Gutenberg Lab — Morning Paper

Trump Administration Freezes $3B in Research Grants, Reallocates from DEI and Climate to Basic Science, AI, and Fusion

Featuring

Jason Calacanis (host, tech investor) × Michael Kratsios (Director, White House Office of Science and Technology Policy; former U.S. CTO)

Bottom line

American science has become politicized and stagnant despite tripling NIH budgets since 1998, with research efficiency falling 80-fold since 1950 (halving every 9 years under Eroom's Law).

3-Line Summary

The Trump administration froze or terminated $3 billion in unspent research grants in 2025, reallocating funds from DEI-related research (25% of NSF grants, $8B over 4 years) and climate science toward basic research, AI-for-science, fusion (2035 target), and quantum computing (2028). While a Nature poll showed 86% of scientists supported Harris vs. 6% for Trump, Kratsios argues "the scientific community has lost its way" through politicization, citing Fauci's COVID policies (masks in schools, protests allowed but not visiting dying grandparents) as "more disservice to science than anyone in modern history." China's science spending grew 19x from 2000 ($33B to $670B) vs. U.S. 3x increase; a decade ago the U.S. published twice as many papers as China, but today China publishes 50% more in nearly every domain.

3 Key Points

1. Eroom's Law and Research Collapse: NIH funding tripled from $14B (1998) to $47B (2024), yet research efficiency fell 80-fold since 1950, halving every 9 years (Eroom's Law). Kratsios: "We have not been innovative in any way on the way that we actually conduct the science" — same methods with more money, no incentive for Congress (district funding) or agencies (admitting failure) to change.

2. $8B to DEI-Related Research: Senator Ted Cruz's analysis found 25% of NSF grants under Biden went to "DEI-related science" = ~$2B/year × 4 years = $8B. Kratsios: "That isn't science" and represents politicization where applicants had to include DEI factors to win funding. This justifies the $3B freeze/termination.

3. RCP 8.5 Climate Scenario Retraction: The most extreme climate projection (RCP 8.5) used in "every media report" and national assessments for 20 years was pulled by IPCC months ago as scientifically unsubstantiated. Kratsios: "If I had questioned it publicly 3-4 years ago, I would've been criticized dramatically" as anti-science — now vindicated.

Editorial Perspective

This is not "anti-science" but a political struggle over the definition of science itself. Kratsios's position — that science should be inquiry, not obedience to authority — is classically defensible, but when implemented as climate budget cuts and DEI research exclusion, it raises the question of who holds the power to define "real science." The median age of 71 for NIH intramural researchers signals this is a battle over both generational succession and resource allocation.

Source: All-In, "Michael Kratsios: Trump's Science Agenda, Anti-Science Claims, Fauci's Damage, DEI & China" (August 24, 2026)

https://allinchamathjason.libsyn.com/michael-kratsios-trumps-science-agenda-anti-science-claims-faucis-damage-dei-china

AI Disclosure: Produced with AI assistance; facts and analysis reviewed by our editorial team.

• --

China's Robot Olympics Is a National Strategy to Normalize Automation—the West Admits Defeat with UBI

Featuring

Jason Calacanis (founder & investor, This Week in Startups) × Lan (co-host, producer)

Bottom line

China is staging government-backed humanoid robot competitions to make automation entertaining and build public acceptance, while the West leads with job displacement messaging like Sam Altman's UBI experiments—a messaging failure that frames AI as a threat rather than empowerment.

3-Line Summary

• At the World Robot Conference in Beijing (Aug 19-23), humanoid robots set records of 9.39 seconds in the 100m sprint (beating Usain Bolt's 9.58) and 2.88 meters in standing high jump (surpassing the human record of 2.45m set in 1993)
• Jason predicts Tesla's Optimus will sell 1 billion units in 10 years to become "the greatest product ever made by humanity," claiming it is 1.5 hardware generations and 5 AI generations ahead of Chinese competitors based on a private demo he attended
• Hugging Face is reportedly in sale talks at over $13 billion (last raised at $4.5B in 2023); Jason argues frontier labs like OpenAI or Anthropic should acquire it to control the open-source threat vector

3 Key Points

1. China's robot PR as psyops: The World Robot Conference has been held annually since 2015 with government backing; the 2026 event featured 51 competitions and over 1,000 participants. Jason frames this as a deliberate strategy to "make people root for the robots before the robots take everyone's jobs," contrasting it with Western messaging that leads with UBI and job loss—effectively admitting humans are obsolete

2. Optimus market dominance prediction: Jason claims the Optimus demo he saw privately was "better than anything at the world games" and predicts 1 billion units by 2036, surpassing the iPhone as the best-selling product in human history. He says the hardware is 1.5 generations ahead of Chinese robots and the AI brain is 5 generations ahead

3. Minimum wage over UBI: As an alternative to universal basic income, Jason proposes raising the federal minimum wage from $7.50 to $20 per hour over 10 years (phased at $1/year). He estimates only 15 million Americans earn under $15/hour, and the consumer cost would be a 4% increase (a $4 latte becomes $4.16)

Editorial Perspective

China's framing of automation as national spectacle while the West leads with UBI represents a failure of agenda-setting in the classic media theory sense. If Altman had opened with "you'll learn any skill for free, for life" and "you'll have the same tutor as rich kids," AI would have been received as liberation rather than threat. Raising the minimum wage is a pragmatic path to lift the economic floor while preserving dignity—a conversation that should precede UBI, not follow it.

AI Disclosure: Produced with AI assistance; facts and analysis reviewed by our editorial team.

Source: This Week in Startups, "China wants you to cheer for the robots taking your job | E2329" (August 25, 2026)

https://4a885955-6823-4b22-a3e1-a526c25516a5.libsyn.com/china-wants-you-to-cheer-for-the-robots-taking-your-job-e2329

The Three Native Advantages of Borderless Founders — Why a16z Now Allocates 40% of Investments to International Entrepreneurs

Featuring

Elena Berger (Host) × Angela Strange (a16z Partner, leads global investing strategy) × Gabriel Vasquez (a16z Partner, co-leads global investing strategy)

Bottom line

In the AI era, having deep ties to both a home market and Silicon Valley is not a disadvantage but a structural advantage, and a16z now allocates 40% of its recent investments to founders with international backgrounds.

Three-Line Summary

• a16z partners Angela Strange and Gabriel Vasquez explain the structural advantages of "borderless founders" — entrepreneurs with international backgrounds who leverage both home market networks and Silicon Valley resources
• AI created a paradox: while the technology democratizes globally, the epicenter of fastest movement has concentrated in the Bay Area, accelerating international founder migration to Silicon Valley
• Three native advantages (talent access, brand, customer networks) create compounding flywheels, and a16z organizes country-specific diasporas to amplify these effects

Three Key Points

1. AI's concentration paradox: Strange notes that AI is "very democratic" and distributes globally, yet simultaneously "concentrates the epicenter of where things are moving the fastest to the Bay Area." This drives Brazilian and Colombian founders who previously wanted a16z to visit their countries to now come to the Bay Area and integrate faster.

2. Stack-building as forced differentiation: Unlike the US where "you can get everything as a service at every layer of the stack," other markets lack fraud-as-a-service or KYC-as-a-service, forcing founders to "build everything," Strange explains. Founders who survive this gauntlet develop deeper technical capabilities and market knowledge than US founders who can assemble solutions from existing services.

3. 40% of investments go to international founders: Strange reveals that 40% of a16z's recent investments were in international founders, with half based in the Bay Area and half elsewhere. Vasquez wants to shift venture returns from the current 90% Silicon Valley / 10% international split to 20-30% international over ten years. This is not a diversity initiative but a fundamental reallocation of capital based on the belief that the historical 90/10 split was inefficient, not optimal.

Editorial Perspective

The advantage of borderless founders is not about overcoming geographic disadvantage — it is about having additional weapons. Access to differentiated talent pools in home countries, trust from governments and large enterprises, and the ability to win customer acquisition races in home markets where "nobody wants to be first" in the US — these are options US founders do not have. Now that AI has democratized the technology itself, this asymmetry becomes even more pronounced. a16z allocating 40% of investments to international founders signals a judgment that historical capital allocation was not optimal but inefficient.

Source: a16z Podcast "How Global Networks Are Reshaping Startup Success" (August 20, 2026)

https://a16z.simplecast.com/episodes/how-global-networks-are-reshaping-startup-success-YCGysOkP

AI Disclosure: Produced with AI assistance; facts and analysis reviewed by our editorial team.

• --