Monthly Roundup #12: November 2023

Things on the AI front have been rather hectic. That does not mean other things stopped happening. Quite the opposite. So here we are again.

PSA: Crumbl Cookies, while delicious, have rather a lot of calories, 720 in the basic cookie. Yes, they display this as 180, by deciding serving size is a quarter of a cookie. This display strategy is pretty outrageous and should not be legal, we need to do something about unrealistic serving sizes – at minimum, require that the serving size be displayed in same size font as the calorie count.

It really is weird that we don’t think about Russia, and especially the USSR, more in terms of the universal alcoholism.

Reminder that there really is an architecture conspiracy to make life worse. Peter Eisnman straight out says: “Anxiety and alienation is the modern condition. The point of architecture is to constantly remind you of it. I feel anxious. I want buildings to make you anxious!” There is also, in response to being asked if perhaps it would be better for there to be less anxiety not more: “And so the role of art or architecture might be just to remind people that everything wasn’t all right. And I’m not convinced, by the way, that it is all right.”

My wife is exploring anime recently. It has its charms, but the rate of ‘this thing multiple friends recommended is actually pretty boring’ is remarkably high. New generations have other concerns.

Avary: growing up is realizing a lot of the anime you watched and loved as a kid is actually problematic af so you’re stuck between exposing yourself with defending it or hating on it with everyone else…

Tom Kitten: Zoomers basically exist in a technological panopticon of continual anxiety about conforming to the latest updates in moral standards & moral panics, but they’re told the alternative is Nazism so many just try to adopt a “haha isn’t it weird” attitude about it.

Can I suggest a third way? You don’t have to say anything. If you love an anime and others are calling it problematic, you don’t have to defend it and you don’t have to condemn it. You can enjoy your anime in peace. I get that there’s a lot more of the ‘silence is violence’ and compelled speech thing going on, but I will need a lot more evidence of real consequences of silence before I stop pushing it as a strategy in such spots.

As a bioethicist, I support requiring students to take ethics.’ Ethics professors continue to show why they are no more ethical than the general population. We badly need ethics, but almost nothing labeled with the term ‘ethics’ contains ethics. Recent events have made this far clearer.

Republicans continue to prioritize not letting the IRS build a free digital tax filing system. I have other priorities, but important to note pure unadulterated evil. Even an ethicists get this one right.

Tipping indeed completely out of control, potential AI edition?

Flo Crivello: TK tried to warn us but you wouldn’t listen.

Molson: I was just asked to tip a hotel booking website.

Lighthaven, a campus in Berkeley, California, is now available for bookings for team retreats, conferences, parties and lodgings. Parties are $25-$75 per person, other uses are $100-$250 per day per person. I have been to two events here, and the space worked exceptionally well as a highly human-friendly, relaxing and beautiful place, with solid catering, good snacks and other resources, and lots of breakout areas. Future events being held here definitely raises my chance of attending, versus other locations in The Bay.

All is once again right with the world, Patrick McKenzie now gets his insurance from Warren Buffet. Because of course he does. Fun thread.

Magnolia Bakery to make weed edibles, but for now only for dispensaries in other states: Illinois, Nevada and Massachusetts.

iCarly revival is cancelled on a cliffhanger, and we will never find out about her mom. Seems somehow highly appropriate? Not that I’ve ever watched. Some cliffhangers are good endings.

Vitalik Buterin reports success in the creation of Zuzalu, a two-month-long intentional community of 200 people formed out of a resort during its off season. Give people a gorgeous and affordable setting, with likeminded people excited by a core group of projects, with people who want to be there and have no other commitments, and it can go great. I encourage more such experiments to happen with more ambition, but no I do not see this as heralding a Network State or other such nonsense.

You don’t have to worry about trying not to offend philosopher Agnes Callard. Oh, you’ll offend her, but not in the usual ways, so there is no avoiding it, and she doesn’t mind. So don’t worry about it. Also, I am guessing that she would get offended by you holding back to try and not offend her, rather than speaking or seeking truth.

Electric car growth is not slowing.

The tragedy is our continued focus on the symbolically superior pure electrics over the what could instead be a much larger number of hybrid vehicles.

It’s true, figuring out how to fund good science is complicated. Choose your fighter.

Stuart Buck: It is a testament to how little we know about metascience that two of the most prominent suggestions are:

1) Fund the person without regard for what the project is;

2) Fund the project without regard for who the person is (that is, anonymize grant applications).

Daniel Eth: Also we see both:

• complaints that scientists perform the actual work before applying for the grant and then use the grant money for their next project

• suggestions that we move from grants to prizes

I agree that we would get better results funding the person and letting them choose the project. I have zero faith in our process to choose valuable projects, instead predicting that the process will choose reliable projects over valuable ones. I also don’t love our ability to choose the best scientists, but I think we are capable of making at least some good decisions there, and if we offer those scientists freedom then many of them would choose valuable projects.

A hybrid approach might be good as well. If you make your mark we let you do what you want, if you don’t then you need to make a great proposal, but you can’t use your reputation for that at all and there are not many slots available, so it needs to be ambitious and awesome, you can’t play it safe. Or at least, that would be the idea.

Wait, non-creative types don’t use every moment as an opportunity to learn and train and figure things out (includes 2 minute clip)? Not all work looks like work, and it is not only the comedian that is (as Seinfeld explains elsewhere) mining actual everything that happens to them for material.

It is the torture that many of the rest of us embrace as well. Why wouldn’t I scan a new restaurant menu for font designs or think about the sound acoustics, isn’t that cool? That does not mean one cannot relax. Relaxing is imperative. But when people’s relaxing involves getting zero useful training data, it confuses me. I love me a relaxing bath, but I’m usually going to be listening to a podcast.

For the true value of working, most of you are slacking off, big time.

I cannot condone anything Meta does but yes, this is a good point:

Arc: I want the Zuck smart camera glasses to succeed mainly because the aesthetic of a bunch of people holding their phones up at events, especially concerts, is horrible, just looks like no one is there for the thing itself (though they may be, they’re distracted!)

Meta also seems to have had people who wanted to take various steps to protect mental health and make their products less addictive, including disabling some filters, and Mark Zuckerberg ultimately said no, citing lack of causal data on the harm, whereas the demand was clear as day. Reports involve noting things like the products providing dopamine hits and satisfying novelty seeking.

This all sounds quite bad, with talk of ‘worse than tobacco.’ But let us remember, also, the story of the man who read smoking was bad for him. So he quit reading.

I fear that if we punish Meta for having reports on harm and considering harm reduction interventions, the main effect will be that they and others censor and render illegible all discussions of harm and harm reduction.

Meta has much to answer for, but let us beware of hitting them on their internal processes until we know what incentives that would create.

Government actually working! IRS makes 83(b) election electronic signatures permanently valid, hopefully avoiding future time bombs exploding on founders when their stock vests. Still insane that the law works this way, but in practice should claim a lot fewer victims.

Even better, IRS launches free online tax filing service for 2024 season. For the pilot program, you need to be in one of 13 states (Arizona, California, Massachusetts, New York or the ones without a state income tax, so either obnoxious-tax city or no-tax city) and have a relatively simple tax return. Let’s fing go.

No10 is listening to my conversations, and I am here for it.

BBC: Smoking age in England should rise by one year, every year, so that eventually no-one can buy them, Rishi Sunak says.

The idea was put forward by a government-commissioned review in 2022.

Makes sense to me. If you are already addicted, and you can’t buy legally, you’ll have to buy illegally. Whereas if you choose to start knowing you will never be able to legally buy, that is a choice. Alternatively, we keep cigarettes legal, and the barrier to sell anything else is you show it is safer and less addictive than a cigarette.

Imagine this as well.

Joe Biden: Today, I’m proud to announce that we are taking our most comprehensive action ever to eliminate junk fees.

The Federal Trade Commission is proposing a new rule that would ban hidden fees across the economy and require companies to show consumers the all-in pricing upfront. The Consumer Financial Protection Bureau is also banning banks and credit unions from charging fees for basic services like checking your account balance.

These actions are going to save Americans tens of billions of dollars.

Danielle Fong: Do this for medicine and the consumer surplus will be in the trillions.

In a world where consumers increasingly see product offerings sorted by headline price, and with related human biases and adjustments of expectations, there is continuous pressure to utilize hidden ‘junk’ fees. Everyone is better off if all the prices are displayed in full up front.

As Danielle notes, the real value is in places like medicine. The junk fees there are even more outrageous, the true prices more hidden and out of our control.

How much should we worry that business will pass on costs to consumers in other ways here? That depends on execution details but my prior is not much. Yes, most of the costs get passed on in the base price, but that is the point. That is good. A lot of the surplus here is reduced cognitive load on the customer, a lot is in better customer choices. We are not actually trying to lower effective prices as such, although I expect a little of that on net.

Give SSI beneficiaries $3,200 stimulus checks, notice they have a bank balance over $2,000, attempt to claw back years of benefits that this balance proves they shouldn’t have gotten, stop the checks while the victims appeal? Yep.

FAA offers us new MOSAIC rulemaking for light-sport aircraft. Eli Dourado reports that the new rules are actually pretty great, giving much greater experimental flexibility on numerous fronts and plausibly hitting the safety sweet spot. You’ll get more passenger seats, higher speeds, more takeoff weight and more.

Eli says this could be transformative, allowing us to to maybe even get our flying cars some day.

Buy American rules are far worse than they look. Not only do you need to source everything from American suppliers (who are in turn sourcing from American suppliers and so on) but also everyone involved needs complex chain of custody systems in order to prove the Americanness of all the objects involved. Often entire projects incur massive delays. If one is determined to shoot oneself in the foot, one can do a lot less damage by using a mix of tariffs and subsidies.

The saga of SpaceX continues.

DabIsBad: >FWS was worried about a rocket hitting a shark No way. Absolutely zero fucking way. Are you kidding me?

>FWS was worried about the infinitesimally low chance of the rocket hitting a shark

>FWS asks SpaceX to calculate the risk of a shark hitting it

>SpaceX asks for data to calculate this

>FWS refuses to give data because it might somehow spread to shark fin hunters

>SpaceX asks if they could get another department of the FWS to do the calculation since they have the data

>FWS says they don’t trust the other department of the FWS

>Took them a bunch of time to resolve this shit

>Another agency asked about chance of rocket hitting a whale

This we have a tight lid on. Training AI models that could kill everyone, not so much. Can we please concentrate our anti-regulatory fire where it is so richly deserved?

Ro Khanna (team blue) says to Matt Gaetz (officially team red, mostly team schmuck), let’s have the house vote on the things like term limits and bans on member stock trading that the public 80% supports but that members for obvious reasons do not want, speculates that an actual vote might shame some into voting for it. Certainly it creates a prisoner’s (legistlato’s?) dilemma, where no one wants to vote against but also no one wants it to pass. Except most of the American people, but who cares about them? Forget it Matt, this is Congress.

Even better would be to go back to letting individual members introduce bills and then the House votes on them. We have computers, this need not disrupt regular business, perhaps if something passes that way you then do a traditional vote for real.

Scott Sumner’s film ratings for Q3. He is always exactly correct in his ratings, on the dimensions he is evaluating. If you want to see the highest Quality films, the objectively best films, he is here to tell you what they are.

Yes, Barbie is a 3.3 on his scale, and MI: Dead Reckoning is a 3.0 (out of 4) – I’d have said 2.8 or 2.9 for MI before learning it impacted Biden, and 3.4 or 3.5 for Barbie, but then I remember Scott is never wrong.

However, as always, one must remember that the dimension Scott is measuring is merely one dimension, which is also composed of its own subdivisions. There are many other dimensions to measure as well, that mostly are of no concern to Scott, but which I care about quite a bit. Is this something I actually want to watch, and would enjoy? Which is highly correlated with Scott’s ratings… once you control for certain factors. And which much more favors relatively recent films.

It is on my project list to do my own film rankings some day. They would look very different. Also at some point when things are quieter, to watch at least all of Scott’s 4.0s and 3.9s.

Note that this search cleanly brings up all his reviews.

I also completely endorse Scott’s horror at the ‘4DX’ theater, and would continue to happily pay a lot to never be in one again.

Dumb Money review: Kind of dumb. Also money. Would invest again. If Scott Sumner saw it he would probably give it at most a 2.5, but again, not why I saw it. Enjoy life.

Saw old movie The Adjustment Bureau, which was okay. Most unrealistic thing was that New York Senate elections were treated as competitive. Was enough about AI to not stand up to scrutiny in any way. Predicted Scott review that would be fair in its own way: Uninventive, seen it all before, philosophically incoherent, 2.2.

Wanted a short-ass dumb action movie one night and watched Avarice on Paramount+, which has no plot, no characters, no acting, no action and no movie. The gimmick is ‘she is an archer and has a bow’ I guess but if you replace it with ‘she carries a gun’ actual nothing changes. Actual 0.0.

Mostly I need to find time to watch more movies, and to always check reviews first and choose my targets more deliberately.

Warner Brothers attempts to bury a third movie, Coyote vs. Acme, as a tax write-off, despite a great concept, a finished product and good audience testing. This was not a ‘this is so bad we must pretend this never happened’ situation, whereas they said Batgirl was unreleasable and my imagination can suffice for a project entitled Scoob! Holiday Haunt. After creatives cancelled a lot of meetings with Warner Brothers, understandably concerned this might happen to their projects as well, WB agreed to let Coyote vs. Acme be shopped around instead. It sounds like fun.

Twitter’s algorithm is severely punishing external links, and on top of that they’ve stopped displaying the headline text of the links. As Nate Silver points out, this might sound like a good short term idea, but it’s a terrible long term plan, and no this is not a purely Elon Musk style of mistake.

Philippe Lemoine: It’s clear at this point that Twitter’s algorithm is now severely penalizing external links, which as Nate explains kind of makes sense in a short-sighted way from their point of view, but as a pretty niche writer relying on Twitter to disseminate his work it massively sucks.

Nate Silver: Yeah and I do think it’s short-sighted. It’s basically replicating the mistakes of Facebook’s News Feed. I’ve always liked Twitter >> Facebook because it isn’t a walled garden and contains my well-curated list of links/sources. (Or even For You, which actually works OK for me.)

It’s sort of a Tyranny of Algorithms thing because it’s easy to know what maxes out time on site in the short term (i.e. forcing you to stay there) but very hard to know in the long term.

Daniel Eth predicted sites will start putting the headline into the graphic. Paul Graham predicts it will be obnoxious.

Meanwhile Simon Willison offers a practical solution to auto-generate social media display cards.

Paul Graham: If Twitter doesn’t revert to the old way of displaying links, sites will start putting text onto the images. And they won’t just put a little text at the bottom, the way it used to be. Your feed will resemble a series of billboards.

The sites will optimize for clicks, so whatever is the most linkbaitish way to put text on images, that’s what we’ll get.

Up until now, I have found Twitter has retained most of its prior usefulness, so long as you use a chronological feed and lists.

I have found the lack of titles on links only mildly annoying. The severe punishment of links does not directly impact my feed. But together they mean less people are posting links, which has been a non-trivial downgrade.

An illustration: Referrals from Facebook and Twitter to top news sites have cratered.

Odd choice to not have the y-axis start at zero here, this decline is dramatic.

Axios blames this on lack of positive selection, that tech companies no longer try to ‘elevate quality information.’ There is doubtless some of that, there is also the factor that these news organizations keep trying to forcibly extract payments, and that links to outside are being intentionally throttled across the board, partly to avoid those payments, partly to keep social media users from leaving the website.

I would support a rule saying that major social media sites cannot throddle posts with outside links, but if they are also going to try and extract payment, not so much.

Also every ‘major news’ site is now behind a paywall. I can read links to the two were I have subscriptions, not to the others. Without a unified subscription option, most people will be unable to follow most links. If I can’t easily share WSJ or NYT links, the value of a subscription goes down a lot, which feeds the cycle. Even I chose one mainstream source to stay grounded on that (WaPo). I also find Bloomberg valuable enough to pay for. Most people don’t even go that far.

Tyler Cowen claims that it is good Elon Musk bought Twitter. Here is his reasoning:

I have disagreed with most of his design decisions, do not like the name change or rebrand, and I have been disappointed by many of his tweets and points of view, often disagreeing vehemently. That said, allowing the videos to be seen on Twitter is the right decision, and it is a very, very important decision.

So I end up glad that he bought Twitter. I also very much like the general feed and also the “Community Notes” features.

I am not sure how widely acknowledged this will be, but someone should say it, and I am happy to be the one. In general, more attention needs to be paid to “getting one big thing right.”

I strongly agree with the underlying principle of ‘getting one big thing right.’ Often this will indeed prove more important than ten or a hundred little things you get wrong, and the reverse would have been true as well.

The question is, what is the particular big thing one must get right? That is itself the big thing one must get right. I do not see the videos as so important one way or another. Freedom of speech in general is a plausible candidate for the big thing, and the past month has reminded us of its importance. Community notes is very good, it could perhaps be the big thing. Or they could together be the next big thing, with notes allowing much more free speech.

An example of a big thing one could get wrong, that might be going wrong, is to give too many interactive advantages to paid users, especially across multiple tiers. Asking for $8 a month is one thing, but I would be very careful with the new additional tiers.

On the bright side Bloomberg’s Aisha Counts says Twitter CEO Linda Yaccarino is claiming Twitter is cash flow positive excluding debt service and expects to be cash flow positive including debt service by 2024.

Nate Silver makes the case that a run by RFK Jr would more likely favor Biden than Trump. This seems right to me, with wide uncertainty bars. Kennedy does not hold positions compatible with the Democrats, and Libertarian runs typically hurt Republicans for obvious reasons.

Free speech is super important. I highly value free speech, and deeply thankful for the first amendment every time I see what happens in countries without one.

Presumably in response to some people not taking kindly to Paul Graham saying things one can totally say, he responded by posting this image, note the date on the older conversation:

Allow me to demonstrate what level we are on by saying this and noticing I do not expect any negative consequences whatsoever for saying it, because come on:

Zvi Mowshowitz hereby says: There are things in our society which are true that you cannot say.

Thus, by this scale, we are at Level 1. Which, by the definition of the scale, is the minimum level that has ever existed in any society, so it is not some huge indictment or big whoop. I do agree that the set of such things is larger now than it was in 2003, but less than in 2020, and again, I feel completely comfortable saying that, and would even if I lacked any power.

I would also not confuse level 1 here with the actual situation, which is that when there is an issue where lots of people are accusing each other of supporting genocide, saying pretty much anything is going to get one side, the other or both mad at you to some extent, that does not mean we can’t say things.

Also, my lord, reversed stupidity is not intelligence, the scale is not the territory, etc:

Rina Artstain: You’re at level -1, where you’re distorting “the truth” to look cool for level 4ers.

Paul Graham: Since higher = worse, level -1, whatever it was, would be good.

The ideal censorship level is zero, the same way the ideal Simulacra level is 1.

(Since everything is also about AI now, note the danger of an AI concluding things like ‘whatever seems like it would metaphorically be the -1 level must be good.’)

If you are surprised that new findings say the Eaterlin Paradox was bogus and money does indeed increase happiness indefinitely on a log scale, I am curious why.

Yes, obviously. No need to read the paper, yet an important paper to ensure exists.

Working Paper: We find that consumer surplus is the primary component of social impact (dwarfing profits, worker surplus, and externalities), suggesting that consumer impacts deserve more attention from impact investors. Existing ESG and social impact ratings are essentially unrelated to our economically grounded measures.

Increasing the minimum wage 10% increases low-end local rents 2.5%-4.5%.

Patrick McKenzie talked anti money laundering law back in February 2023, recommended for those interested if you haven’t seen it.

His new related piece, also recommended, is Seeing Like a Bank. The core idea is that banks could in theory solve your problems using bespoke professionals who can track information and understand the underling dynamics and use reason, but those people cost a hundred times as much the way doctors and lawyers do, and most bank issues very much do not require or justify such action and involve That Guy calling in time and again, so instead you get a three-tier calling system gating everything. But, if you know the right shibboleths, you can get the professional involved on your behalf and make things actually work. Banks also could in theory have software that worked well and kept good track of everything, but mostly they don’t and it will be a while before they do.

Historians do not know their economics, and thus do not know their history. Economists say the Great Depression was a story of aggregate demand contraction, monetary contraction, protectionism and failed Federal Reserve policy. History textbooks talk about income inequality, under-consumption and a stock market crash. Of course, the economists are also making big mistakes here, especially their failure to emphasize the gold standard and various contractionary policies implemented throughout, as always I recommend Scott Sumner’s The Midas Paradox.

No, it’s not free under $5, but if you want justifications for spending money you don’t have, ‘girl math’ and ‘boy math’ are now here for you. Why rationalize away your money all by yourself when you can have professionals help you do it, for free? Given your hourly, that means you’re making money. Which you can now spend. That’s math.

With a little software work, Mercury has routed around the requirement that every payment to a new person effectively requires a new W-9, because all the info in the W-9 is already known.

Paper claims the Tax Cuts and Jobs Act of 2017 substantially increased corporate investment, those with mean tax change increased domestic investment 20%, with equilibrium long-run effect of 6% greater domestic capital and 9% greater total capital, while net tax revenue declines less than 2% of baseline corporate revenue due to feedback effects. As Tyler Cowen notes, this will not settle the matter, very few will be convinced. In my model, those who want higher corporate taxes would not change their position even if they were indeed convinced, they care about some mix of fairness and equality and perception of fairness and equality, so efficiency and positive-sum arguments will land on deaf ears.

Is it this simple?

Alec Stapp: The massive divergence between consumer sentiment and economic reality is mind-boggling.

Joe Weisenthal asks: Inflation has been coming down, and the unemployment rate is low. Yet measures of consumer sentiment are still in the tank. Curious if anyone has any thoughts or theories? Would love to hear some takes.

Joe Light: A lot of the things I buy, outside of food and transit, are things I haven’t had to shop for in years, and each time I run into them, it’s like the inflation happened right at that moment

Matt Stroller: The CPI doesn’t include interest charges, aka cars and houses.

Tren Griffin: That people will focus on the *rate of changein core inflation rather than how much they pay for real world goods and services now, is a triumph of hope over experience. What did I pay for potatoes or laundry soap five years ago? How much higher is my grocery payment now?

Threadass: My rent went up 50% in one year, my utilities are about 30% more expensive, groceries and household goods are expensive, and avg interest rates on mortgages have doubled. I know many friends and (now former) colleagues who have been laid off in the past 6 months or so.

George (other thread): The economy can be summed up by an experience I had at a recent family reunion. Everyone was complaining about how shit the economy was and how expensive everything was

I pointed out that for the first time ever, every adult present had a good paying job they liked. Three people present had just been bragging about doubling their salaries. 2 people had just gotten back from their first ever Europe trips.

The raises and the jobs were things they felt they had earned. The prices going up were the government’s fault.

Kind of yeah. Things are no longer rapidly getting more expensive. Things are still much more expensive than they used to be, and interest rates rising means rent, mortgages and car payments are way up. It’s kind of a big deal. And for whatever reason, the government is responsible for the price level, and isn’t responsible for your raise or new job.

The results of the UAW strike illustrate the real effective rise in wages and the cost of living, with the minimum raise being 25%. Yes, there was some catchup there, but also they quote someone jumping from $20/hour to $35/hour. In my experience hiring people, I have also had to pay a lot more than I did a few years ago, far more than the official inflation rate.

Relatedly, Tyler Cowen notes income stability in America has been rising for decades. Yet workers do not seem less nervous. I posit this is partly because workers are more risk averse, partly both are caused by us otherwise giving people less margin for error, giving people less affordance to survive being down on your luck and still keep your life on track. Having a volatile income is not that scary if you think you will still get a family in the end. I do agree with Tyler that if you are not too risk averse, there are great rewards at start-ups and for other forms of risk taking here.

I also would presume that people notice the specter of future changes from AI.

Danielle Fong: ambient grift fields surround pools of persuadable capital (crypto, nonprofits, esg) and to first order, work to kick out alternative legitimate investments for that capital in time this parasitic behavior starves the host; not always killing it, but limiting growth, liveness.

Andrew Rettek: Yep. At equilibrium, this grift brings those pools of capital down to the average rate of return. Some things die there or die getting there.

Danielle Fong: having witnessed this happen three times, the decay time constant varies with the smart money (eg ~1 week for lottery winners, ~months to years for crypto wealth, ~1 decade for top tier big tech, generational for legendary investors)

Andrew Rettek: This is the EMH at work, and it will win eventually. It can only be slowed by some other force at least as strong. Agency is best.

Danielle Fong: everything I have ever witnessed, you always have major nonequilibrium. it’s practically never that case that every company or technology that could exist, should — especially because the frontier of the possible, important, or labor/resource competitive is always changing.

At equilibrium I would say that such effects do not merely bring pools of capital down to the average rate of return. It goes far lower than that, the equilibrium is that a vulnerable pool of money will be entirely extracted. Something about a fool and his money.

Banks increasingly give people the option to pay to access money from a check right away. Patrick McKenzie notes that notes that banks used to instead turn to their risk department, and also that you could ask for a waiver and someone in the bank was authorized to credit your account today.

Letting people pay to get the money now seems an odd combination of price discrimination against the poor that does not seem great, and adverse selection against the bank that seems far worse. The whole point of placing a hold on the money is to guard against cases where the check is fishy and the customer does not have the money when the check bounces. When do you think people will pay to cash it now? When they suspect the check will not clear and they intend to use the money right away. The market for this service does not seem like it should clear?

529 savings plans are technically owned by the account owner, not the beneficiary, because otherwise colleges would impose an effective tax rate of 95% by reducing aid awarded. Kids should essentially never save for college with their own money for the same reason, unless the plan is to actually pay the whole thing, because you will not make college any cheaper. It still boggles my mind that we allow colleges to demand to know your assets and then perform ~100% asset confiscation.

Thread by Chris Conlon on the FTC’s case against Amazon. The core case is that Amazon causes prices to be too high, which is perhaps the most absurd allegation I have heard in a long time. Conlon says the FTC does perhaps have a case on the part where charging a lower price elsewhere leads to a demotion in Amazon’s rankings, although as he notes what would be the remedy there? Also how exactly is the consumer harmed by that?

Ben Thompson covers the FTC complaint, says this is the key paragraph.

This case is about the illegal course of exclusionary conduct Amazon deploys to block competition, stunt rivals’ growth, and cement its dominance. The elements of this strategy are mutually reinforcing. Amazon uses a set of anti-discounting tactics to prevent rivals from growing by offering lower prices, and it uses coercive tactics involving its order fulfillment service to prevent rivals from gaining the scale they need to meaningfully compete. Amazon deploys this interconnected strategy to block off every major avenue of competition — including price, product selection, quality, and innovation — in the relevant markets for online superstores and online marketplace services.

Amazon’s coercive tactics to thwart competition lie within the Amazon website. You are totally free to offer products elsewhere. Amazon’s ‘anti-discounting’ is, as noted below, that they will price match. The rest of the complaint is found to be even less compelling, and the FTC framing every ordinary business offering such as Amazon Prime, bundling or (gasp) store-branded goods as inherently suspicious does it no favors.

Thread on the details of that FTC complaint against Amazon. What are the actual objections?

Douglas Farrar: Amazon pursues a pay-to-play scheme forcing sellers to buy ads. Worse, many of these ads are junk ads that aren’t relevant to what users search for. Jeff Bezos and co. call these Junk Ads “defects,” and sellers pay big bucks for them.

These ads do two things. First, they make the customer experience much worse, which is why you might search for “water bottles” and end up with offers to buy “buck urine”

They know this harms consumers, but they don’t care because maximizing advertising profit at all costs “has effectively become ‘law’ even if it has many flaws” according to a senior Amazon executive

The second thing all of Amazon’s ads do is raise prices for consumers. Sellers are forking over big money in a pay-to-play scheme. That cost goes directly to you according to Amazon’ own executives.

Forcing ad buys seems worse than a straightforward fee. Amazon gets the revenue either way, and it ‘pays for’ the ad buys by having the customer experience get worse. As described, this is a perverse process that Amazon has every reason to fix. Corporations make these kinds of errors all the time, no idea why it requires an FTC action other than a political hit job. There’s some strange underlying assumption that Amazon shouldn’t have control over or profit from its own website.

Amazon’s own economists even pointed out that by flooding search results with paid ads, Amazon steers shoppers towards higher-priced products and makes comparison shopping harder on the platform.

By the way, you as you can see above Amazon is still trying to keep some of the material in the complaint out of the public eye…

Amazon is really excited about how much money these ads are earning them by the way. In their Q3 report last week they said that revenue from ads grew 25%, which is faster than AWS.

Amazon being excited by this is bad news for Amazon if the ads are like this. Ads you have to buy are a zero sum game at best, again is this about disguising how much it cots to list on Amazon?

Amazon harmed price competition in other ways too. Former Amazon exec Jeff Wilke came up with the idea for an algorithm to avoid a “perfectly competitive market” where Amazon’s rivals lower their prices!

The concept is that Amazon will match any price drops by rivals, but not move to lower prices first. A price match. You’re calling out a price match. Ben Thompson notes that Amazon will also punish the seller (via lower prominence on the website) if they offer a higher price for an identical product at Amazon and a lower one elsewhere. So Amazon is demanding that prices on its website not be too high, at the cost of not featuring the related product if it isn’t listed at a competitive price, and that’s bad because it leads to high prices. Got it.

Those foolish enough to compete with Amazon by lowering prices met the full force of Amazon’s monopoly power. Take Jet dot com, which tried to compete by lowering seller fees.

Amazon also harms competition and extracts monopoly rents from sellers by requiring them to use Fulfilled By Amazon to have Prime eligibility. They once launched a program called Seller Fulfilled Prime which was a huge hit.

Amazon shut SFP down because they said deliveries weren’t on time. But new info today shows sellers using SFP met the delivery requirement set up by Amazon more than 95% of the time.

Yeah, 95% is… not actually that great, especially if the 5% is non-trivially late, and if there is not proper accountability and repair when something goes wrong. Also, your 95% on time is not my 95% on time.

Eric Boehm: Just read the line in the FTC’s own lawsuit, as helpfully screenshotted by Farrar: “Sellers enrolled in SFP met their promised ‘delivery estimate’ requirement set by Amazon more than 95% of the time in 2018.”

How many of those sellers were actually meeting the standards for Prime-level shipping? According to Amazon spokesman Tim Doyle, it was about 16 percent.

Were there also other motivations? Amazon’s rule is, as Ben Thompson points out, that if you pay Amazon for guaranteed delivery services, that it provides those services.

This freaked Amazon out. One Amazon exec said splitting Prime eligibility from FBA was an “oh crap” moment. Another one said a growing third-party logistics marketplace “keeps me up at night.”

There is plenty more I didn’t cover. And you can view all the newly unredacted material at the link below, thanks to the heroic work by the FTC attorneys working on this huge case.

None of that matters if Amazon’s shipping standards were rarely met.

The case rests, as Thompson notes, on the accusation that Amazon is a monopoly. Amazon does not seem like a monopoly to me, except that it has a monopoly on selling via Amazon. Which I would hope does not count?

Freddie DeBoer asks how the NBA can survive its current era of player empowerment. If star players keep demanding trades to the same handful of teams in attractive locations – such as Miami, New York and Los Angeles – and their teams feel forced to then do so for what amounts to pennies on the dollar, how do the teams in less fun places ever win?

Rules enforcing parity in pay and limiting hiring of superstars help mitigate the damage somewhat, but they also lock in the result. If we each get to pay the same fixed pool of money, but superstars are the best deals to begin with and can choose to take pay cuts in order to play in places they prefer and with better teams, solve for the equilibrium. Even if everyone plays it straight without loopholes, it only ends one way.

This is compounded by the NBA being highly skill-intensive with best-of-seven series that lack large sources of variance, with results being dominated by the superstars.

I see two plausible ways this goes. Option one is to embrace it, or at least accept it. The Indiana Pacers of the world will be second-division teams that occasionally threaten to break through. Perhaps that is fine. The Lakers and Heat generate more excitement anyway. Like in college football, each team can have its own baselines and goals. People come to the NBA to see star personalities playing exceptional ball, to enjoy that the players are off the hook. So release that hook.

Alternatively, we could go back to a world in which the players largely don’t control their fate or where they play. Except no, we can’t and we won’t. So option one then.

Which I really do think is fine. Then again, the NBA rarely holds my interest.

The MLB strike zone is in theory a function of player height and stance. In practice, with human umpires, any attempt to abuse this for technical loopholes would run afoul of the ‘who are you kidding’ clause, and the umpire would call strikes anyway. With robot umpires, there is no such clause. They are now testing, in the minor leagues, an automatic system that measures from knees to a baseball above the belt. For now, in the minors, that should work, because no one cares about winning and the goal is to make it to The Show, if you game the robot umpire it won’t do you any good. So the system might work fine.

However, if you move that system to The Show, the munchkins will be out in force, and you won’t like the results. Another illustration that the rule that a computer must follow to the letter often needs to be very different from the human rule. This is on top of the fact that umpires widen or narrow the strike zone based on the count to favor whoever is behind on it, and also that they favor the home team a bit and veterans and stars a bit, and sometimes the team trailing, all of which they are very much not supposed to do but very much do anyway.

I would be in favor of actively codifying all those quirky preferences, as long as they are symmetrical. Home teams winning more than half the the is good for the game if played fairly, as are the other adjustments. But the fans would not stand for it.

Remember the principle of Leaders of Men, to focus on what matters most.

Don Van Natta Jr. (ESPN): If you need two timeouts to put 12 men on the field for a FG try that hands your opponent a second chance to beat you, you should be fired.

Such incompetence is indeed a hint that perhaps some firing-worthy things are happening. On the margin it can cost you games. But we need periodic reminders that you pick your coach mostly on whether they can draft or recruit, sculpt, script and motivate a team. The things that fans rightfully yell about as dumb or incompetent? The fans are right, but ask how much they matter.

Zac Hill gives his perspective on what is wrong with Magic: The Gathering’s Standard format. He sees Standard as about board presence and doing powerful individual things players like to do, giving a ‘standard’ play experience. Variety is good, but the ‘good stuff’ midrange style strategies need to remain the heart of the format. When the dominant play patterns depart too much from this, and the best decks you see all the time are all about walls of text and detailed interactions or they fog until they take infinite turns or their curves only go up to three, the heart is gone.

I think there’s something to be said for that, although I do not take it as far as he does. I would instead place a lot more responsibility on the ‘too many words’ problem. Cards now are too complex, are too fiddly and noisy, and there are too many cards released. Players cannot keep up or keep track, driving them into evergreen formats that change more slowly or where universal awareness is not expected. Cards often not ever being in Standard at all compounds this problem.

Having Standard rotations every three years instead of two seems if anything anti-helpful. Yes, your core concept can last longer, but you also need to keep 50% more things in your head at once, and there is less room for new concepts to breathe. As Sam Black notes, a lot of Standard is letting new cards shine. You want it small.

Sam Black goes farther, saying that Standard no longer makes sense. When Magic was designed around Standard, especially when it included mechanically unified blocks, Standard was great. Now that many cards are never in Standard, and the releases offering the best value include many non-Standard-legal cards, what are we even doing? There is no grassroots support for Standard, no reason for there to be any, and tournament support alone won’t cut it.

Sam’s suggestion is to sunset Standard without replacement, as Magic is trying to do too many different things, and Standard is no longer key to the business model. I do not think, long term, that this works. I think Standard is serving several purposes other formats don’t. It lets new cards shine, allows play at a more modest power level, offers a compact set of cards players can hope to fully learn or have access to as a gateway to being competitive. More than that, it is constantly changing.

Commander, Modern, Pioneer and all the rest are mostly static. Sure there are changes, but also there mostly are not. It is cool to visit those worlds on occasion, but Magic requires a place players can innovate, 1-on-1 with 20 life, without being up against years or decades of refinement.

That does not strictly have to be Standard. I have always loved Block Constructed, although to have that you would need to have blocks, and Wizards says people don’t like blocks. So mostly we are talking rotation speed, also known as talking price, and figuring out how to navigate the desire to aggressively power creep with some new releases, to create something sustainable.

I admit I do not know how to do that. I fear that Magic has in important ways peaked, having picked its low-hanging fruit of low-complexity cards and simple concepts, and now picking the fruit of connection with other franchises. That the future is a game built around Commander, which means a whole sort of general mish mash of stuff that people do for fun but that holds little interest for me outside of perhaps occasional limited play or something like Premodern.

Magic: The Gathering unifies its two types of boosters into new play boosters, designed for draft and also sometimes with multiple rares.

Saffron Olive: Big news on booster packs. Both draft boosters *andset boosters are going away with Murders of Karlov Manor in a few months to be replaced by play boosters (which are essentially set boosters, but made to be draftable).

Overall this seems like a good change. Makes thing simpler – having draft, set and collector boosters seems excessive – plus the idea of drafting something akin to set boosters is interesting thanks to multiple rares and wildcard slots.

The biggest drawback is that it will make limited more expensive, at least in paper, since play boosters will sell for set booster prices and there will no longer be a cheaper “draft booster” option.

Seth Burn: Obviously this will mean different things to different players. As someone who buys draft booster boxes specifically to draft, or play team sealed (remember that), this change is absolutely miserable.

Drafts are not cheap. This will make them less cheap. Or perhaps it will make them slightly more cheap? A draft will cost more money, but before draft boosters gave you worse value than set boosters, so if you get full value from the cards you could come out ahead. If all you care about is drafting, not as good.

Magic has gone from three rarities, to four rarities, to having relatively more copies of its higher rarities. Which seems like going around in a circle. The rarer cards will be less bomb-like, which seems good. All of that with some price raises along the way, but one has to keep up with inflation somehow I suppose.

I am still out. I experimented briefly with Wilds of Eldraine limited, but the online experience failed to hold my interest, the logistics of playing in person are not great for me, and the continuous investment cost of keeping up is way too high. So perhaps a draft or two right around release when reading the cards as I go is still fine, some occasional light Premodern, and that’s probably about it.

It is simultaneously amazing the ways people talk about and frame questions about the dangers of self-driving cars, and also amazing that it is all not so much worse.

In this case, I mean, wow.

SAN FRANCISCO — A pedestrian crossing a busy intersection was struck by a regular car Monday night and hurled beneath a Cruise autonomous vehicle where she was trapped for several minutes until firefighters freed her, according to emergency responders and a video of the crash viewed by The Washington Post.

Nothing in the article suggested what the driverless car could have done, even in theory, to be safer, other than not being on the road in the first place.

Except, actually there was something. It seems that after the woman flew under the car, the driverless car dragged her along a bit in a way it shouldn’t have, causing more damage. This is not itself shocking or worrying to me, because ‘another accident flies a person under the car’ is exactly the kind of not-in-the-training-set situation that you’re not going to handle so well the first time.

Except, actually there was a bigger problem, you idiots.

Cruise’s initial tweets the day after the crash didn’t mention its vehicle dragging the woman as it pulled over. Indeed, the DMV says that when it met with Cruise the day after the incident, the company only showed footage from the initial crash, leaving out the part where the victim got dragged under the vehicle as it pulled over.

That got Cruise suspended in California, and then Cruise voluntarily suspended their other operations. Intentionally misleading officials about accidents is rather not okay. This is an existential threat to Cruise. If they actually did what they are accused of doing, they brought it on themselves, and have single-handedly pushed back development of self-driving cars for years.

I say if because California has a history of pulling tricks around this issue. I would not rule out, based on what I know, that Cruise did cooperate and they are twisting this all around somehow. Otherwise, yeah, this is something a company should fry for.

Cruise also has this little other issue.

Dan Elton: On Cruise’s driverless cars: “Those vehicles were supported by a vast operations staff, with 1.5 workers per vehicle. The workers intervened to assist the company’s vehicles every 2.5 to five miles” Huh, is that true? (I don’t believe anything I read in the @NYTimes these days).

Timothy Lee: People (rodney Brooke’s on bluesky and Gary Marcus) are describing the the 2.5 to 5 mile figure as “stunning” but I am not that surprised. You encounter a lot of stuff in 5 miles of driving in San Francisco.

That doesn’t mean cruise cars would have made a mistake once every 5 miles without help. It means that once every 5 miles there was a situation where the car wasn’t 100 percent sure what to do.

I bet in most of those cases it had the right idea with 99 percent confidence. But that’s not good enough when lives are potentially on the line. We shouldn’t fault companies for being careful.

With that said these details should not be secret. Waymo and cruise should tell the public a lot more about what is going behind the scenes.

Dan Elton: Wow. Cruise’s cars are being remotely operated 2-4% of the time? Am I reading this correctly? I feel like we’ve been lied to.

Cruise CEO (on Reddit): Cruise CEO here. Some relevant context follows.

Cruise AVs are being remotely assisted (RA) 2-4% of the time on average, in complex urban environments. This is low enough already that there isn’t a huge cost benefit to optimizing much further, especially given how useful it is to have humans review things in certain situations.

The stat quoted by NYT is how frequently the AVS initiate an RA session. Of those, many are resolved by the AV itself before the human even looks at things, since we often have the AV initiate proactively and before it is certain it will need help. Many sessions are quick confirmation requests (it is ok to proceed?) that are resolved in seconds. There are some that take longer and involve guiding the AV through tricky situations. Again, in aggregate this is 2-4% of time in driverless mode.

In terms of staffing, we are intentionally over staffed given our small fleet size in order to handle localized bursts of RA demand. With a larger fleet we expect to handle bursts with a smaller ratio of RA operators to AVs. Lastly, I believe the staffing numbers quoted by NYT include several other functions involved in operating fleets of AVS beyond remote assistance (people who clean, charge, maintain, etc.) which are also something that improve significantly with scale and over time.

Gary Marcus: How I am feeling, too. Lied to.

This is how it always starts. You have to bootstrap to get data, given how crazy the public is about every little thing that goes wrong. How else would you do it?

Except, why are they still doing this? Why does the CEO think this situation is long term stable? In some sense this is 97% less human driving, but this is very much not 97% of the way there nor does it provide 97% of the value. In the beginning, This Is Fine. Given how long they were on the road? Very much not fine.

Waymo vehicles at a minimum generated fewer insurance claims and generally have a strong safety record. Seems robust, even taking into account that Waymo has no reason to ever file an insurance claim, or to not pay money before letting anyone else write one. What is happening behind the scenes at Waymo? I don’t doubt some amount of support, but I presume they are much farther along in reducing the need.

Mr Beast provided clean water for up to 500,000 Africans. Some people were (checks notes) unhappy about this. He will keep doing such things. Anyone upset about this should notice the skulls on their uniforms. Doing good that earns you back the costs is the most effective of altruisms.

But yes, such people do exist.

Annamarie Forcino: “oh poor me, I’m getting canceled because I’m so nice 🥺” when in reality any criticism comes from the fact that people shouldn’t have to rely on a rich youtuber to get access to clean water.

Sydney Humanism Group: I understand that it can be frustrating to feel misunderstood or unfairly judged. However, it’s important to remember that being nice doesn’t exempt anyone from accountability.

Chris Freiman:

I will take a bold stance that it is good to help people even if those people should not have to rely on you.

I will also note that I am confident this bold stance is very popular.

Effective altruism too is rather popular among those who have heard of it. Which it had better be, if your whole plan is to dedicate everything to helping others as much as possible given fixed resources.

I’d certainly hope that the definition would be popular, to those who don’t know about the movement. And indeed, seems to go pretty well, and better with those who are familiar with the concept?

Stefan Schubert: YouGov finds that most Americans who have heard of effective altruism approve of it. Unclear how much to infer from it, though.

22% say they’re familiar with effective altruism; fairly strongly related to education level.

Next time an EA thinks they need to hide their affiliations, notice that this about as good as approval numbers ever get.

Scott Alexander writes up the experiment with impact markets. Judges found that the projects collectively were not worth what investors paid for them. Of 18 invested projects, 17 had negative ROI until Austin and Scott decided to overpay for two more, leaving 15 with negative returns.

The remaining one project got a 25x return, but that was because it was sold for $300 total, and it was pure community building at the University of Maryland future. Which I am assured is technically a Big 10 school. But for recruitment to be valuable, there needs to be something else effective going on.

Effective Altruism’s major organizations are largely under a single legal umbrella, which is inhibiting risk taking and information sharing quite a bit. This seems quite bad and I do not see sufficient corresponding upside.