Bass Pro Shops, Gorillaz, and Ping Pong(ing Congress)
What we’re reading, June 17, 2026
Here’s what caught our attention over the last week:
The driving around test (Euro edition) — Matt Clancy
You should read Dario Amodei's latest essay — Dylan Matthews
It's 2017 again at the NIH — Jordan Dworkin
40+ ways to plug in a data center — Willow Latham-Proenca
Should European housing politics be Americanized? — Alex Armlovich
Oops, I ordered the wrong antibody again — Saloni Dattani
Updates from Congress — Nisha Austin
The driving around test (Euro edition) — Matt Clancy
As part of the month-long debate about the extent of the European/American gap in living standards, Luis and Pieter Garicano suggested a “driving around test”: “Go to the periphery of any modern American city and see a level of new-built material wealth that is extremely uncommon in Europe.” With the World Cup underway in America, some European tourists are doing their own “driving around” tests and going viral, though these driving around tests are hardly scientific (the US may be wealthy, but free hotel rooms and meet-and-greets with our favorite singers do not come standard with road trips).
Nonetheless, for those worrying about the US-EU living standards gap, there’s plenty to read from the last few weeks. To start, Simon Johnson’s interview with economist Philippe Aghion covered a wide range of plausible factors explaining the gap: a lack of a single market, insufficient funding for long-run research, lack of DARPA-like agencies, a culture that doesn’t celebrate risk-taking, and excessive red tape. For more on this theme, see also Alex’s post below about European YIMBYs (or the lack thereof). While Aghion believes these factors are real, he pushes back when Johnson characterizes Europe as troubled:
I wouldn’t say “troubled.” Europe is not on the verge of a major crisis. Instead, the danger is a long period of stagnation or sluggish growth.
Unfortunately, the authors of Europe 2031 (the latest AI scenario microsite) worry that we are in an era when stagnation will impose heavy costs on Europe in the years ahead. One theme of the US-EU living standards debate has been on the importance of having a thriving domestic tech sector. So far, many of the benefits of the US tech sector have been shared by Europe, which, after all, has as much free access to Google and Facebook as it wants. The Europe 2031 authors paint a scenario where this assumption breaks down, and access to AI becomes nationalized as its security implications grow and compute for inference remains scarce. The microsite’s timing is eerie, coming shortly before the US government banned access to Anthropic’s latest model for foreign nationals (though, at least in this case, the ban meant that US citizens also lost access).
You should read Dario Amodei’s latest essay — Dylan Matthews
If you’re staying on top of AI news, you probably don’t need me to tell you to read Anthropic CEO Dario Amodei’s latest essay. That said … you should read Dario Amodei’s latest essay. It’s the most detailed look at the company’s public policy attitudes to date, and particularly interesting to the AGF team because of how much Amodei focuses on clinical trials. As he writes, AI seems likely to "greatly increase the rate at which new drug candidates enter the regulatory pipeline"; but that pipeline is so long and slow that it will be hard for customers to enjoy those drug candidates in a timely way. That implies that we need major reforms to the way the Food and Drug Administration and European Medicines Agency evaluate drug candidates, to make sure new, promising therapies aren't languishing for years before approval.
It’s 2017 again at the NIH — Jordan Dworkin
It’s 2017 again: the Gorillaz are topping charts, Star Wars is in theaters, and the NIH is discussing a potential cap on the number of grants a single PI can hold at once. Last week the agency published an RFI floating potential caps at two, three, or four simultaneous research project grants, with the stated goals of spreading funding more widely across labs, institutions, and researchers, and supporting more early- and mid-career scientists.
The agency’s last attempt at this – the 2017 Grant Support Index, which would’ve capped support at roughly three R01s’ worth of funding – was dropped within a month following pushback from elite universities. Last week’s RFI cites a handful of metascience papers in support (specifically showing diminishing marginal returns to additional funding, and reduced disruption as a function of team size), though in my view the metascientific jury is still out on the cost-benefit of a hard cap; getting more money to younger scientists seems like the strongest case for the cap, but cases against include diverting funding from meritorious proposals, increasing process burden, and making large-team science more difficult to sustain. Since 2017, the agency has tried a few other approaches for achieving these goals, including a dedicated fund for early- and mid-career scientists (which directly replaced the GSI when it was dropped) and a softer version of a cap at NIGMS, in which explicit justification is required for funding above a certain threshold. STAT has a good overview of the background and the various perspectives. Responses to the RFI are due August 3.
Should European housing politics be Americanized? — Alex Armlovich
"Should European housing politics be Americanized?" asks Samuel Hughes. American policy frames usually map badly onto European debates, but housing is an exception: Europe's housing shortages are arguably worse than America's, with European house prices now above US levels and roughly 80% of the postwar increase attributable to regulatory restrictions on building. Despite that, YIMBYism is effectively nonexistent on the continent, and housing arguments often default to rent control, zero-sum expropriation without new construction, and public housing. Hughes's sharpest move is historical: zoning was arguably invented in Europe, and he walks through Berlin's 1905 villa-colony scheme to show the same map performing the same exclusionary function 120 years later. I went in a little nervous about how a British writer would explain American YIMBYism abroad, but Sam’s argument was novel and deftly executed.
40+ ways to plug in a data center — Willow Latham-Proenca
It may not be the most scintillating reading for the general audience, but the Lawrence Berkeley National Lab is out with an extremely cogent, clear rundown of over 40 proposals to speed up how fast large loads – data centers, of course, but also industrial users or fleet EV charging – can connect to power. These range from the frequently-discussed (BYOG, interruptible service) to the arcane (ERIS-like service for loads, one of several proposals to adapt tools from the generation interconnection toolkit to the load side). As usual, the national lab that brought you quantitative evidence that new load doesn’t (necessarily) increase prices is calibrated on how much to panic – they compare load forecasting errors from the 1970s over-forecast to now, and find we might actually be doing better. Worth noting that proposals to speed up construction of the transmission needed to connect many of these loads to electrons – which can take up to three times as long as the planning and study process, in the report’s estimation – are out of scope here.
Oops, I ordered the wrong antibody again — Saloni Dattani
The science sleuth Sholto David recently uncovered a mix up that’s affected hundreds of papers in biology. Scientists who were intending to work with the p16-INK4a antibody, to flag a tumor suppressor linked to cancer and aging, accidentally ordered another antibody p16-ARC instead, which tags a component of the cell’s internal skeleton.
Some of the consequences were mostly harmless, where scientists ordered the correct antibody but described the wrong one in their papers. Others were unfortunate, as lots of experiments may not have worked because of the mistaken antibodies. But some are potentially fraudulent, where in some experiments, scientists reported results using the wrong antibody that could have only worked if they were using the right ones.
The whole naming confusion reminds me of another related episode in biology, where hundreds of papers were affected by an error in which the gene SEPT1 got automatically converted into the date ‘1st September’ by Microsoft Excel. Eventually, the Gene Nomenclature Committee solved the problem by simply renaming the gene SEPTIN1. Let this be a warning about the dangers of dating.
Updates from Congress — Nisha Austin
The first major federal housing bill in roughly 30 years just took a big step forward. After months of ping-ponging between the House and Senate, all four corners (the chairs and ranking members of the Senate Banking and House Financial Services committees) released unified bill text yesterday for the 21st Century ROAD to Housing Act. Senate leadership is expected to bring it to a floor vote this week. We’ll be watching.
On energy permitting, there are two bipartisan moves worth keeping an eye on. Senators Cotton and Cortez Masto introduced a Senate version of the FREEDOM Act, which would prevent agencies from revoking permits or halting construction on fully permitted energy projects, and is a direct response to the administration’s stop-work orders on offshore wind. The Senate version rolls in much of the House’s geothermal package from earlier this month and is notably cleaner than the House’s: it strips the most stringent agency penalties and moves fees into a freestanding “Permitting Performance Fund,” which should quiet concerns that the original structure would create a cost spiral for agencies. Meanwhile, Senator King added an amendment to the NDAA targeting the DOD’s increasingly creative use of flight path certification to block onshore wind projects. Neither energy bill is a sure thing, but both signal that permitting certainty has bipartisan legs even in this environment.
Here are a few other highlights and announcements from our team and grantees:
Jordan published Our grandchildren’s AI-science bottleneck, on why the fields AI accelerates least may end up determining the pace of discovery for everything else.
Matt published a blog post about the Atlas of Innovation, an interactive guide to choosing between prizes, grants, contracts, and other tools for funding innovation, developed with IFP and the Market Shaping Accelerator.
Saloni wrote up What’s new in biology for Works in Progress, rounding up recent developments in the field, and covering everything from antibody mix-ups to new cancer research.
Abundance New York launched a voter guide ahead of the city’s elections.
California YIMBY endorsed Xavier Becerra for governor.