women than men (source).
May want to click through to better see the graphs. This was definitely not the best possible way to graphically represent this data.
A clear preference on the one side for compensation, on the other for reduced hours.
Note that having children leans male here, and overall there is much more balance.
Another stark contrast. One is free to ask what outcomes such preferences would produce.
What kind of incompetent person would have to rate limit something like Twitter?
Mike Solana: damn they copied EVERYTHING huh
Mosseri (on Threads): Spam attacks have picked up so we’re going to have to get tighter on things like rate limits, which is going to mean more unintentionally limiting active people (false positives). If you get caught up [in] those protections let us know.
Disclose.tv: JUST IN – Zuckerberg’s Twitter clone “Threads” has seen a 70% decline in daily active users since its July 7 peak — Forbes
Tyler Altman asks why we no longer carve statues into building facades, after doing it for thousands of years. Some good partial answers:
Alesia Charles: I want to say neo-Puritanism, but I don’t have the time or energy to explain what I mean right now.
Gotriq: Because a global society has no shared icons to display.
Sarah Constantin: “why is architecture not decorative any more” is a story about the early 20th century “why don’t institutions like train stations and banks try to look overtly inspiring and glorious” is a different story, probably about the 1960s
Rob Miles: My guess is economics. Everything got cheaper but the kind of skilled labour needed to do this got much more expensive. The cost of covering a building with custom statues, as a percentage of total building cost, became massively more than it used to be.
I would combine them, especially the economics response from Miles, extending that to include standardization and a system of regulations and permissions. You used to be able to go ahead and do any old awesome thing without asking permission. Now you would need to get all sorts of approvals, risking destroying the entire timeline or approval of the project. That simply is not worth the trouble.
Except that it is. We need to go back. We need to build awesome statues again. We need uniqueness, character, art, awe, and we can damn well afford it. The problem goes well beyond the statues – before AI went crazy I was planning an extensive series inspired by Christopher Alexander’s A Pattern Language.
This is not a lost cause I intend to spend my ammunition on given the opportunity costs, but it must be said.
Almost all the gains from A/B testing, and new ideas, come from a handful of very good ones. At Bing, the top 2% of ideas resulted in 75% of gains. Oh no, inequality! The takeaway is that if your gains mostly come from very big wins, you do not need to do extensive testing of new ideas – if it isn’t naked eye obviously (ten times) better then you do not lose much ignoring it.
How is this for irony? Of course the Paperwork Reduction Act causes paperwork.
Samuel Hammond: Wouldn’t it be awesome if the US government did A/B testing too? Unfortunately that could constitute a survey of the American public and require a lengthy approval via the Paperwork Reduction Act.
Men repay their student debt a lot faster than women, who on average barely repay it at all, whereas men were repaying it during the pause with no interest payments like a bunch of chumps.
The loans given out in 2009 were typically in the 5%-8% interest range. This amount of failure to repay during the late 2010s reflects some combination of extreme illiquidity, extreme insolvency and extreme economic illiteracy. Whereas for loans from 2020, indefinitely paying the legal minimum seems highly reasonable. I’d love to see the crosstabs on this in terms of income and family size and marital status and economic literacy.
George Mack offers a thread of 25 charts illustrating what he says are neglected trends. Here are some of them, noting that they are important but not obviously neglected.
25. The Importance Of Exercise The mortality gap between a smoker and non-smoker: 40%. The mortality gap between the bottom 25% of exercise fitness and top 2%: 400% The best marketing minds of their generation created anti-smoking ads — where are the pro exercise ads?
I would caution there on correlation versus causation. If you are healthy you are more likely to exercise. I am sure there is a big and real effect, but it isn’t 400%.
I’d also add this one:
Nabeel S. Quershi on puzzles. He notes that the better you are at chess the more time you spend falsifying moves versus generating them. There is some amount of wisdom to get from that, but also generating potential moves in chess is mostly the easy part, then things get exponential. In his third section ‘conviction that there is an answer’ he notes the following puzzle:
To me (and I’d assume anyone else who did math competitions) this is trivial, the sequence has limited memory so it will repeat in a loop, all you have to do is find one containing a 6, and I did not do well on the USAMO or Putnam. He solves this by not giving up based on knowing there is a solution, and concludes that scientists and founders have faith that there is an answer out there.
The next level is to be able to do the search where you look for the path to victory with the same determination as if you knew it existed, without assuming that the path exists, and indeed instead also trying to prove that no such path exists, so that you figure out why you can’t prove that yet. Or so that you do prove it, and can stop. My biggest failure as a founder was, quite literally, not knowing when to quit.
In section five he discusses science brain versus founder or CEO brain, where scientists need to be driven by curiosity to find the value, whereas CEOs and founders need to get their heads in whatever game is most valuable right now. I wonder if part of what I have, for both better and worse, is a strange hybrid of both.
Also curious: Despite my being (last time I checked) only a B-player at best, in the  footnote, which was meant to be an example of complacency, my first candidate move was the right one and instinctively looked like it would work, whereas the move he says is ‘automatic’ didn’t occur to me since I was confident I’d found a forced win.
The second example is even weirder to me, even if you discount the fact that the question didn’t say ‘integers.’ The moment you write out a 1 you have to know that’s wrong. That’s not complacency, that’s a complete failure to do sanity or error checks.
Sports Go Sports
Winning turnover battles wins football games, but a lot of that is that teams that are behind take more risks (and are also worse) and this causes more turnovers. Thus, being +1 in turnovers means you win 71% of NCAA football games, rising to 80% for +2 and 88% for +3, but that greatly overstates their causal effects. Lots of good work here, especially on expected turnovers for the game circumstances of various quarterbacks, but alas Mike Bursik does not run the trickier calculation of how often an extra turnover results in victory, or even better that measure as a function of the size of the game’s point spread. Further research is needed.
Why is Saudi Arabia spending billions on Western sports? The natural explanation could be that there are not so many ways to spend the kind of money Saudi Arabia makes on oil in order to make long term investments. By drawing in sports, they get positive attention, make people more excited to work and live there, provide compliments for other investments, where they are capacity limited. Or that they are using it to buy connections and networking and create investment ties, for similar reasons.
A rival explanation is this is ‘sportswashing’ designed to distract us from their failure to uphold proper Western values, or to buy indulgences for it. This does not make sense to me. Sports instead is a giant beacon, drawing more attention, making people suspect they are up to something. It didn’t work for Qatar and it won’t work for Saudi Arabia. No one is going to forgive them for anything because they have good soccer.
An alternative explanation that I have not heard, but that seems obvious, is what if this has nothing to do with the national interests of Saudi Arabia? This is a kingdom, not a democracy, ruled by one family. They want sports because they want sports, and by framing it as for the country, they justify spending billions for the same reason every Western billionaire buys sports teams. Because it’s fun.
Tyler Cowen looks at NBA contracts, expecting prices to go up as Saudis and others compete for our talent and basketball gets more valuable. I found the discussion weird, because it was not structured around the salary cap. Overpaying for a player can thus cripple your team. The Saudis could potentially make existing deals look good if they substantially drain talent out of the NBA with a new league, although that could also lower revenue and thus the salary cap, and would likely take some years to get set up.
The Pac-12 has now completely collapsed. Colorado fled to the Big 12, then Washington and Oregon went to the Big 10 (which it seems is keeping that name despite having 18 teams) which triggered Arizona, Arizona State and Utah to also flee for the Big 12. As Nate Silver puts it in the link above, there is a complex optimization problem here, and no obvious solution to it. Individual teams are following their incentives, which is leading towards a world of two super-conferences, the SEC and Big 10, and then everyone else. Or it is leading to a second reorganization, where the best teams cut out the underperformers and form some sort of super league.
Rece Davis (ESPN): I’ve said, “what happens when the next iteration of this is ‘why are we sharing money with programs that don’t contribute as much and aren’t as attractive’”. The golden ticket a few hold now might disappear.
Stewart Mandel: I wrote this 4 years ago, before all the recent major moves. All roads are clearly headed this way, it’s just a matter of when/how many will be included. Also: I could see it being more AFC/NFC with ESPN and Fox.
If it were strictly this with no way in for other teams that would be unacceptable. However, if you included some form of relegation and some way for wild cards to enter the playoff, kept rivalry games and ensured everyone had a free slot to schedule a non-premier team every season, then it seems fine? Say, the top two division champions outside the league get to play wild card games against the bottom two Premier teams to steal their slots (which is great for drama) and if they succeed and win their challenge game (or get so far along that they don’t have to play one, and instead the two bottom finishers play to see who gets relegated, the problem there is you have to factor in TV ratings somehow without being obvious that you are doing so), then they get promoted to the Premier league.
Conferences seem forced in the long run to treat all their members mostly or entirely equally, temporary junior status of Washington and Oregon notwithstanding, so whoever is a bigger draw will inevitably defect to bigger and better things, geography and rivalry be damned. There is no stable equilibrium.
There also is no way to expel a school from a conference. Which means that a conference with a number of underperforming schools, like the Pac-12, loses its best schools, then either downgrades a tier like the Big 12 did or it falls apart entirely.
I suppose the long term plan is that the SEC and Big 10 expand more, all the best teams play each other for a while, then after a while there is enough dead weight in both that the top teams start thinking about breaking away, and the cycle continues? We get a kind of slow-motion relegation, except chaotic and without giving the games drama and breaking things apart in ways that don’t make sense.
It is a shame. Things were better for everyone – fans, players, schools, everyone – when we had conferences with geographical and cultural identities that mattered. What to do about this?
My presumption is that for now there is nothing that can be done. The money is what it is and money talks.
But the money could be different. That will be especially true in the 16 team playoff era. At that point, there should be little need to game the system on that front, your scheduling will probably mostly be a wash on the new bubble or reward tougher schedules, and no one capable of going deep will be that likely to miss.
What everyone will want are the best individual games, as often as possible. So what system would we design to do that? Well, what do we want?
We want rivalry games, especially those with rich histories.
We want top matchups, top 25 vs. top 25 or even better, with fewer cupcakes.
We want things to make geographical sense when possible.
We want the games to matter, tell stories, have stakes.
We want to maximize TV ratings.
That all is good for pure fans, and also makes more money. And it has to be something that is incentive compatible, given the money. That’s what make it so hard.
I’ve gone over this many times over the years. Thought about options. Relegation. Swiss pairings, either within or between years. Legislation. Different revenue schemes. I don’t have a good answer. Not one that might actually happen.
If I was Czar, and given the power to go wild, here is what I might do to your city.
Restore geographical conferences with 12 teams per conference, with a 1st tier (ACC, SEC, Big 10 and Pac 12) and 2nd tier conference (American, Big 12, Mac and Mountain West) in each of the four classic regions, then other conferences are at-large. Any teams that are not in contiguous states with teams in the conference, or in the region itself, are expelled, after that initial downward placement where needed is by average strength of record over the last four years. Any teams not in either regional conference go into other conferences, which are free to organize as they see fit.
Each conference divides into two 6 team divisions, has one championship game between the top teams from each division. The bottom teams from each division play the safety game.
Each year there are two relegation matches in each region, played as special Bowls named after the conference the teams are playing for. The conference champion of the 2nd tier conference plays the loser of the conference safety game from the 1st tier conference. The loser of the safety game of the 2nd tier conference plays the highest strength of record conference champion from the region. They play for the conference spot, and trade their old one if they win.
Joint revenue from non-TV sources is split evenly by each conference. Revenue from TV and streaming contracts is half split evenly, and half given out proportional to combined TV ratings of your games, including bowl games and playoff games.
Players can get paid via one four year contract, with no limits or caps, and can supplement via NIL. Any player whose eligibility is used up automatically gets a full scholarship for that year if they don’t already have one, and anyone who is paid a dollar gets a full ride until graduation. Players can use the transfer portal, but they can’t use it to get a pay bump and there is some sort of financial friction introduced (a transfer fee, or something else) because otherwise everyone is a free agent and the teams lose their identities. Teams that have good revenue bases but that have been underperforming would get more revenue from TV, so they could use that to pay players better and catch up. If they can’t, that’s on them.
Teams in first divisions play 12 regular season games, with 8 conference games (all 5 in their division, 1 Swiss-style across divisions based on previous year’s records to open the season, 2 across Swiss-style played after 4 of the 5 games in-division), 1 rivalry game (if this is a conference game, it counts as that, and the team gets another Worthy Opponent), 1 Swiss-style game against another conference, 1 free non-conference game against a Worthy Opponent of their choice that they schedule (either a team that has been Top 25 in the last 2 years, or was in a top conference in the last 4), and 1 Challenge Game against a second division team. Teams in second divisions play up instead of down, and are allowed to schedule anyone as their Worthy Opponent.
Conference championship week now includes championship safety matches as well. After championship weekend, an eight team playoff is formed, consisting of the four first division champions, the top two ranked second (or lower) division champions, and two highest ranked wild cards, as chosen by the committee. The four division one champions get the top four seeds and host the first round, committee ranks 1-4 and ranks 5-8.
The quarterfinals take place the week after the conference championships. Bowls can give a bid collectively to both teams, and the higher seed decides which one to accept – whoever loses the game goes to the bowl accepted, let them have fun.
The semifinals are the Rose Bowl and the Sugar Bowl. The Rose Bowl is Pac 12 vs. Big 10, or if one of them didn’t make it at all, whoever beat the champion. The Sugar Bowl is ACC vs. SEC, or their replacements, same deal. They take place on New Years Day, not New Years Eve. New Years Eve is reserved for the relegation bowls, although others are free to schedule against them if they dare. The finals is a week later and takes place at a rotation of the other former New Year’s Six bowls.
Gamers Gonna Game Game Game Game Game
Game of the month was a new Tier 2 roguelike my friend Alan Comer introduced me to called Brotato. It’s kind of a cross of Vampire Survivors and Slay the Spire. You get 20 waves of enemies, a hero with six auto-firing weapons and lots of helpful items, and tons of stats those items and your leveling bonuses move up and down. Here is a very Jorbs item tier list from Jorbs for Brotato, which you should not look at until after you’ve formed your own opinions. I continue to be a strong fan of ‘play games as blind as possible.’
I’ve also been playing Siralim 3, then I found out there’s a fourth one called Siralim Ultimate but I’m going to at least finish the story on this one anyway. It is a monster catching and breeding while grinding randomly generated levels game with 6-on-6 battles. I love the sense of humor and I’ve been enjoying the steady progress and ability to chill, main story is something like 25-30 hours long. My guess is that the best version of this is a Tier 3 game, good for those who want this experience but not to general tastes.
I put down Diablo 4, and with time to reflect on that experience, I’d say this was a Tier 3 game, while being modestly better than Diablo 3 which I would also have at Tier 3, whereas Diablo 2 was Tier 2 and remains the best game in the series. The best and only Tier 1 ARPG remains Path of Exile and it is not close. If you want a rather uninspired and completely generic but exquisitely crafted, relaxing and less demanding ARPG experience, Diablo 4 will give you exactly that.
Both Final Fantasy 16 and Baldor’s Gate 3 are in my ‘play eventually’ queue. I’ve heard nothing but great things about Baldor’s Gate 3, but I want to wait for some upgrades and patches and for when I have the right mindset, and maybe finally play the first two first. I’ll eventually get to FFXVI as well but I am in no rush there either, and am going to continue replaying the old ones first if I continue enjoying that, although FFV is proving to be a grind once again.
I’m going to keep posting gorgeous Premodern deck pictures. This is arty with Mr. T. Huge.
Sam Black has thoughts on the post-Land Tax era of Premodern.
Here’s good old Phyrexian Dreadnaught that won a NYC monthly.
World of Warcraft Classic introduces hardcore characters. Seems remarkably well thought out in its implementation detais, and puts the focus back on the part of the game that is most fun, which is leveling up a new character, while making every moment meaningful via threat of death. I won’t be playing, but I love the idea. It is telling that this applies to WoW Classic and not to modern WoW, which has destroyed the leveling experience so that players can ‘get on with it’ faster.
Jacob Wilson: Tried my best to fill out the MC Barcelona survey.
Or, alternatively, if it’s more than that, you’d prefer not to answer.
Vic Jenny notes one perversity of the free-to-play model, where the goal of the game is to keep playing it.
Vic Jenny: The “goal” of arena for me is to play for free and stack up gems while rank up. A subscription model basically takes this goal away, that would drive me to away more than the subscription money would.
Right now Arena offers you essentially four options.
Play a lot of drafts with or without contructed and clear daily quests and win enough that you can break even or lose small on currency while building your collection, keep up with set releases, play for free.
Clear daily quests via playing constructed, don’t draft much, play for free but stay non-competitive or fringe because you don’t keep up with set releases.
Don’t do either of these things, draft when you want, pay if you need to.
Don’t do either of these things, play constructed by paying a metric ton.
Invoking options one also means letting the release schedule dictate what you play when. As a set is first released, that’s when constructed is most fun, yet you need to stick to limited for a while. Otherwise you lose on both ends, your limited pays out less usefully and you need to pay up front.
When I played Arena, I was good enough that my lifetime spend was only $50, which was for the Eldraine package when I saw how nuts the set was and I didn’t want to risk being stuck drafting it forever. I would have happily paid more, but like most free-to-play games, the game mostly does not give you a reasonable way to pay a reasonable price.
The issue with a subscription is you risk being stuck, where you no longer build your collection, so you then have to keep subscribing forever to play constructed. It also would draw many whales away from very large fees, so it’s likely a nonstarter, that’s the whole point of not giving you a reasonable way to pay. And also if you let it be month-to-month, a lot of players would come back for set releases, then quit again, over and over, which is a good way to lose those players when they decide not to come back one time.
Still, it is sad that Magic is so all-or-nothing. Now that I have quit for a few years, if I wanted to come back and not spend tons of either (A) time or (B) money, I couldn’t play most constructed competitively. I wish there was a solution, but the economics seems to be saying there is not.
My position is I would happily pay quite a lot per month for full access if I (1) wanted to be playing Magic again and (2) there were goals to strive for, a strong outer loop. Right now, if you exclude currency grinding, all that’s left are the Arena Opens (which are fine but kind of mid and uninspired) and the ladder grind. I need another reason, or another mode, at that point.000The “goal” of arena for me is to play for free and stack up gems while rank up. A subscription model basically takes this goal away, that would drive me to away more than the subscription money would.
Jorbs thread on the infinites in Slay the Spire. I think Jorbs is right that Slay the Spire is now ‘too easy’ for those with a deep mastery of its systems. For the rest of us, not so much. I very much play Act 1 in order to win Act 4, but not in a ‘not afraid I will suddenly die in Act 1 or Act 2’ kind of way, or especially in a ‘I can take Acrobatics on floor 1 and not be terrified’ kind of way. I do go infinite sometimes and win that way, but there is a lot more to winning at Ascension 20 than that, you need to know how to get there without dying first. The point that games should quickly end once you get an infinite is very on point, as several other deckbuilders have hammered into me the hard way.
I think what Slay the Spire needs, if this was still possible, is… Ascension 25, where perfect (limited information) play is only ~50% win rates at best, rather than 80%+, and also probably a few nerfs to some of the infinite-related Watcher cards balanced by strengthening a few others that don’t help you go infinite. The new restrictions would be designed to kneecap anything that would take forever or shortcut the game too much. You could even have the 25 restriction be something that outright kills infinites, such as ‘When you play a card for the (fifth?) time in a battle, it exhausts.’ Perhaps the 23 could be ‘limited buff stacks’ and 24 could be ‘limited debuff stacks.’
Cate Hall at the World Series of Poker Main Event asks: Why is everyone so grumpy playing poker? You’d think someone forced them to play.
I’ve always said if you can’t have fun playing Magic, don’t play. The same applies to poker, unless you have an hourly rate that justifies a different strategy. Some future year I will go to the World Series of Poker to play the main event, and I will presume I am lighting $10,000 on fire, and I will be 100% fine with that and do my best to have a total blast and play as if only first place matter. Otherwise, I realize there is some +EV in the main event in particular because of the extra value of doing well, and that might even exceed the taxes involved, but if you are sitting there miserable do yourself a favor and do something else next time.
World Series of Poker was a lot of fun to watch either way, a fine way to relax, although I definitely worry it doesn’t contain efficient training data. Also new agony of defeat just dropped, great run Nate Silver (clip here). Pretty good way to go, absolutely nothing you can do.
Here’s Part I of his report on the tournament. Due to its length, being able to keep your B-game reliable for weeks on end is key, as is exploiting weak players. My worry with lighting the $10,000 on fire is not that I couldn’t learn relatively quickly to play reasonable poker, it is a combination of my expectation that I would still have tells that the better players would figure out over many days, and more than that simply that I wouldn’t be able to play for 10 hours, well into the night Vegas time, for two weeks straight, I’m too old. It’s a young person’s game.
87% of all games released prior to 2010 not available for commercial purchase, unless you buy used I suppose, or of course if you use a technically illegal emulator or other source. It would be better if this was more like 75%, there are many good games from the past but far more not good ones. I endorse the general consensus that it is on game companies to give you a reasonable way to buy their old content, or else they are giving you the green light to get it another way. The law should reflect that, after some reasonable period of time software would become official abandonware.
I like watching people play games but this flat out boggles me: “How to Make Money by Losing $300,000 a Year on Slot Machines (WSJ)” via people watching your livestream.
Scott Lincicome: In my day, only politicians and cronies could monetize failure. Now, the darn internet lets ANYONE do it. Sad!
Steve Lieberman: He’s fun to watch, no swearing, very inclusive, and welcoming. Also very importantly shows loses. He shows the bad sides like going hundreds of spins before a bonus. I haven’t won more watching his vids but I’ve lost less and play longer.
Yeah, I don’t know what to say to any of that except token-style ‘I’m out.’
The Lighter Side
Aleph: An epiphany I had earlier today: Most people know nothing about the past, so for them movies like Oppenheimer and Napoleon don’t have known endings “So after the battle of Waterloo-” “Shut up man, spoiler warning please”
Andrew Rettek: This happens. A teacher of mine in high school got mad at a student for spoiling Titanic, and some friends in college complained about spoilers for 300.
It is legit. If they do not know, let them have this one.
Luke Stein: Ice cream store manager explained that after they labeled their waffle cones gluten-free (which they always were), many customers stopped ordering them so now they split them into two stacks, one labeled “waffle cones” and another labeled “gluten-free waffle cones”
Oren: If I see a “gluten free” label on anything, I’m gonna assume that it’s a modified (probably not as tasty) version of the original product. Perfectly reasonable.
Luke Stein: Yup.
Whenever possible, it is wise to avoid anything that is labeled ‘free’ of anything, or of any hint that something is an imitation of the form and function of a different thing. Or that tells you something you didn’t think you had to confirm – if you tell me the fish is fresh, or the bagels are baked fresh daily, that is better than the alternative but it is very much not a good sign.
An ongoing question to find the restaurant with the greatest number of brothers. The over comes in.
Brianna Ashby: You’re telling me I’m supposed to just move on with my day after this?
Odd that we haven’t been able to create better filters than this.
If you missed the clip of the woman leaving the airplane, watch (30sec).
As Misha reminds us, it was foretold. Also it had already happened, of course.
Anomalyuk: My first teachers didn’t believe I could read at 4 when my mother told them, it was only after a few months of school when they caught me reading the letters to parents they handed out that they found out.
Misha: I actually got in trouble in kindergarten because I could read and write Russian and my teacher didn’t believe that was real writing
Emmett Shear:this is truly incredible. My friend’s kid is four and he can sight-read. It’s not ambiguous at all. I mean, how does someone achieve this level of wrong?
[and so on, several more such cases.]
No Context Brits: Never deleting this app.
Alec Stapp: How is this real? (I assure you that it is.)
The budget sent to Evers increased the revenue limit for school districts by $325 per pupil for “the 2023-24 school year and the 2024-25 school year.” Using his veto pen, Evers turned that into: “for 2023-2425.”
He did so by cutting out “the,” “24,” “school year and the 20,” a “-” and the second “school year.”
This is nothing. It used to get so much worse…
Cremieux: Legislators were displeased. In 2006, he did this again to raise local government levy limits from 2 to 3.86%, again, against the wishes of the legislature. They were, again, displeased.
So in 2008, the Wisconsin Constitution was changed to prohibit “crossing out words and numbers to create a new sentence from two or more sentences”. The governor didn’t seem to care because this left him with considerable veto power.
Namely, he could “cross out words within a sentence to change its meaning, remove individual digits to create new numbers, or delete entire sentences from paragraphs.” Governor Evers just used this power to extend public school funding to 2425 instead of 2024-2025.
Richard Nixon: We put a lot of work into writing and analysis and so forth. When this happened we chose to merely show you this exchange next to the football picture. It is the most successful thing we’ve ever done. Hundreds of thousands. Some things just cut to rhe quick.
Right Wing Cope: Marco Rubio cannot read.
One does need to be careful with such calculations. Much of what we buy is optionality and slack, and reliability in moments of maximum leverage. Other people need to engage in extended deep work with increasing returns to scale. Thinking about people’s hourly rate tells you little about the value of their marginal hour.
Tyler points to this point:
“No one at Shopify would expense a $500 dinner,” Nejatian said in an interview. “But lots and lots of people spend way more than that in meetings without ever making a decision.
A central reason you don’t get to expense a $500 dinner is that the incentives involved are dangerous and destructive. A meeting does not create a similar principle-agent problem. While that does not fully invalidate the point, most things in business that are worth doing, and most things not in business too, are not close calculations, and are instead overwhelmingly valuable. There is tons of surplus. Often it is wise to obfuscate the ‘cost’ of such things, since the benefits are similarly elusive.
Aella: sometimes ppl accuse me of iamverysmartism, comments like “she must carry around a thesaurus to try to sound smart” no dude wtf you must not know many nerds. this is what happens when you grow up reading books bro. imma start accusing you of iamverydumbism see how you like that.
Troya Müzesi:A message just sent to our e-mail address says: “You can get your surprise gift by clicking the link below” As you guessed, of course, we did not open that link. We last made this mistake 3,200 years ago. #TroyaMüzesi #Museumoftroy #Troyamuseum #trojanwar #trojanhorse #trojan
Alex Cohen: My wife just texted me this, I’m sleeping on the couch tonight.