In this episode of All-In with Chamath, Jason, Sacks & Friedberg, Michael Kratsios and David Friedberg examine the state of American science and research. They discuss how U.S. science has become politicized over recent decades, particularly during the Covid-19 pandemic, and argue for a return to foundational scientific principles based on open debate and empirical evidence. The conversation covers concerns about federal research funding allocation, including how current grant processes may discourage bold innovation and how DEI priorities have affected funding decisions.
Kratsios and Friedberg also address the U.S.-China competition in scientific output, noting China's substantial increase in R&D spending and research publication. They explore challenges in the domestic STEM pipeline, including declining numbers of U.S.-born PhD recipients in computer science and barriers facing early-career scientists. The episode presents various reform proposals, from alternative funding models to mission-driven targets in fields like space exploration and fusion energy, aimed at restoring scientific productivity and maintaining America's competitive position.

Sign up for Shortform to access the whole episode summary along with additional materials like counterarguments and context.
Michael Kratsios and David Friedberg discuss the urgent need to depoliticize American science, reform federal research funding, and strengthen the nation's competitive position against China's rising scientific capabilities.
Kratsios and Friedberg argue that U.S. science has become deeply politicized over the past 15–20 years, moving away from its foundational principles of inquiry and skepticism. This trend peaked during the Covid-19 pandemic, when authorities like Dr. Anthony Fauci promoted policies such as student masking that weren't always supported by empirical data. Disagreement was frequently labeled as anti-science, turning scientific debate into a matter of loyalty. Friedberg notes that many scientists stopped asking hard questions and defaulted to dogma, while those who challenged conclusions on Covid-19 measures or climate modeling scenarios like RCP 8.5 were dismissed as conspiracy theorists.
To restore trust, Kratsios and Friedberg call for a return to gold standard scientific methods based on questioning, experimentation, and humility. Scientific authority should depend on open debate, not conformity. Kratsios also highlights how political ideology around DEI has affected funding, with about 25% of NSF grants over four years—roughly $8 billion—going to DEI-related projects, which he views as diverting research priorities away from scientific merit.
Despite federal biomedical research funding tripling since 1998—from $14 billion to $47 billion—scientific productivity per dollar has fallen roughly 80-fold since 1950, illustrating what Friedberg calls Airthur's Law: research efficiency halves every nine years. Kratsios identifies the core problem as a lack of innovation in how federal science is managed and financed.
The NSF and NIH grant processes discourage bold, high-risk research, with peer reviewers favoring safe, conventional projects. Grant durations are often mismatched to actual project needs, stifling both quick experiments and long-term ambitious work. In response, the administration is piloting reforms including "golden tickets" that allow reviewers to unilaterally back innovative proposals, varied grant durations from six-month sprints to five-year projects, and funding scientists based on talent rather than institutional affiliation.
Kratsios suggests supplementing traditional federal funding with alternatives, noting that the funding landscape has reversed since World War II—private sector investment now accounts for 70% of R&D versus 30% federal. He advocates for diversified allocation models that redirect university overheads to research and support independent scientists, ensuring funding follows the best researchers regardless of location.
China's R&D spending rose from $33 billion in 2000 to $670 billion in 2021—a 19-fold increase compared to the U.S.'s threefold increase. Friedberg notes that China now publishes roughly 50% more scientific papers than the U.S. across nearly all domains except some life sciences.
Kratsios contrasts China's centralized, top-down approach—which has struggled to develop EUV lithography despite making it a national priority—with the U.S. market-driven model, where competition among government, universities, industry, and investors fosters diverse innovation paths. He describes China's industrial strategy of identifying technologies, creating cheaper alternatives, dumping them in U.S. markets, and establishing dominance—particularly targeting control over critical technology minerals. The U.S. must ensure independent supply chains through diversified sourcing and domestic production to avoid exploitative dependencies.
Kratsios observes that the percentage of U.S.-born PhD recipients in computer science has inverted over three decades: once 70% American, now only 30%, with seven out of ten STEM PhD candidates not being American citizens. This shift reflects policies that discourage advanced STEM education, including the elimination of NSF's gifted-student programs, removal of advanced math courses in secondary education, and UC's decision to drop SAT requirements—which has led to more Berkeley students needing remedial math.
Early-career scientists face low pay, uncertain prospects, and grant pressure. Kratsios highlights efforts like NSF's GRFP, which funds 2,600 promising PhDs annually with portable funding, making universities compete for talent. He argues the U.S. must both rebuild the domestic STEM pipeline and create legal pathways for international scientists trained in the U.S. to stay, as other nations now actively recruit global talent with attractive salaries and housing.
Kratsios notes that today's R&D funding ratio—70% private, 30% government—reverses the post-WWII model. He argues government should fund early-stage, pre-competitive discovery science, while private industry focuses on commercialization. Friedberg points out that private industry prioritizes near-term revenue, leaving fundamental sciences like deep space exploration and pure physics underfunded, despite their potential for unforeseen future applications.
The administration has set ambitious targets: American boots on the moon by 2028, a nuclear reactor in space by 2028, relevant quantum computing by 2028, and fusion energy by 2035. The "Genesis Mission" aims to use AI to double U.S. scientific output in materials science, chemistry, math, and physics. Kratsios advocates a balanced portfolio approach combining mission-driven investment with robust funding for basic discovery science.
He emphasizes that funding should prioritize merit over institutional prestige, fostering redundancy and resilience across the scientific community. Currently, funding too often tracks the institution rather than the individual, limiting opportunities for talented researchers outside elite organizations. Kratsios also notes that the median age of NIH researchers is now 71, indicating a lack of advancement pathways for early-career scientists. He sees an opportunity for institutional reforms that support younger scientists and promote generational mentorship, which will revitalize American science with fresh perspectives and innovative directions.
1-Page Summary
Michael Kratsios and David Friedberg argue that over the past 15–20 years, science in the United States has become deeply politicized, moving away from its foundational principles of inquiry and skepticism. This trend peaked during the Covid-19 pandemic, when public and scientific authority figures—exemplified by Dr. Anthony Fauci—began promoting policies and positions, such as student masking in schools, that were not always supported by empirical data. Disagreement with these positions was frequently labeled as anti-science, undermining the very basis of scientific inquiry, which relies on continuous questioning, hypothesis testing, and debate.
Kratsios highlights that Dr. Fauci and other authorities turned scientific disagreement into a matter of loyalty or faith, where questioning guidelines or policies was equated to opposing science itself. This led to confusion and a collapse of trust, as the public witnessed conflicting standards, such as allowing mass protests while barring people from visiting dying relatives. Friedberg adds that many scientists themselves stopped asking hard questions and defaulted to dogma, aligning behind conclusions without sufficient empiricism. People who challenged these conclusions, whether on Covid-19 measures or subjects like vaccine skepticism or specific interpretations of climate change, were frequently dismissed as conspiracy theorists, anti-science, or politically motivated outcasts.
Friedberg extends this critique to climate research, citing the now-removed RCP 8.5 scenario used in climate modeling. Although this extreme scenario lacked credibility and was ultimately dropped from formal predictions, it nonetheless dominated media coverage and influenced public perception and funding for years. According to Kratsios and Friedberg, anyone publicly questioning or criticizing the use of RCP 8.5 would be attacked as anti-science. The focus on such extreme, ideologically driven scenarios did a disservice to scientific integrity, distorting both the research agenda and public understanding.
Kratsios and Friedberg call for a return to gold standard scientific methods based on questioning, experimentation, data collection, iteration, and humility. The scientific process, Friedberg notes, depends on continuous inquiry—a cycle of asking questions, empirically testing them, collecting data, and refining conclusions. Scientific authority should never depend on conformity or position but on open debate and willingness to reconsider prior understanding.
Adhering to humility means recognizing the provisional nature of scientific knowledge and embracing debate rather than suppressing dissent. Friedberg argues that true scientific progress depends on letting ideas emerge and compete fairly, not on ideological alignment. A healthy scientific community welcomes disagreement as part of the method, with objectiv ...
Politicization of Science and Restoration of Integrity
The U.S. has dramatically increased its federal investment in biomedical research over recent decades, yet concerns are mounting over diminishing returns, inefficiencies, and a lack of transformative innovation. Current funding models and their institutional constraints have led to declining scientific productivity, signaling the need for reform in how federal research dollars are allocated.
Federal biomedical research funding has more than tripled since 1998. NIH funding rose from $14 billion in 1998 to $27 billion in 2003, reaching $47 billion by 2024. Despite this surge in investment, the number of breakthrough treatments has not increased proportionally. In fact, as David Friedberg points out, scientific productivity per dollar spent has fallen remarkably—by roughly 80-fold since 1950. This trend is captured in Airthur's Law (named analogously to Moore’s Law, but in reverse), which posits that research efficiency halves every nine years. Even as federal funding grows or remains constant, the expected scientific returns diminish, highlighting critical inefficiencies in the system.
Michael Kratsios identifies a major issue: a lack of innovation not in the science being done, but in how federal science is managed and financed. Rather than rethinking funding structures, federal agencies like NIH and NSF have continued to operate as they always have—simply adding more money in hopes of proportionally bigger outcomes. Questions are rarely asked about alternative methods for conducting and funding science, or opening funding up to different kinds of scientists and institutions. This inertia stifles both creativity and efficiency in federal scientific research.
The grant process at key agencies like NSF and NIH reinforces cautious, conventional research over bold, high-risk proposals. Grant applications undergo peer review, but reviewers often favor projects falling within a "strike zone"—ideas likely to be approved and deliver reliable, low-risk results. As Friedberg explains, researchers depend on grant funding not just for their work but also for their salaries and labs, so they are incentivized to propose safe projects more likely to receive approval. This makes transformative or out-of-the-box science less likely to be funded.
Typical grant durations—usually about 18 months—are often mismatched to actual project needs, stifling both short, high-velocity experiments and long-term, ambitious projects. This further pressures scientists to conform to the system's limitations rather than pursue transformative research that might require more custom timeframes or risk tolerance.
In response to these stagnating models, the NSF is piloting initiatives to make the system more supportive of unconventional, high-potential science. One such experiment is the introduction of "golden tickets": peer reviewers are given a limited number of tickets they can use to unilaterally back a proposal of their choice without consensus from the review committee. This mechanism, which has shown promise in Denmark and other countries, aims to encourage reviewers to select more innovative or unconventional projects and to attract higher quality reviewers to the process.
Grant programs are also experimenting with varied funding durations. New models allow for fast-tracked, six-month "sprints" for quick-turnaround projects, as well as multi-year grants (up to five years) for more complex work. The intention is to let scientists propose the duration that best matches their research, allowing for increased flexibility and more meaningful progress.
Another shift under the new vision is to fund scientists based on their talent, potential, and merit rather than their institutional affiliations. For example, programs like the NSF’s Graduate Research Fellowship Program (GRFP) are structured so recipients can take their funding to any institution, fueling competition and allowing researchers to work where they are best supported. Friedberg advocates for another alternative, likening it to venture capital: provide talented scientists with substantial, long-term resources and the freedom to pursue their own ideas, trusting that significant breakthroughs will emerge from their independence.
Historically, following World War II, the U.S. ...
Reforming Federal Research Funding Allocation and Improving Returns
The competition between the United States and China in scientific output and technological leadership has intensified over the past two decades. China’s rapid investment in research and development (R&D), its growing volume of scientific publications, and its aggressive industrial practices are spurring the U.S. to re-examine its own innovation ecosystem, industrial policy, and supply chain strategies.
David Friedberg highlights the dramatic surge in China’s investment in R&D, rising from $33 billion in 2000 to $670 billion in 2021—a 19-fold increase. In the same period, U.S. R&D spending increased only about threefold. This massive increase signals China’s determination to expand its scientific and technological capacity and challenge U.S. leadership in these areas.
Friedberg further notes that a decade ago, the U.S. published about twice as many scientific papers as China. Now, China publishes roughly 50% more scientific papers than the U.S. across nearly all domains, with the exception of some life sciences. The credibility of these publications—many of which are peer-reviewed—suggests China is making real breakthroughs and, in some areas, establishing a lead over the U.S.
The transformation in scientific publishing reflects broader changes in technological and economic competition between the two countries. As China increases its output and quality in science, this shift mirrors China’s ambitions in global economic and technology landscapes.
Michael Kratsios contrasts the U.S. and Chinese approaches to scientific innovation. He notes that China’s top-down, centralized scientific model allows a single agency to set priorities and allocate resources. However, this model faces serious limitations, as shown by China’s struggle to develop EUV lithography technology since the U.S. imposed export controls in 2019. Despite making this a top national priority, China has so far been unable to achieve a breakthrough in this crucial technology.
Kratsios emphasizes that the U.S. market-driven approach, where competition among government agencies, universities, industry, and venture investors generates diverse paths to innovation, prevents the system from becoming entrenched or stagnant. He views this free-market ecosystem—supported by vibrant government funding, active industry involvement, and dynamic venture capital—as a major driver for U.S. breakthroughs. Kratsios points out that organizations like DARPA, with a clear alignment to national security mission requirements and operational independence from a single science agency, are better able to achieve mission-relevant research and development. This structure has historically yielded significant and relevant technological advances.
U.S.-China Competition in Scientific Output and Technological Leadership
Discussion by Michael Kratsios and David Friedberg highlights deep concerns about the weakening pipeline for STEM talent in the U.S., the policies that have discouraged Americans from pursuing advanced STEM, and the failure both to incentivize domestic scientists and to retain international talent trained in the U.S.
Kratsios observes a striking reversal over the past 30 years in the percentage of U.S.-born PhD recipients in computer science: while 70% were American and 30% foreign three decades ago, the trend has now inverted, with only 30% being American and 70% foreign-born.
Kratsios highlights this shift, noting that seven out of ten STEM-related PhD candidates in the U.S. are not American citizens. This marks a significant gap in cultivating and retaining domestic scientific talent, despite the country’s longstanding tradition of investing in education.
The loss of American interest in pursuing STEM, coupled with the growing reliance on foreign talent, raises concerns about national vulnerability in science and technology. As more international candidates fill PhD spots, the U.S. risks losing its long-term scientific and technological edge.
Kratsios recounts how the National Science Foundation previously ran programs to encourage gifted and talented students in high schools and middle schools to pursue STEM. The discontinuation of these programs cut off an important source of future American scientists.
He also points to trends, particularly in California, where school systems have eliminated advanced courses like algebra for high school students. This restricts opportunities for high-performing students to develop foundational math skills critical for scientific careers.
Friedberg notes that the University of California’s decision to drop SAT requirements has led to more students needing remedial math at prestigious universities like UC Berkeley. Previously, top math SAT scores were essential for admission, ensuring a higher baseline of preparedness for rigorous scientific training.
Kratsios describes early-career scientists as among the most underpaid for the value they provide. Young researchers finishing PhDs enter academic careers with low salaries, considerable pressure to secure grants, and uncertain long-term job prospects.
Efforts to address this, such as the National Science Foundation’s Graduate Research Fellowship Program (GRFP), fund about 2,600 of the country’s most promising aspiring PhDs each year. These fellowships provide portable funding, making universities compete to attract awardees and enabling researchers to choose environments where they can excel.
Strengthening the Stem Pipeline and Attracting Talent
Current trends in science funding in the United States highlight a pressing need to revisit government priorities, with a special emphasis on supporting early-stage research and diversifying who and what gets funding.
Michael Kratsios notes a dramatic shift in the funding landscape for research and development (R&D) over the past 70 years. Today, roughly 70 percent of R&D is financed by the private sector, while just 30 percent comes from federal government sources. This stands in stark contrast to the post-World War II era when the government was the primary funder. Kratsios argues for a return to government-led support for early-stage, pre-competitive research, especially discovery science. These are areas where private sector incentives are weak or nonexistent, making it an essential role for public investment.
David Friedberg points out that while hundreds of billions of dollars are now flowing into fields with obvious commercial applications, like AI, quantum computing, or life sciences, private industry naturally focuses on research likely to generate near-term revenue. This leaves fundamental sciences, such as deep space exploration, the origins of the universe, and pure physics, underfunded. History demonstrates, Friedberg notes, that advances in these disciplines often yield applications years later that cannot be foreseen in the present.
Kratsios describes an ideal model where the government funds the discovery phase of science, which creates foundational knowledge, while the private sector takes the lead on commercializing these discoveries. This approach leverages the strengths of both public and private sectors, ensuring society benefits from fundamental science that industry alone would neglect.
Michael Kratsios outlines ambitious federal goals: American astronauts back on the moon by 2028, a nuclear reactor in space by 2028, the first elements of a moon base by 2030, a scientifically relevant quantum computer by 2028, and commercial fusion energy by 2035. Achieving these targets demands multiyear, coordinated government action and investment.
Kratsios highlights the “Genesis Mission,” a flagship initiative aiming to double the United States’ scientific output by directing artificial intelligence at the nation’s hardest scientific problems. He contends that AI will be the greatest amplifier of discovery in history, accelerating research in materials science, chemistry, mathematics, and physics. This initiative now involves participation from across the federal government.
Kratsios advocates a “New Golden Age” model: along with focused, mission-driven government investment in large-scale national targets, the U.S. should maintain robust funding for basic discovery science. This balanced portfolio ensures readiness for both near-term societal challenges and foundational advances that power future industries.
A key tenet of Kratsios’ vision is that funding decisions should prioritize the merit of individual scientists and ideas, rather than the institutional prestige of their host organizations. He stresses the importance of supporting scientific excellence regardless of a researcher’s background, home ins ...
Rethinking Science Funding and Diversifying Funding Flows
Download the Shortform Chrome extension for your browser
