How the White House Plans to 10x Scientific Productivity | Michael Kratsios | EP #276
Peter Diamandis interviews Michael Kratsios, Director of the White House Office of Science and Technology Policy, at the White House about America's AI Action Plan, the Genesis Mission (AI for scientific discovery), and the 'Science in a New Golden Age' report. They cover why scientific productivity has declined despite bigger budgets, reform mechanisms like long-duration grants, fast-track reviews, golden-ticket funding, and an AI-native science marketplace using prediction markets and DAOs, plus six national technology missions spanning AI, quantum computing, fusion, lunar exploration, robotics, and semiconductors. The conversation also touches on AI's public-perception problem, job-loss fears, open vs. closed source AI strategy, chip export controls on China, humanoid robotics policy, regulatory 'innovation zones,' and education reform. Kratsios reveals he now believes the Genesis Mission's productivity target should be raised from 2x to 10x given recent AI progress. The episode closes with a short Fountain Life segment with Dr. Don Saladino on early cancer detection via full-body MRI screening.
Regulatory philosophy: 'born free' vs 'born in captivity' technologiesOther▶ 0:03:27
Kratsios frames tech regulation in two buckets: 'born free' technologies (like the 1990s internet) where government should step back, and 'born in captivity' technologies (commercial drones, AI-powered medical diagnostics) that need affirmative government action before their benefits can reach Americans. He cites the EU AI Act as an example of premature regulation, finalized before ChatGPT existed.
AI's public-perception problem and fear reductionAI▶ 0:08:59
Diamandis cites polling that 3/4 of Americans fear AI and 71% oppose data centers near their homes (more than oppose nuclear plants). Kratsios traces the fear-first narrative to the 2023 Bletchley Park AI Safety Summit and argues AI has 'a massive PR problem'; both discuss using healthcare stories, manufacturing build-outs, and Diamandis's new Future of Vision X Prize film competition to counter dystopian sci-fi narratives.
AI and jobs: labor market data and reskillingEconomy▶ 0:12:53
Kratsios says he's long-term optimistic about AI and employment, noting automation fears from the first Trump administration never materialized. He points out companies often blame ordinary layoffs on AI because it plays better in the press, and describes a new Department of Labor initiative (from the AI Action Plan) to collect better data on AI's actual labor-market impact to guide reskilling programs.
Open source vs. closed source AI and the American AI Exports ProgramAI▶ 0:19:59
Kratsios says the US must lead in both open and closed source AI; the domestic open-source ecosystem 'could be doing better' versus well-funded Chinese open models. The American AI Exports Program (via Commerce, now whole-of-government) bundles US chips, models, and applications into turnkey national AI stacks backed by Export-Import Bank and DFC financing, explicitly designed to prevent a repeat of Huawei's subsidized global telecom expansion.
Diamandis asks whether restricting Nvidia chip exports to China backfired by spurring Chinese domestic chip development, drawing a parallel to launch-vehicle and satellite-defense export restrictions. Kratsios disagrees, calling the controls 'one of the savviest decisions' of the administration and crediting 2019-era EUV lithography export controls with widening the US lead over China's best chips.
Humanoid robotics policy and supply chainRobotics▶ 0:27:16
Diamandis notes the US has a handful of humanoid robot companies versus China's 150+, aggressively state-supported. Kratsios says the administration just restricted importation of non-US humanoid robots (grandfathering already-shipped units) to force domestic supply-chain buildout, drawing a parallel to a similar December drone/UAS restriction that reportedly boosted US drone supply-chain investment.
National technology missions: Moon, lunar base, space nuclear reactor, quantum, fusionSpace▶ 0:17:07
Kratsios lists six national technology missions: putting a man back on the Moon by 2028, building the first lunar base elements by 2030, launching a nuclear reactor into space by 2028 (for Mars-capable propulsion), the Genesis Mission (AI for science), an executive order to build a scientifically relevant quantum computer by the end of Trump's term (Kratsios is most excited about pharmaceutical applications), and next-generation semiconductors. Fusion energy is also discussed as newly attracting unprecedented private investment (37 venture-backed fusion companies).
Declining scientific productivity and Golden Age reform mechanismsOther▶ 0:31:38
Kratsios explains 'Eroom's Law' (Moore's Law backwards) — discovery per dollar has fallen despite bigger budgets and better tools — because science hasn't changed how it's conducted, citing NIH's ~$45B budget alongside rising drug costs. The Golden Age report proposes long-duration (5-year, funded-on-day-one) grants, fast-track reviews (a few pages, reviewed in under a month, inspired by Tyler Cowen's COVID-era rapid grants), 'golden ticket' funding where individual review-panel members can unilaterally fund risky proposals, and new meta-science units at NIH and NSF to test which funding mechanisms actually work.
AI-native scientific marketplace: agents, DAOs, and prediction marketsAI▶ 0:42:36
Diamandis reads a Golden Age report passage describing a future scientific marketplace where AI agents identify promising research leads, hire autonomous labs, verify cryptographically signed results, exchange data and funding via microtransactions and smart contracts, and where prediction/bounty markets and reputation systems replace slow institutional grant coordination. Kratsios calls this 'possible' and links it to record philanthropic scientific capital (e.g., OpenAI Foundation near a quarter-trillion-dollar valuation) and the historical flip from 70% federal to ~70% private R&D funding since 1950's Endless Frontier era.
Longevity and health span missing from the Golden Age reportLongevity▶ 0:48:33
Diamandis points out longevity is absent from the report despite average US life expectancy (~78-79) far outpacing health expectancy (63), leaving 16 years in poor health. He references the $101M XPRIZE Healthspan (830 competing teams) aiming to reverse functional age by 20 years, and Kratsios concedes it should probably have been included, promising to explore NIH programs on it.
Diamandis describes a closed-loop science vision — AI proposes a hypothesis, robots run the experiment, the model reads results and designs the next one, 24/7 with no humans in the loop — citing Lyla Sciences' million-square-foot buildout. Kratsios says the administration wants to retrofit federal labs with this capability and expects universities to be early adopters, ultimately democratizing lab access for high schoolers and college students without needing institutional connections.
Education reform: AI-first schools and four-year PhDsOther▶ 0:57:05
Diamandis argues the US education system is still built for the Industrial Revolution and should be 'AI first.' Kratsios says the administration prioritizes parental choice (from AI-heavy models like Alpha School to no-tech classrooms) but acknowledges most Americans get a 'broken middle' serving neither well, and highlights a same-day announcement of four-year PhD programs pairing students with private-sector companies to prepare them for industry, not just academia.
Regulatory innovation zones and data center policyEconomy▶ 0:52:51
Diamandis raises the idea of state/regional 'singularity zones' with relaxed regulation to attract tech entrepreneurs, referencing Kratsios's 2017 UAS drone Innovation Pilot Program and the newer EIPP eVTOL testing program as precedents, plus interstate 'regulatory arbitrage' (self-driving companies leaving California for Arizona). Kratsios connects this to the president's ratepayer protection pledge requiring data center operators to build/bring/buy their own power rather than raising local electricity costs.
“The president um directed through an executive order is to create a scientifically relevant quantum computer by the end of his term.”
Your call:
openMichael Kratsios: The Genesis Mission will double the productivity of the entire US scientific enterprise, though Kratsios says the administration should now aim for a 10x target instead.
EP #? · · due: within the next decade (target under review, being raised to 10x) · ▶ watch
“We essentially want to double the productivity of the entire scientific enterprise in the United States over the next decade... to me I think we should probably aiming for 10.”
Your call:
openElon Musk (relayed by Peter Diamandis): Double-digit US GDP growth is coming, followed by triple-digit growth, en route to a 'post-capitalist' economy where AI and robotics deliver anything anyone could want.
EP #? · · due: double-digit growth in 18-24 months, triple-digit growth in 5 years, post-capitalist economy by 2036 · ▶ watch
“we're going to see double digit growth in the GDP in the next 18-24 months triple digit growth in 5 years and then... by 2036 basically post capitalist... anything you could possibly want will be delivered by AI and robotics.”
“Demis talking about curing all diseases within the next decade... He's dead serious.”
Your call:
Numbers that matter
3/4 of Americans fear AIDiamandis citing current polling on AI sentiment
71% of Americans oppose data centers near their homesHigher than opposition to nuclear power plants nearby
80% of people in China are pro-AIContrast with US sentiment, per Diamandis
150-plus humanoid robot companies in China vs. a handful in the USDiamandis comparing US and Chinese humanoid robotics industries
NIH budget has ballooned to almost $45 billionKratsios noting rising biomedical research spending alongside rising drug costs
45% of a researcher's time is spent on grant administrative workNational Academies study cited by Kratsios
Median age of an intramural NIH scientist is 71Compared with Nobel laureates' prize-winning work typically done in their mid-40s
37 venture-backed fusion companies (Diamandis's last count)Illustrating unprecedented private investment in fusion energy
Nvidia's order to Corning is the largest in Corning's historyJensen Huang relayed anecdote about AI-driven supply chain demand
OpenAI Foundation valuation is almost a quarter-trillion dollarsCited as evidence of unprecedented philanthropic/scientific capital available today
In 1950, ~70% of US R&D was federally funded and 30% private; today that ratio has flipped (majority private, ~30% federal)Kratsios on the changed science-funding ecosystem since Vannevar Bush's Endless Frontier
XPRIZE has launched $600 million of prizes, driving about $30 billion of R&DDiamandis on incentive-prize leverage
The Ansari X Prize was $10 millionCited example in the Golden Age report of incentive prizes
US government spends about $200 billion a year on R&D, ~$45 billion of it at NIHKratsios sizing government R&D budget relative to potential prize sizes
A single government incentive prize in the range of $100 million is 'certainly doable'; $1 billion 'may be a lot' for one prizeKratsios responding to Diamandis's proposal of $1 billion prizes
Worth digging into
🕳️ The AI-native scientific marketplace (agents, DAOs, prediction/bounty markets funding research directly)
Described by Kratsios's own report as a 'complete fundamental reimagining of the entire scientific process' but discussed only conceptually; no concrete pilot or timeline exists yet.
🕳️ Autonomous self-driving labs retrofitted into federal and university labs
Kratsios frames this as democratizing lab access down to high schoolers, but acknowledges the hardware to run a wide range of experiments doesn't yet exist at scale; Lyla Sciences' million-square-foot buildout is the only cited concrete example.
🕳️ Genesis Mission productivity target revision from 2x to 10x
Kratsios states live in the interview that he now believes the official 2x target is outdated and should be 10x, but this hasn't been formally revised in the published report.
🕳️ Humanoid robot and drone import restrictions as an industrial-policy tool
The administration just imposed import limits on non-US humanoid robots (following a similar December drone restriction), betting this will replicate the drone supply-chain investment surge; the actual investment data for robotics hasn't yet materialized.
Diamandis references a new state-by-state pro-tech ranking (Texas #1) and Kratsios draws on 2017 UAS pilot program precedent, but no new federal 'innovation zone' program was announced in this interview beyond existing drone/eVTOL pilots.
🕳️ Longevity/healthspan being absent from the six national technology missions
Kratsios admits on-air this was an oversight ('now that you mentioned it, we probably should have included it') despite XPRIZE Healthspan and prominent longevity scientists pushing the case that aging is the highest-leverage economic problem to solve.