China fact of the day
In 2022 more of the top-tier ai researchers working in America hailed from China than from America.
Of course the Canadians do their part to help make this come true. Here is a much longer survey from The Economist about science in China. Via Z.
Nat Friedman discusses Sora, and image wisdom
NF: Yeah, I interviewed the Sora creators yesterday, the day before on stage at an event and it was super interesting to hear their point of view. I think we see Sora as this media production tool, that’s not their view, that’s a side effect. Their view is that it is a world simulator and that in fact it can sort of simulate any kind of behavior in the world, including going as far as saying, “Let’s create a video with Ben and Daniel and Nat and have them discuss this,” and then see where the conversation goes. And their view is also that Sora today is a GPT-1 scale, not a lot of data, not a lot of compute, and so we should expect absolutely dramatic improvement in the future as they simply scale it up and thirdly that there’s just a lot more video data than there is text data on the Internet…
And then Andrej Karpathy, I was talking to him the other day too, and he said, “There’s something strange going on-”
[Ben Thompson] And a picture is worth a thousand words by the way, so the number of tokens there is just astronomically larger.
NF: He was exploring this idea that the world model and image and video models actually might better than in text models. You ask it for a car engine, someone fixing a carburetor and just the level of detail that can be in there is extraordinary, and maybe we made a mistake by training on the text extracted from Common Crawl and what we should do instead. I asked him for his most unhinged research idea. He said what we should do instead is train on pictures of web pages and when you ask the model a question, it outputs a picture of a web page with the answer and maybe we’d get way more intelligence and better results from that.
That is from his dialogue with Ben Thompson and Daniel Gross, gated but worth paying for.
Claims about Taylor Swift
Taylor Swift’s record-shattering Eras Tour is continuing to supercharge consumer spending as it enters its U.K. leg, suggesting that the Bank of England may not be out of the woods yet in its fight against inflation.
As hundreds of thousands of dedicated Swifties flock to London in August to see the singing sensation during her final U.K. dates, the economic boost could be enough to defer a possible September interest rate cut, according to investment bank TD Securities.
“We still anticipate a BoE cut in August, but the inflation data for that month might keep the MPC (Monetary Policy Committee) on hold in September,” the bank’s macro strategist, Lucas Krishan, and its head of global macro strategy, James Rossiter, wrote in a note Friday.
Here is more from CNBC, via D. And recently Taylor Swift fans caused seismic activity at an Edinburgh show.
Sunday assorted links
Finland bleg
Helsinki, Turku, and Tampere — what please are your recommendations? I thank you in advance for your sage counsel.
Claims about Africa and its politics
From Ken Opalo:
A common misperception economic policies in African states tend to be statist, far-Left, or anti-market. This is not supported by the data (see examples of Kenya, Nigeria, Cote d’Ivoire, and South Africa below). In actual fact, most governments in the region tend to be eager adopters of allegedly apolitical “best practices” that are essentially center-right economic orthodoxy. If you add to this their social policies, the modal African government is essentially Center-right. What these countries often fail on is implementation (partially because said policies are seldom useful in context and/or due to weak state capacity).
And this:
Third, the lack of governing experience and decades of state repression have led to a perversive anti-statist discourse and politics on the African Left. Among Leftist intellectuals, the colonial origins of the African state has been used as a reason to perpetually delegitimize state-building (many of the same intellectuals suffered state repression). In this rendering, the African state can never overcome the original sin of colonial origin; and should be abolished and replaced with a Pan-African state (which presumably would be better at deploying coercion and providing public goods and services). At the same time, many economically-ascendant Africans who are broadly sympathetic to Leftist politics harbor anti-statist sentiments when it comes to the economy and tend to overstate the statist origins of African economic underdevelopment — this partially reflects the ideological hold of economic orthodoxy in the region. The reality, of course, is that African countries are terribly under-governed. Data on security and law enforcement, registration of births and deaths, education attainment, taxation, expenditure absorption, economic regulation, etc. all point to the fact that the contemporary African state is too small and too weak to meet the challenges of modern economics and politics.
Here is much more. I am not sure how much the left- vs. right-wing framing applies to Africa at all — sometimes I think the better category is “prioritizes things going well, or not,” as part of the author’s remarks would seem to indicate.
*The Wrong Stuff: How the Soviet Space Program Crashed and Burned*
By John Strausbaugh, an excellent book. Here is one good passage of many:
Putting dogs on top of rockets was nothing new. Since so little was known about the effects that blasting off in a rocket might have on th ehuman body and brain — the g-force of acceleration, the disorientation of weightlessness, the impact of radiation, the g-force of deceleration — the Soviets and the Americans both had been using various species of animals to test conditions since the 1940s. The Americans started sending up fruit flies aboard their White Sands V-2s in 1947. An anesthetized rhesus monkey they named Albert II…went up eighty-three miles in a V-2 in 1949. Unfortunately, his parachute failed to oepn on reentry and he was smashed to death on impact with the ground. The Americans continued to send up primates in the 1940s and 1950s. Something like two-thirds of them died. They used many other species as well, maybe the oddest of which was black bears, who were strapped into a rocket-powered sled at a facility with the deceptively sweet name the Daisy Track to test the physical effects of ultra-rapid acceleration and deceleration.
Recommended.
Saturday assorted links
1. Iran to give awards to religious leaders who convince their congregants to have more babies.
3. How much of life on earth is “dormant”?
4. Akira Endo, RIP. He discovered statins.
5. ““Walking” Tree That Looks Like An Ent Just Won New Zealand Tree Of The Year.”
6. What rate of tariff needed to replace income tax revenues?
Where are the female composers?
Ludwig van Beethoven, Johann Sebastian Bach, and Frederic Chopin are household names, but few will recognize Francesca Caccini, Elisabeth Lutyens or Amy M. Beach, who are among the top-10 female composers of all time. Why are female composers overshadowed by their male counterparts? Using novel data on over 17,000 composers who lived from the sixth to the twentieth centuries, we conduct the first quantitative exploration of the gender gap among classical composers. We use the length of a composer’s biographical entry in Grove Music Online to measure composer prominence, and shed light on the determinants of the gender gap with a focus on the development of composers’ human capital through families, teachers, and institutionalized music education. The evidence suggests that parental musical background matters for composers’ prominence, that the effects of teachers vary by the gender of the composer but the effects of parents do not, and while musician mothers and female teachers are important, they do not narrow the gender gap in composer prominence. We also find that the institutionalization of music education in conservatories increases the relative prominence of female composers.
That is from a new research paper by Karol Jan Borowiecki, Martin Hørlyk Kristensen, and Marc T. Law, via K.
Denmark recalls Korean ramen for being too spicy
Three fiery flavours of the Samyang instant ramen line are being withdrawn: Buldak 3x Spicy & Hot Chicken, 2x Spicy & Hot Chicken and Hot Chicken Stew.
Denmark’s food agency issued the recall and warning on Tuesday, urging consumers to abandon the product.
But the maker Samyang says there’s no problem with the quality of the food.
“We understand that the Danish food authority recalled the products, not because of a problem in their quality but because they were too spicy,” the firm said in a statement to the BBC.
“The products are being exported globally. But this is the first time they have been recalled for the above reason.”
It’s unknown if any specific incidents in Denmark had prompted authorities there to take action.
The Danish Veterinary and Food Administration said it had assessed the levels of capsaicin in a single packet to be “so high that they pose a risk of the consumer developing acute poisoning”.
“If you have the products, you should discard them or return them to the store where they were purchased,” it said in a statement.
It also emphasised the warning for children, for whom extremely spicy food can cause harm.
The notice has sparked heated discussion online with many amused reactions from lovers of spicy food. Many have made assertions about the Danes’ low tolerance for spice.
Here is more from Frances Mao. Via Sam Mendelsohn.
Another Friday assorted link(s)
Trump on AI. It feels like someone just showed him a bunch of stuff for the first time? And here is Trump on UFOs.
Friday assorted links
1. New AI video generator, open for public use.
2. The Chicago Cubs really were hurt by playing day games (a semi-new Substack from a GMU Ph.D. student). And his (Nicholas Decker) observations on South African unemployment.
3. Major deposits of rare earth minerals discovered in Norway, never underestimate elasticity of supply.
4. Summary take on the Mediterranean diet, and other diets.
5. “The Israelis are using trebuchets. The Chinese and Indians are fighting with clubs and knives along the LOC. There is trench warfare in Ukraine. Retro-fication of warfare continues” Link here.
6. Search instead of scaling? And further results in that general direction. Many of the “AI winter” types are not taking these directions seriously enough. And from Rohit: “This is wild, and also why it is so hard to say what llms can’t do, because providing it a template to do a thing is equivalent to teaching it to do it.”
Enhancing FDA Information Sharing for Neglected Tropical Diseases
Many countries look to the US FDA for guidance on approval decisions. In fact the FDA will sometimes receive and evaluate drugs and vaccines whose primary market is in less developed countries. Fexinidazole, for example, is a drug for treating African trypanosomiasis, i.e. sleeping sickness. We don’t get many cases of sleeping sickness in the US but there are many such cases in the Democratic Republic of Congo.
Thus, the US FDA is providing a useful service, both to US pharmaceutical firms and especially to developing countries. That’s great. But Jacob Trefethen of OpenPhil notes that for odd bureaucratic and legal reasons we redact a lot of information that could be useful to the countries that actually will use these treatments. Here, for example, is an excerpt from the approval decision for Fexinidazole:
What’s especially strange here is that as far as Trefethen, or I, can tell, no one wants this! The FDA has no reason to hide this information, the company submitting the proposal surely wants as much information as possible sent to the countries where they will ultimately need to get approval (remember this is successful applications!) and the medical agencies in the developing countries would like to get context to have confidence in the FDA’s decisions. Instead, it seems that these drugs are getting caught in rules intended to protect pharmaceutical firms in other contexts. Thus, Trefethen makes two suggestions:
let’s create a track for products on the Neglected Tropical Disease list, sharing assessments with few or no redactions with the WHO Pre-Qualification (PQ) system, and allow PQ to share those documents further with regulators in partner countries.
Such an approval track already exists in the EU:
[The EU] have an approval track for products that are mostly going to be used elsewhere. If you apply using that track, they loop in regulators from those countries too. They share the documents assessing your clinical data and inspecting your manufacturing site with the WHO prequalification (PQ) team – the team whose stamp of approval speeds things up for many countries with less experienced national regulators. Gavi and the Global Fund need a product to be prequalified in order to buy it through the UN procurement agencies (e.g. UNICEF, for children’s vaccines).
Even without an approval track there are other small changes in priority and emphasis that could improve information sharing. The FDA is not unaware of these information sharing issues, for example, and there are procedures in place for confidentiality agreements with other countries. Trefethen suggests these could be given greater priority.
FDA leadership should set aggressive goals to complete more two-way Confidentiality Commitments with lower- and middle-income country regulators.
Extend the scope of existing commitments, when they’re limited, to allow sharing in more areas – especially related to drug approvals.
Extend 708(c) authority to more country agreements, not just those with European countries, to allow sharing of full documents that include trade secrets.
I’ve long advocated for peer approval, Trefethen gets into the weeds to point to specific ideas to make this a more useful idea, especially for developing countries. See Trefethen for more ideas!
Jimmy’s Jobs of the Future episode
From Jimmy:
Here is the post on LinkedIn and here is the post on X / Twitter. Any engagement / amplification would be much appreciated. Here is the link to the YouTube episode.
I had good fun doing this episode…
How to think about UAPs
I have a new Bloomberg column on that topic, I would most of all stress the closer:
When I think about all this, I try to keep two questions separate. First, is there a major puzzle to account for? And second, what is the best explanation for that puzzle? It helps to focus on the first question in isolation, since we can’t seem to keep our heads on straight when it comes to the second.
By admitting that there is a real puzzle to be solved, the Senate Intelligence Committee has moved decisively to answer the first question. Once we clarify exactly what the puzzle is, maybe we’ll be able to make some progress explaining it.
As for background:
The Senate Intelligence Committee isn’t buying it. The Intelligence Authorization Act , which it passed last week, among other things calls for review of the All-Domain Anomaly Resolution Office . The bill would also limit research into what are now called UAPs (for unidentified anomalous phenomena) unless Congress is informed and add whistleblower protections for anyone who might wish to step forward and speak their minds.
Recommended, I make other points as well.