IFP Update: December 2025

📣 Announcements

• We’re hiring for our editorial team. If you’re an eagle-eyed editor with a passion for counterfactual policy impact, apply here. The deadline is January 19th, with a minimum salary of $110,000 (and potential for more based on experience). DC or NYC preferred, but open to remote candidates. Remember: We are offering a $3,000 bounty to whoever refers our hire.

• The deadline to apply for our PhD-level course on the Economics of Science, Ideas, and Innovation is today. Apply here.

• Will Poff-Webster joined as Director of Infrastructure Policy. Will previously worked on the Hill in Senator Schatz’s office on infrastructure and technology issues.

• Olivia Jimenez joined as Communications Manager. Olivia previously worked on talent sourcing and special projects for the AI Security Institute in the UK.

• Arthur Tellis joined as Senior Technology Fellow. Arthur previously worked on the Senate Armed Services Committee and at the Pentagon on cost assessment and program evaluation.

• Arnab Datta has been promoted to Director of Policy Implementation, taking on a larger role working across our policy teams.

✍️ Published Work

• Should the US Sell Hopper Chips to China? Saif Khan, Tao Burga, Tim Fist, and Georgia Adamson analyze the implications of exporting NVIDIA’s H200 chips to China. The H200 is nearly 6x as powerful as the H20, which already requires an export license. They find that unrestricted H200 exports would shrink America’s AI compute advantage from 21-49x down to 6.7-1.2x.

• How to Accelerate the American Scientific Enterprise. In response to OSTP’s request for information, Andrew Gerard, Matthew Esche, Dan Turner-Evans, and Connor O’Brien submitted five recommendations for accelerating American science. They argue for reducing visa delays for STEM researchers, empowering program officers like DARPA does, and creating agency-based metascience units.

• Proxy Praxis: How Surrogate Endpoints Can Speed Drug Development. Ruxandra Teslo launches a new series examining how surrogate endpoints (measurable biological signals that predict clinical benefit) can dramatically reduce the time and cost of clinical trials. She explores why the system is backwards: new, better endpoints face steep barriers while old, weak ones persist due to precedent.

📰 Media

• IFP’s H200 export analysis was widely cited in news coverage of President Trump’s decision to allow chip sales to China: The Wall Street Journal, The New York Times, The Washington Post, Bloomberg, Reuters, Politico, The Atlantic, CNBC, and The Economist.

• STAT News mentioned our work on clinical trial abundance:

  • “The Institute for Progress, for example, has been championing what it calls ‘clinical trial abundance’ — a pragmatic reform agenda to make trials cheaper, faster, and more numerous. Their case for abundance lays out a vision where the U.S. embraces iteration rather than fearing it, returning clinical research to something closer to the scientific method: run, learn, adjust, repeat.”

• The Washington Post cited our work on H-1B reform:

  • “The Economic Innovation Group and the Institute for Progress, both pro-innovation, pro-growth policy research organizations, have proposed replacing the lottery with a straightforward compensation-based ranking for applicant selection.”

• And Bloomberg Law quoted Jeremy Neufeld:

  • “Jeremy Neufeld, director of immigration policy at the Institute for Progress, said the rule will disrupt employers’ recruitment pipeline at top US universities since those graduates will now have a decreased chance in the lottery. Those students typically get paid more than other H-1B applicants but are registered at a lower wage level, he noted.”

🏛️ Statecraft by Santi Ruiz

• 99.8% of Federal Employees Get Good Performance Reviews. Why? Santi interviews Trump’s head of talent about the federal workforce. The conversation covers the 300,000+ people leaving government through the Deferred Resignation Program and Reductions in Force, and the problems with using headcount alone as a metric.

• Did the CHIPS “Everything Bagel”...Work? Santi talks with members of the original CHIPS leadership team about how they got all the deals done, and whether the much-discussed “everything bagel” requirements actually slowed things down. The takeaway: “Nobody on the team thought we were going to get all these deals done.”

🏗️ Construction Physics by Brian Potter

• Stagnant Construction Productivity Is a Worldwide Problem. Brian looks at construction productivity trends across countries using EU KLEMS data and finds that stagnant productivity isn’t just a US problem. Even sustained, large-scale building programs like China’s or Sweden’s factory-based construction haven’t changed this.

• How Bell Labs Won Its First Nobel Prize. Brian tells the story of physicist Clinton Davisson, who won the 1937 Nobel Prize for demonstrating electron diffraction while working at AT&T, before Bell Labs formally existed. He explores what this early research tells us about the culture that made Bell Labs so successful.

• How Accurate Are Learning Curves? A new paper analyzes 87 technologies and finds that learning rates in the first half of each curve don’t correlate well with those in the second half. The core insight (that costs fall with cumulative production) still holds, but the specific rates are far less predictable than commonly assumed.

• Should US Homebuilders Emulate Sweden? Brian examines the common argument that US housing should adopt Sweden’s 85% prefab rate. Despite impressive factories, Swedish construction costs have risen at about the same rate as US prices since 1995, and the average Swedish home costs over 70% more per square foot.

🏭 Factory Settings by Mike Schmidt, Todd Fisher, and Sara Meyers

• The CHIPS Investment Process: Move Fast and Break Nothing. Sara O’Rourke, former COO of the CHIPS Investment Office, explains how the team redesigned the typical government grant process to emulate private equity. The new approach included optional pre-applications with formal feedback, letting CHIPS push companies toward more ambitious projects. By the end, they had allocated $34 billion across 20+ projects.

• An Inside View of NEPA in Practice. Mike Schmidt details how the National Environmental Policy Act created significant risks of delay and litigation for CHIPS. Environmental assessments typically take 9-12 months, while full impact statements average 4.4 years. Congress ultimately passed the Building Chips in America Act in October 2024 to exempt most projects.

• How CHIPS Executed on an Ambitious but Vague Mandate. Todd Fisher explains how the team translated the CHIPS Act’s broad mandate into an executable strategy. Four early decisions shaped success: using rolling applications, proactively allocating capital across industry segments, establishing a 5-15% funding range anchored in IRR analysis, and adding upside-sharing provisions to close information asymmetries with companies.

🔬 Macroscience by Andrew Gerard

• To Get More Effective Drugs, We Need More Human Trials. Ruxandra Teslo and Jack Scannell argue that debates about pharmaceutical productivity focus too much on basic science and FDA approval while ignoring clinical development. They make the case for “Clinical Trial Abundance,” noting that Australia completes Phase I trials 40-50% faster than the US.

• How Bad Is It When the Government Cancels Active Research Grants? Andrew Gerard examines the consequences of the Trump administration terminating thousands of in-progress university grants. He outlines the risks (reduced research quality, harm to the PhD pipeline, weakened university finances) and proposes resilience measures including Congressional fixes like automatic payouts for terminated grants.

👋 Tweet for the Road

Caleb Watney@calebwatney

chatter on the street

6:39 PM · Dec 19, 2025 · 6.3K Views

---

3 Replies · 5 Reposts · 40 Likes