Month: January 2023
The economics of why Noma is closing
Here is the take of yours truly:
Tyler Cowen, a professor of economics at George Mason University in Fairfax, Va., and a devoted restaurant-goer, says that people are misinterpreting Redzepi’s intentions with the closure. Cowen doesn’t think the chef is arguing that he can’t make money with Noma and its grand artistic ambitions. It’s just that he can make more money doing other, perhaps less stressful, things.
“He’s so well-known now, he can just do private events, cook for billionaires, special weddings and work two months a year or whatever and make more than he’s making in the restaurant,” Cowen says. “He’s the one who’s going to earn from here on out. Why slave every night till like 2 a.m. in a restaurant when you can set your own schedule and price discriminate, charging the super wealthy?”
Here is the longer WaPo article by Emily Heil and Tim Carman, presenting other views as well.
Wednesday assorted links
1. Arnold Kling on the Great Re-Evaluation.
2. Lynne Kiesling now has the Knowledge Problem Substack. And Matt Yglesias profile.
3. Some new nasal entry results from Fast Grants-funded research.
4. “A Chinese state-owned bank in Hong Kong is offering customers one shot of an mRNA vaccine if they make a deposit of HK$4mn ($512,000)…” (FT link)
5. An Indian view on who benefits and loses from GPT.
6. Good thread on recent growth miracles.
7. The Americans at Stanford pull the “harmful language” list.
8. Blake Hounshell triibute, RIP, he was great to work with.
Testing Freedom
In the latest Discourse Magazine I discuss the FDA’s long-standing fear and antipathy toward personalized medical tests and how this violates the 1st Amendment.
In 1972, the FDA confiscated thousands of home pregnancy tests, declaring that they were “drugs” meant to diagnose a “disease” and thus fell under the FDA’s regulatory dominion. The case went to the U.S. District Court for the District of New Jersey, and Judge Vincent P. Biunno ruled that the FDA had overstepped. “Pregnancy,” he said, “is a normal physiological function of all mammals and cannot be considered a disease … a test for pregnancy, then, is not a test for the diagnosis of disease. It is no more than a test for news….” As a result of Judge Biunno’s ruling, home pregnancy tests are easily available today from pharmacies, grocery stores and online shops without a prescription.
These days, debates over home pregnancy tests from the 1970s seem anachronistic and paternalistic. Yet the same paternalistic arguments appear again and again with every new testing technology. In the late 1980s, for example, the FDA simply declared that it would not approve at-home HIV tests, regardless of their safety or efficacy. As with pregnancy tests, the concern was that people could not be trusted with information about their own bodies…the first rapid at-home HIV test was developed and submitted to the FDA in 1987 [but] it took 25 years before the FDA would approve these tests. (Now, you can easily buy such a test on Amazon.)
…The FDA has a vital role in ensuring that tests are clinically accurate—tests should do what they say they do. Tests don’t need to be perfectly accurate to be useful (think of thermometers, personality tests and tire pressure gauges), but if a test advertises that it measures HDL cholesterol, it should do that within the tolerances the firm promises. The FDA has the technical knowledge to ensure that tests work, and that’s a skill that Americans value from the agency.
What Americans don’t want is to be told they can’t handle the truth. Yet when it came to at-home tests such as pregnancy tests, HIV tests and genetic tests, that’s exactly the reasoning the FDA used—and continues to use—to suppress information. The FDA should ensure that tests are safe, but “safety” means physical safety. The FDA may not declare a product unsafe because it might produce dangerous knowledge. Patients have a right to know about their own bodies. Our antibodies, ourselves. The FDA has authority over drugs and devices but not over patients.
Judge Biunno had it right back in 1972 when he said that diagnostic tests produce “news.” Test results, therefore, are a type of speech that fall under the First Amendment right to freedom of speech. The Supreme Court has repeatedly rejected restrictions on freedom of speech based on “a fear that people would make bad decisions if given truthful information”; thus, FDA restrictions on tests based on such fears are unconstitutional. The question of whether consumers will respond “safely” to test results is no more relevant to the FDA’s regulatory authority than the question of whether readers will respond safely to political news published in The New York Times. The FDA does not have the constitutional authority to regulate news.
My ChinaTalk podcast with Jordan Schneider
Here is the transcript, here is the podcast. Excerpt:
Jordan Schneider: You mentioned growing up reading classic novels and scholarship. What do you think will be relevant and not relevant about that sort of stuff in our new AI world?
Tyler Cowen: I suspect the classic texts will re-emerge in value. Reading Plato, Kant, or Adam Smith gives you a sense of a vision and big-picture thinking that the ais won’t be able to give us for some while — maybe never. [If] simply scanning the internet for facts, the AI might give you a very good digest — which you’ll consume in less time — and you’ll then seek out the thing the AI can’t give you at all.
That will, again, be radically original big-picture thinking.
Recommended, interesting throughout. We also talk about education, therapy, China, the person I envy most, the demand for pets, working for the Aztec empire, my own secret book project, and much more.
Do markets expect unaligned AGI risk?
Here is a new essay by Trevor Chow, Basil Halperin, and J. Zachary Mazlish, all favorite thinkers of mine, excerpt and these are their words I will not double indent:
“In this post, we point out that short AI timelines would cause real interest rates to be high, and would do so under expectations of either unaligned or aligned AI. However, 30- to 50-year real interest rates are low. We argue that this suggests one of two possibilities:
- Long(er) timelines. Financial markets are often highly effective information aggregators (the “efficient market hypothesis”), and therefore real interest rates accurately reflect that transformative AI is unlikely to be developed in the next 30-50 years.
- Market inefficiency. Markets are radically underestimating how soon advanced AI technology will be developed, and real interest rates are therefore too low. There is thus an opportunity for philanthropists to borrow while real rates are low to cheaply do good today; and/or an opportunity for anyone to earn excess returns by betting that real rates will rise.
In the rest of this post we flesh out this argument.
- Both intuitively and under every mainstream economic model, the “explosive growth” caused by aligned AI would cause high real interest rates.
- Both intuitively and under every mainstream economic model, the existential risk caused by unaligned AI would cause high real interest rates.
- We show that in the historical data, indeed, real interest rates have been correlated with future growth.
- Plugging the Cotra probabilities for AI timelines into the baseline workhorse model of economic growth implies substantially higher real interest rates today.
- In particular, we argue that markets are decisively rejecting the shortest possible timelines of 0-10 years.
- We argue that the efficient market hypothesis (EMH) is a reasonable prior, and therefore one reasonable interpretation of low real rates is that since markets are simply not forecasting short timelines, neither should we be forecasting short timelines.
- Alternatively, if you believe that financial markets are wrong, then you have the opportunity to (1) borrow cheaply today and use that money to e.g. fund AI safety work; and/or (2) earn alpha by betting that real rates will rise.
An order-of-magnitude estimate is that, if markets are getting this wrong, then there is easily $1 trillion lying on the table in the US treasury bond market alone – setting aside the enormous implications for every other asset class.”
TC again: I am pleased that they wrote a separate companion piece on Cowen’s Third Law.
J. Barkley Rosser, RIP
Alas Barkley has passed away. He was a friend of mine and to many of us at GMU, a well-known economist, a feisty commentator here at MR and much more…here is one profile.
What we know about road deaths during the pandemic
The study verified that the absence of traffic jams played some role in allowing drivers to reach dangerous speeds on too-wide roads, but the researchers also found that the most significant differences between their forecast and real-world death totals happened in the dead of night, when most roads have always been congestion-free.
Between 10 p.mm and 1:59 a.m., deaths were nearly 22 percent higher than expected; during the typical morning rush hours, by contrast, deaths were actually 6.3 percent lower than the model anticipated they’d be. The late afternoon and evening rush hour, meanwhile, “did not differ significantly from the forecast.”
…2020 also saw an increase in hit-and-runs, which clocked in at 31.2 percent higher than originally forecast.
…According to AAA, “about 70 percent of the entire increase in driver fatal crash involvement [between May and December of 2020] was specifically among males under the age of 40.” Tefft suspects that increase may have been particularly driven by the minuscule subset of young, male motorists who were emboldened to do risky things on the road when the world shut down, though the data doesn’t tell him exactly why.
The article has further points of interest.
Tuesday assorted links
1. There are no elected officials left in Haiti.
2. Was the T. Rex actually pretty smart?
3. The Microsoft deal with OpenAI will be big.
4. Non-drinking is on the rise amongst the English.
6. Anti-war Russians, living in Latin America.
7. How the human walk evolved for endurance, not speed (Wired).
8. “A Dutch supermarket chain introduced slow checkouts for people who enjoy chatting, helping many people, especially the elderly, deal with loneliness. The move has proven so successful that they installed the slow checkouts in 200 stores.” Link here.
Why did the gender wage gap stop narrowing?
During the 1980s, the wage gap between white women and white men in the US declined by approximately 1 percentage point per year. In the decades since, the rate of gender wage convergence has stalled to less than one-third of its previous value. An outstanding puzzle in economics is “why did gender wage convergence in the US stall?” Using an event study design that exploits the timing of state and federal family-leave policies, we show that the introduction of the policies can explain 94% of the reduction in the rate of gender wage convergence that is unaccounted for after controlling for changes in observable characteristics of workers. If gender wage convergence had continued at the pre-family leave rate, wage parity between white women and white men would have been achieved as early as 2017.
That is from a new NBER working paper by Peter Q. Blair and Benjamin Posmanick. Might the gender wage gap be one economics topic where a naive, mood-affiliated view on it best predicts a bunch of other bad views on totally separate topics?
How much did pre-ACA Medicaid expansions matter?
This paper examines the impact of Medicaid expansions to parents and childless adults on adult mortality. Specifically, we evaluate the long-run effects of eight state Medicaid expansions from 1994 through 2005 on all-cause, healthcare-amenable, non-healthcare-amenable, and HIV-related mortality rates using state-level data. We utilize the synthetic control method to estimate effects for each treated state separately and the generalized synthetic control method to estimate average effects across all treated states. Using a 5% significance level, we find no evidence that Medicaid expansions affect any of the outcomes in any of the treated states or all of them combined. Moreover, there is no clear pattern in the signs of the estimated treatment effects. These findings imply that evidence that pre-ACA Medicaid expansions to adults saved lives is not as clear as previously suggested.
That is a new NBER working paper from Charles J. Courtemanche, Jordan W. Jones, Antonios M. Koumpias, and Daniela Zapata.
Here are some relevant pictures. Now, would you expect subsequent Medicaid expansions to have higher, lower, or the same marginal value?
What should I ask Rick Rubin?
I will be doing a Conversation with him, here is Wikipedia:
Frederick Jay Rubin is an American record producer. He is the co-founder (alongside Russell Simmons) of Def Jam Recordings, founder of American Recordings, and former co-president of Columbia Records.
Rubin helped popularise hip hop by producing records for acts such as the Beastie Boys, Geto Boys, Run-DMC, Public Enemy, and LL Cool J. He has also produced hit records for acts from a variety of other genres, predominantly heavy metal (Danzig, System of a Down, Metallica, and Slayer), alternative rock (The Cult, Red Hot Chili Peppers, The Strokes, and Weezer), and country (Johnny Cash and The Chicks).
In 2007, Rubin was called “the most important producer of the last 20 years” by MTV and was named on Time‘s list of the “100 Most Influential People in the World“.
So what should I ask?
And I am excited for his new book The Creative Act: A Way of Being.
Where are all the workers?
The subtitle of the paper is “From Great Resignation to Quiet Quitting”, here is the abstract:
To better understand the tight post-pandemic labor market in the US, we decompose the decline in aggregate hours worked into the extensive (fewer people working) and the intensive margin changes (workers working fewer hours). Although the pre-existing trend of lower labor force participation especially by young men without a bachelor’s degree accounts for some of the decline in aggregate hours, the intensive margin accounts for more than half of the decline between 2019 and 2022. The decline in hours among workers was larger for men than women. Among men, the decline was larger for those with a bachelor’s degree than those with less education, for prime-age workers than older workers, and also for those who already worked long hours and had high earnings. Workers’ hours reduction can explain why the labor market is even tighter than what is expected at the current levels of unemployment and labor force participation.
Dain Lee, Jinhyeok Park, and Yongseok Shin wrote that new NBER working paper, important work for understanding our current time.
Monday assorted links
1. AEA vs. EJMR? (Bloomberg).
2. Chollet with some GPT skepticism.
3. Noma in Copenhagen is closing (NYT). “…Mr. Redzepi admitted to bullying his staff verbally and physically, and has often acknowledged that his efforts to be a calmer, kinder leader have not been fully successful.”
4. “Using a 3-second sample of human speech, it can generate super-high-quality text-to-text speech from the same voice. Even emotional range and acoustic environment of the sample data can be reproduced. Here are some examples.” Link here.
5. Joshua Kim comment on my higher education worries. I think he is saying they don’t get enough money!?
The Extreme Shortage of High IQ Workers
At first glance it seems peculiar that semiconductors, a key item of national strategic interest, should be produced in only a few places in the world, most notably Taiwan, using devices produced only in Eindhoven in the Netherlands by one firm, ASML. Isn’t the United States big enough to be able to support all of these technologies domestically? Yes and no.
Semiconductor manufacturing is the most difficult and complicated manufacturing process ever attempted by human beings. A literal spec of dust can ruin an entire production run. How many people can run such a factory? Let’s look at the United States. The labor force is approximately 164 million people which sounds like a lot but half of the people in the labor force have IQs below 100. More specifically, although not everyone in semiconductor manufacturing requires a PhD, pretty much everyone has to be of above average intelligence and many will need to be in the top echelons of IQ.
In the entire US workforce there are approximately 3.7 million workers (2.3%) with an IQ greater than two standard deviations above the mean. (Mean 100, sd, 15, Normal dist.) Two standard deviations above the mean is pretty good but we are talking professor, physician, attorney level. At the very top of semiconductor manufacturing you are going to need workers with IQs at or higher than 1 in a 1000 people and there are only 164 thousand of these workers in the United States.
164 thousand very high-IQ workers are enough to run the entire semiconductor industry but you also want some of these workers doing fundamental research in mathematics, physics and computer science, running businesses, guiding the military and so forth. Moreover, we aren’t running a command economy. Many high-IQ workers won’t be interested in any of these fields but will want to study philosophy, music or English literature. Some of them will also be lazy! I’ve also assumed that we can identify all 164 thousand of these high-IQ workers but discrimination, poverty, poor health, bad luck and other factors will mean that many of these workers end up in jobs far below their potential–the US might be able to place only say 100,000 high-IQ workers in high-IQ professions, if we are lucky.
It’s very difficult to run a high-IQ civilization of 330 million on just 100,000 high-IQ workers–the pyramid of ability extends only so far. To some extent, we can economize on high-IQ workers by giving lower-IQ workers smarter tools and drawing on non-human intelligence. But we also need to draw on high-IQ workers throughout the world–which explains why some of the linchpins of our civilization end up in places like Eindhoven or Taiwan–or we need many more Americans.
What is an optimum degree of LLM hallucination?
Ideally you could adjust a dial and and set the degree of hallucination in advance. For fact-checking you would choose zero hallucination, for poetry composition, life advice, and inspiration you might want more hallucination, to varying degrees of course. After all, you don’t choose friends with zero hallucination, do you? And you do read fiction, don’t you?
(Do note that you can ask the current version for references and follow-up — GPT is hardly as epistemically crippled as some people allege.)
In the meantime, I do not want an LLM with less hallucination. The hallucinations are part of what I learn from. I learn what the world would look like, if it were most in tune with the statistical model provided by text. That to me is intrinsically interesting. Does the matrix algebra version of the world not interest you as well?
The hallucinations also give me ideas and show me alternative pathways. “What if…?” They are a form of creativity. Many of these hallucinations are simple factual errors, but many others have embedded in them alternative models of the world. Interesting models of the world. Ideas and inspirations. I feel I know what question to ask or which task to initiate.
Oddly enough, for many queries what ChatGPT most resembles is…don’t laugh — blog comments. Every time I pose a query it is like putting a blog post out there, or a bleg, and getting a splat of responses right away, and without having to clog up MR with all of my dozens of wonderings every day. Many of those blog comment responses are hallucinations. But I learn from the responses collectively, and furthermore some of them are very good and also very accurate. I follow up on them on my own, as it should be.
LLMs are like giving everyone their own comments-open blog, with hallucinating super-infovores as the readers and immediate response and follow-up when desired. Obviously, the people with some background in that sector, if I may put it that way, will be better at using ChatGPT than others.
(Not everyone is good at riding a horse either.)
Playing around with GPT has in fact caused me to upgrade significantly my opinion of MR blog comments — construed collectively — relative to other forms of writing.
Please do keep in mind my very special position. The above may not apply to you. I have an RA to fact-check my books, and this process is excellent and scrupulous. Varied and very smart eyes look over my Bloomberg submissions. MR readers themselves fact-check my MR posts, and so on. Having blogged for more than twenty years, I am good at using Google and other methods of investigating reality. At the margin, pre-LLM, I already was awash in fact-checking. If GPT doesn’t provide me with that, I can cope.
And I don’t take psychedelics. R-squared is never equal to one anyway, not in the actual world. And yet models are useful. Models too are hallucinations.
So if GPT is doing some hallucinating while at work, I say bring it on.