The ROAD Act's 124 to-do items, space mirrors, and cancer vaccines (maybe)

What we’re reading, July 15, 2026

Here’s what caught our attention over the last week:

1. 124 implementation actions and counting — Alex Armlovich

2. 200,000 comments and the state of US science — Jordan Dworkin

3. Does upzoning work? — Matt Clancy

4. Designing a cancer vaccine — Saloni Dattani

5. Western transmission and DOE needs — Willow Latham-Proenca

6. Space mirrors — Dylan Matthews

Rendering of Reflect Orbital's proposed constellation of satellites. Source: Reflect Orbital

124 implementation actions and counting — Alex Armlovich

Aaron Shroyer and Sarah Brundage at NAAHL have the essential federal implementation guide for the ROAD Act. None of the law's ~47 housing sections are wholly self-executing: NAAHL counts 124 discrete implementation actions across eleven agencies, with 58 due within the first year. HUD owns 88 of them, including nine rulemakings (environmental review streamlining, the chassis definition, offsite construction financing, HOME reform). It might take awhile: HUD finalized roughly four significant rules per year last term, and its current agenda already has 54 rules queued. The staffing math is worse: the three HUD offices holding three-quarters of the workload are down 27–30% from FY25, and the General Counsel's office that must clear everything is down 35%. Meanwhile a dozen new programs (pattern-book grants, the Innovation Fund, PRICE, single-stair Point-Access Block grants) exist only on paper until Congress appropriates funding, and several sunset within three to seven years…so every year of delay compresses their operating window. Think tanks and university research specialists in housing should be ready to step up and help HUD’s limited headcount move these research, guidance, and rulemaking procedures forward.

200,000 comments and the state of US science — Jordan Dworkin

This Monday was the deadline for public comments on OMB’s proposed rewrite of the rules governing federal grantmaking. The response was immense, with more than 200k comments submitted over the past few weeks. As expected, responses poured in from scientists, universities, societies, and publishing outfits. But because of the scope of the OMB proposal (it affects over $1 trillion in annual federal assistance), concern came from well beyond science, with the National Association of Counties filing its own comment, and the National League of Cities and Moody’s raising the potential negative impacts on project financing and credit for state and local borrowers. In Marginal Revolution, Alex Tabarrok highlighted broader, increasing alarm from organizations and thinkers not inclined towards hyperbole.

For Tech Policy Press, Chris Marcum and Abigail Haddad analyzed the ~50,000 comments posted by last week, and found roughly 94% opposed and 6% in support, with opposition largely grassroots (only ~1/8 of opposing comments came from a form letter) and comments in support almost entirely organized (nearly 9/10 supportive comments were from a single form-letter campaign). The dominant objections were the politicization of grant decisions and review, and the expanded termination authority. OMB has said it wants to finalize the rule by October 1 and has declined requests for an extension of the comment period. But litigation seems likely.

Even in the absence of the new rules, US science is continuing to strain under the pressure of this administration. NSF is on track for its worst grantmaking year since 1985 (~2,200 awards so far in FY26 compared to an FY21-24 average of almost 6,000); PhD admissions across top research universities fell 15% this year after a drop of 11% last year, with Caltech cutting its incoming class by 40%; and last week NSF moved to ban collaborations with nearly all Chinese research institutions, dropping an existing risk mitigation approach (TRUST) in favor of a flat prohibition.

Does upzoning work? — Matt Clancy

Does upzoning lead to new homes? Does it reduce the cost of homes? Michael Wiebe has an accessible review of the academic literature around upzoning in his living literature review, examining empirical studies on major upzoning cases in Auckland, Sao Paolo, Wellington, New York, and Zurich. Upzoning consistently increases the housing supply, and when studies are set up to measure the impact on prices, it brings costs down (at least relative to what would have happened without upzoning - sometimes that just means prices increase less fast). But the link between upzoning and these outcomes is a complicated one (as Alex has written about before); upzone in a place where building new housing requires a lot of demolition, and you’ll need to upzone by a larger amount to see a comparable effect.

Designing a cancer vaccine — Saloni Dattani

The great blogger known as owlposting has an awesome new podcast episode with Alex Rubinsteyn and Ben Vincent, two scientists who work on cancer immunotherapy, on ‘How to design a cancer vaccine’. The general idea behind a cancer vaccine is to prevent cancer cells from evading the immune system by training patients’ own immune cells to recognize them and kill them off.

Unfortunately, many past attempts to develop personalized cancer vaccines have failed. This is in part because they were designed based on the genetic profiles of cancer cells, rather than the actual proteins those cancer cells expressed on their surface. Measuring which proteins were actually presented would typically require samples to be collected differently than they usually are (fixed with formalin, rather than frozen) and the measurement was expensive and, in older versions, not very accurate. There were other problems too: mRNA vaccines were a new platform and faced their own regulatory scrutiny and immune cells were hard to steer well. They tend to over-respond to just a few target mutations, which might distract immune cells from the real ones, or they might be set loose on the wrong ones, which could be dangerously toxic.

But now, things might be different. Prediction and measurement have gotten better, the mRNA platform has been tested in billions of people, and there’s a clearer regulatory path for personalized drugs. I really enjoyed the episode and feel cautiously hopeful. It’s also worth listening to hear about how weird cancer cells are: they can, for example, reawaken ancient viral DNA buried in our genomes or even switch on genes that are normally only active in the sperm or placenta. You can listen here.

Western transmission and DOE needs — Willow Latham-Proenca

It’s been a busy few weeks in transmission – at the close of June, 11 western states agreed to improve coordination and permitting for transmission across their shared footprint (much of the western grid outside of California is operated by a patchwork of utilities, unlike the northeast and midwest, where regional transmission organizations plan and operate the grid). The pact follows a study by the Western Transmission Expansion Coalition earlier this year, which identified more than 12,600 miles of new or upgraded transmission needed in the region over the next decade, a quarter of which are yet to be even planned for.

The DOE also released its most recent draft report on national transmission needs. It’s no surprise that high-voltage interregional and cross-interconnection links continue to be massively underbuilt, with consequences for both energy prices and grid reliability - for example, the study found transmission congestion costs reached $12 billion in 2024. The draft also highlights another geography where transmission is critical – DOE estimates that the southeast (another non-RTO zone) is missing out on about $10/MWh in congestion relief value – essentially, cheaper power can’t get to where it needs to go – as a result of underbuilt connections to neighboring regions.

Space mirrors — Dylan Matthews

Solar panels only generate power when the sun is shining, and everyone knows you can’t make the sun shine on demand. What the startup Reflect Orbital proposes is: maybe you can? Cofounded by a SpaceX veteran, the company proposes to launch satellites with targetable mirrors that can redirect sunlight to the dark side of the earth. The main idea is to produce more frequently-available solar generation, but the company also proposes using the technique to assist in search and rescue and other disaster response.

The proposal is taking a step forward, with the FCC approving a demonstration satellite. This piece from Hiroko Tabuchi does a good job of laying out the appeal of the idea (plentiful solar power!) and the real possible downsides, like disruptions to astronomical observations and animals’ circadian rhythms.

Here are a few other highlights and announcements from our team and grantees:

• The ROAD to Housing Act was signed into law — the first major federal housing legislation in roughly 30 years. Alex’s blurb above covers what comes next on implementation.

• Caleb Watney published A Long Sequence of Small, Correct Decisions, announcing his new role at Coefficient Giving leading a team focused on strengthening the institutions that will navigate the transition to powerful AI, and overseeing the Abundance and Growth Fund and CG’s government relations.

• Matt published the Q2 2026 progress report, a look back at the fund’s work over the last quarter.

• California YIMBY published the latest Homework newsletter, highlighting YIMBY-endorsed candidates advanced in 17 of 19 competitive open seat races in recent elections, a strong showing for Open New York and California YIMBY.

• California YIMBY also published a summary of Brian Potter’s working paper on the divergence of housing prices from construction costs across US cities.