Wednesday assorted links
1. Where are all the trillion dollar biotech companies?
2. Ne Zha 2 is bombing in America.
3. More John Cochrane on the Fed.
4. Some new problems with the minimum wage.
5. “Complaining about feminization does not a man make”.
6. Did the rise of OpenAI stop Google from being broken up?
7. Geoffrey Hinton moves away from the Doomer position.
8. Car sketch with labels, noting GPT is probably not the best model for this task, at least not the sketch part of it.
Read it and weep
A global bond sell-off deepened on Wednesday, driving the yield on the 30-year US Treasury to 5 per cent for the first time since July, as investors’ fears over rising debt piles and stubbornly high inflation dominated trading.
Longer-dated bonds bore the brunt of the selling, with the yield on the 30-year Treasury up 0.03 percentage points at 5 per cent and Japan’s 30-year bond yield hitting a record high of 3.29 per cent.
In the UK, long-term borrowing costs climbed further after reaching their highest level since 1998 on Tuesday. The 30-year gilt yield rose 0.06 percentage points to 5.75 per cent.
Here is more from the FT.
Where are the trillion dollar biotech companies?
In today’s market, even companies with multiple approved drugs can trade below their cash balances. Given this, it is truly perplexing to see AI-biotechs raise mega-rounds at the preclinical stage – Xaira with a billion-dollar seed, Isomorphic with $600M, EvolutionaryScale with $142M, and InceptiveBio with $100M, to name a few. The scale and stage of these rounds reflect some investors’ belief that AI-biology pairing can bend the drug discovery economics I described before.
To me, the question of whether AI will be helpful in drug discovery is not as interesting as the question of whether AI can turn a 2-billion-dollar drug development into a 200-million-dollar drug development, or whether 10 years to approve a drug can become 5 years to approve a drug. AI will be used to assist drug discovery in the same way software has been used for decades, and, given enough time, we know it will change everything [4]. But is “enough time” 3 years or half a century?
One number that is worth appreciating is that 80% of all costs associated with bringing a drug to market come from clinical-stage work. That is, if we ever get to molecules designed and preclinically validated in under 1 year, we’ll be impacting only a small fraction of what makes drug discovery hard. This productivity gain cap is especially striking given that the majority of the data we can use to train models today is still preclinical, and, in most cases, even pre-animal. A perfect model predictive of in vitro tox saves you time on running in vitro tox (which is less than a few weeks anyway!), doesn’t bridge the in vitro to animal translation gap, and especially does not affect the dreaded animal-to-human jump. As such, perfecting predictive validity for preclinical work is the current best-case scenario for the industry. Though we don’t have a sufficient amount and types of data to solve even that.
Here is the full and very interesting essay, from the excellent Lada Nuzhna.
Emergent Ventures winners, 46th cohort
William Wu, NoVa, piano, Prokofiev and Rachmaninoff.
Olga Niekrasova, Kyiv, animated film.
Anhelina Leshak, Ukraine/GWU, general career support.
Myles Fritts, Florida, fish and genetics.
Luzia Bruckamp, LSE, education and fertility.
Maximilian Jager, Frankfurt, LLMs to outline the German regulatory code.
Reyansh Sharma, London/Cambridge, open source math for the AIs.
Donnacha Fitzgerald, improve genome editing, London, from Ireland.
Sam Glover and Stella Tsantekidou, free speech in the UK.
Lyubov Guk, London/Ukraine, immigrant entrepreneurs.
Lily Geidelberg, London, AI and your calendar.
Ben Johnson, London, new policy and research developments for R&D.
Aden Nurie and Ian Cheshire, Tampa and Wallingford, “We’d like a grant to help us turn ChatARV into the best AI algorithm for finding comparable properties on the market.”
Jessie Chen, Singapore/Cambridge, privacy-preserving machine learning models that can work on encrypted data without exposing sensitive information.
Anton Leicht, Berlin, AI policy research.
Stablecoins competing with banks?
Another concern is that stablecoins could displace bank deposits and undermine commercial banks. Many banks are taking steps to “tokenise” their deposits, allowing them to be used easily for blockchain-based transactions. And they are reducing their fat fees on cross-border payments. In other words, faced with new competition, they are cutting costs and increasing efficiency. This is as it should be.
Here is more from Eswar Prasad at the FT. Or can the banks simply not compete and will lose lots of deposits?
Tuesday assorted links
Policy uncertainty matters right now
More than half of Fourth District business contacts said that “uncertainty and the potential impact on demand” was the most important factor influencing their capital expenditure plans for the rest of 2025. Contacts most often reported that their inventory levels were the same as they were one year ago, though more firms reported lower inventory levels than higher ones. More than half of firms anticipated that they would work through their current inventories in one to three months. Most survey respondents said that one-fourth or less of their current inventory had been subject to additional duties in 2025 because of changes to trade policy.
Here is more from the Cleveland Fed.
The polity that is Brazil
Yet perhaps the biggest reason spending is high is that the constitution requires it. The charter mandates an extraordinary 90% of all federal spending. Notably it ties most public pensions to wage growth, and requires health and education spending to rise in line with revenue growth. If Brazil were to end most tax exemptions and undo these two policies, its debt-to-gdp ratio, which is already above 90%, would be almost 20 percentage points lower by 2034 than it would be without any reform, reckons the IMF. To deal with all this, what is really needed is to amend the constitution.
High spending and a tangle of subsidised credit schemes also reduce the effectiveness of monetary policy. That means the central bank must increase rates even higher to control inflation. Brazil’s real interest rate of 10% is among the highest in the world. Such rates cripple investment and drag down growth, while well-connected businessmen can get their hands on artificially cheap rates.
Among those who must pay the full rate is the government itself.
And:
Tax exemptions total 7% of gdp, up from 2% in 2003 (see chart 2). Dozens of sectors receive tax breaks or credit subsidies on the basis that they are national champions, or from “temporary” help that has never ended. Brazil’s courts cost 1.3% of gdp, making them the second-most expensive in the world, with much of that going on cushy pensions and perks. Some $15bn a year, or 78% of the military budget, is spent on pensions and salaries. The United States spends just one-quarter of the defence budget on personnel.
Here is more from The Economist.
What determines business school faculty pay?
We examine the determinants of business school faculty pay, using detailed data on compensation, research, teaching, and administrative service. We estimate that a top-tier journal publication is worth $116,000, with significant variation across disciplines. Second-tier publications are worth one-third as much, and other publications have no impact. Further analysis of salaries and cross-discipline publication records suggests that researchers are compensated based on the journals they publish in rather than the departments they belong to. Conference presentations and teaching evaluations have significant but smaller effects than top-tier publications. Faculty administrators earn a premium, with department chairs receiving 11-35% and deans 58-94%. Post-Covid-19, real faculty pay has fallen more than in comparable fields and the sensitivity of pay to research performance has weakened.
That is from a new paper by Michelle Lowry, Daniel Bradley, April M. Knill, and Jared Williams. Via Arpit Gupta.
Monday assorted links
I podcast with Jacob Watson-Howland
He is a young and very smart British photographer. He sent me this:
Links to the episode:
Youtube: https://youtu.be/4nMg0Qg7KRI
Spotify: https://open.spotify.com/episode/2wDyCaXhGN5ruN1SNkcaBt?si=3c054eb0396f4787
Apple: https://podcasts.apple.com/gb/podcast/watson-howland/id1813625992?i=1000724051310
Jacob’s episode summary is at the first link.
What are the markets telling us?
That is the topic of my latest piece for The Free Press, as after all stock valuations are high. Excerpt:
First, the One Big Beautiful Bill Act, passed earlier this year, slashed corporate taxes. Before Trump’s first term, the corporate tax rate was 35 percent; in his first term Trump cut it to 21 percent, and this year he and the Republican Congress extended aspects of that first-term tax bill. Factors such as 100 percent bonus depreciation and expanded interest deductions give many companies the ability to lower their tax bill further, though not in a way that can be expressed readily by any single number.
With these lower tax burdens, companies should be worth much more. At the same time, the American taxpayer now owes more than $37 trillion in debt. So someone has to pay higher taxes over time, to satisfy those obligations. That someone probably is you, so you might want to take that into account in your overall assessment. But you should feel good about the companies, and somewhat less good about yourself and your children, given the tax hikes in the offing, sooner or later. You are paying for some of those higher stock market values.
Plus the ten or so percent decline in the value of the dollar can be seen as a loss of confidence in the United States, for a bunch of the obvious reasons. But will the country fall apart? Probably not:
That said, some of the worries are exaggerated. I’ve seen lots of posts on X lately saying that Trump is destroying the independence of the Federal Reserve system, and that this change will bring much higher rates of price inflation. Maybe. But that is not what the markets are saying. There are different measures of expected inflation, and most stand between 2 and 3 percent, with some of the more important market measures clustered near 2.5 percent. You might think that is higher than it should be, but it is hardly a sign of pending hyperinflation. It’s also not different from how things were toward the close of Joe Biden’s term.
And this:
Finally, are you worried about fascism coming to America? The collapse of democracy? Did you read about the Trump critics Jason Stanley, Timothy Snyder, and Marci Shore moving from America to Canada?
Well, head on over to the prediction markets—for instance, Kalshi. Currently the Democrats are favored at 53 percent to win the next presidential election. In other words, it costs 53 cents to buy a security that pays off a dollar if the Democrats win. In turn, that means the market thinks the Democrats have a very slight advantage in 2028. Does that sound like the collapse of democracy to you?
I return here to the same logic: If you think the market is being naive and foolish, why aren’t you placing your bets in the opposite direction? At the very least you could be wealthier under the forthcoming fascism you predict, and you might even be able to donate your winnings to antifascist causes.
Comments that fail to understand the distinction between optimal ex ante prediction and ex post outcomes should be shamed!
Sunday assorted links
Those new service sector jobs
Poets with Mercor, for $150 an hour. And that is just a start.
*Adolescence*
This is a British TV show, in four episodes, available on Netflix. It is the first TV show I have wanted to watch in about four years. I cannot review it without immediate spoilers, but I can tell you it takes place in a school, in a criminal justice system, and in a family. Northern England, so do turn on those subtitles. Here are some reviews.