TBPN Podcast – Episode Summary

Date: November 25, 2025
Hosts: John Coogan & Jordi Hays
Episode Focus:

• Polygenic embryo selection drama: Nucleus Genomics vs. competitors
• Michael Kratsios on the “Genesis Mission” for AI and science
• Nvidia’s public defense against TPUs and accounting skepticism
• Klarna and the future of stablecoins
• Updates from innovators in delivery drones, security, and tech
• Insightful discussions on AI’s present and future (OpenAI, Anthropic, Google Gemini 3)

Main Theme & Purpose

This episode delivers an intensive look into cutting-edge biotech controversy—especially around polygenic embryo selection services—AI infrastructural policy (via Project Genesis), and the ongoing chess game between Nvidia and Google in the race for AI hardware dominance. Interviews with industry leaders (including Nucleus CEO Keon, critic Cremieux, ex-White House CTO Michael Kratsios, Klarna CEO Sebastian) provide firsthand perspectives on both headline-grabbing disputes and large-scale technological visions.

Major Discussion Segments & Insights

1. Anthropic, OpenAI, SSI, and the AI Scaling Debate

00:00–13:04

• Open with reflections on the viral drop of the Ilya Sutskever–Dwarkesh Patel podcast, providing rare unscripted “hot-mic” insight into a leading AI researcher’s worldview.
• Quote:

  “His hot mic moment is like, ‘wow, it's exactly like science fiction!’ It's all real.” – John (04:59)

• Discussion of the “ages of AI”:
  • 2012-2020 as research-driven
  • 2020-2025 as scaling-driven
  • “Return to research” paradigm now necessary (cite Karpathy, Sholto, etc.)
  • AI progress will demand new breakthroughs, not just more GPUs
• Key insight: Consensus in the AI world that we’re hitting diminishing returns on blind scaling; fresh research paradigms are needed.

2. Nvidia Responds to TPU & Accounting FUD

13:04–21:59

• Nvidia issues public statements defending their lead against Google’s TPUs and refuting Michael Burry’s Enron-esque accusations.
• Quote:

  “Having the largest company in the world sending tweets to defend their main product is not very reassuring.” – Jordy (13:52)

• Deconstruction of media narratives and industry comms missteps.
• Highlight: Nvidia emphasizes system-level advantages, ecosystem lock-in, and market demand outpacing all competition.

3. Policy Moves: Genesis Mission – AI for Science

22:07–44:00, 119:44–145:44

• Readout of Trump's executive order launching the Genesis Mission to push scientific AI in U.S. national labs.
• Interview with Michael Kratsios (White House):
  • Genesis = “single largest marshaling of federal resources for scientific discovery since Apollo.”
  • Plan: National lab data + private AI/compute = new AI-driven scientific breakthroughs across bio, energy, defense, semiconductors, more
  • Strong emphasis on public-private partnership, U.S. export of AI stack, and keeping up with global competition

4. Trait-Based Embryo Selection: Nucleus Genomics Debate

44:00–87:39

a. Background & Context

• Review of Scott Alexander’s Astral Codex Ten deep-dive on trait-based embryo selection
• Distinction between early “designer baby” hype vs. current reality of trait/risk scoring
• Nucleus’s recent move to offer trait prediction services (IQ, height, eye color, etc.) sets off industry-wide storm; now embroiled in IP and scientific validity controversies

b. Cremieux Interview (Critic, 45:23–64:48):

• Outlines scientific and ethical issues leveled at Nucleus:
  • Allegations of misleading marketing about prediction accuracy
  • Use of small numbers of SNPs (e.g., 12 instead of >1M) for trait prediction: “physically impossible” to achieve claimed accuracy
  • Plagiarism in white papers, copying of competitor methods
  • Potential noncompliance with CLIA (regulator) for updated results without disclosure
  • Questionable customer testimonials, possible use of stock/AI images
• Quote:

  “They are effectively making up the results. They have actually... provided customers with score reports that have to be fake.” (49:11)

c. Key Issues Raised:

• Trust and transparency essential in life-altering tech
• Potential for Nucleus to sour the water for the emerging industry as a whole
• Redemption path would require radical transparency, accountability, and regulatory openness

d. Keon Interview (CEO Nucleus, 65:30–87:39):

• Denies all scientific wrongdoing; claims “models are public," invites independent validation
• Defends use of anonymized reviews and AI images due to HIPAA
• Says product predates public launch; reviews could reflect beta users; committed to clarifying marketing “sloppiness”
• Asserts industry competitors are attacking because “Nucleus has excellent science and excellent marketing”
• Acknowledges the need for better communication and transparency, especially on updated model results
• Quote:

  “Our science is public... go test it. In fact, our science is public, the competitors is not.” (67:16)

• Quote:

  “It's a very high stakes industry... It is appropriate that you will just kind of need to deal with this for a while.” – John (77:44)

e. Aftermath:

• Ongoing dispute about public availability of Nucleus models
• Critics point out Typeform signup for model downloads was down during taping

5. Panel: Klarna’s Stablecoin (Sebastian, 147:29–157:55)

• Klarna CEO Sebastian talks fintech’s next wave: Klarna launching a USD-pegged stablecoin (USDK) for peer-to-peer payments, internal treasury use, and cross-border settlement
• Blockchain rails via Bridge/Tempo/Stripe partnership
• Quote:

  “I think crypto plays a role... the value... is all these excess profits that we see in banking, they're going to come back to consumers.” (155:33)

6. Nvidia & Google, Debt-Driven AI Economy, Odd Lots with Joe Wiesenthal

87:44–112:18

• Reflection on Nvidia's unusual PR offensives, fragility in a world of industry/press narratives rapidly spiraling on social media
• AI's share of GDP growth, risk of overhyped tech stocks
• Transparency about venture capital vs. new debt-financed phase of large-scale AI buildout
• Importance of system-level view over focusing on benchmark/out-of-box specs

7. Delivery Drones, Zooming Past Sci-Fi: Interview with Zipline (Keller, 169:19–187:53)

• Zipline signs $150M+ contract to expand drone delivery across Africa
• In the US, explosive week-on-week growth
• “We are now planning for next year... to actually double our growth rate” (185:58)
• Unique insight: Infrastructure investments are now about creating real jobs/manufacturing in the US, "trade not aid"
• Drones transforming food delivery (“first customers burned their mouths…”)
• Hyperlocal, high-frequency delivery changing restaurant economics and consumer habits

8. Tech Security, AI Agent Credentialing (1Password, David, 158:22–168:37)

• 1Password's expansion into developer tools, headless browser integration, Atlas/Perplexity partnerships
• Increasing importance of credential hygiene as AI assistants become more integrated
• Resiliency and security layers more critical as attacks become more sophisticated

9. Screen Time & Attention Defense (Clearspace, Royce, 188:27–199:17)

• Clearspace aims to fight addictive attention loops and gives parents tools (e.g., “pushup to scroll” family challenges)
• Apple/Android “screen time” tools marked as insufficient—need for intentional, agentic control

10. Meta & Social Media Regulation, Parental Controls

(199:18–201:36)

• Debate on tech companies’ responsibility to protect users (esp. teens), and whether maturity ratings (like MPAA for movies) make sense for apps

Notable Quotes

• On the future of AI research:

  “We need another breakthrough and then simultaneously consensus be like, but like, we're definitely going to get that breakthrough in like the next decade. It's... hard to predict.” —John (11:13)

• On biotech trust:

  “This is as high stakes you get: People are selecting your future child.” —Jordy (204:20)

• On marketing in biotech:

  “If you're buying a phone case and the company uses an AI-generated hand model, that's one thing. But if the same tactic is used marketing the health of your future baby, that's a disaster.” —John (203:37)

• On AI infrastructure:

  “We want to create something that is so compelling and so exciting... that they [other countries] will want to use our [US] technology.” —Michael Kratsios (142:37)

• On Zipline adoption:

  “In a few weeks, Zipline will literally have the majority of homes in our service areas ordering from the service. That is... not a joke.” —Keller (183:39)

• On fintech transformation:

  “Crypto brings switching costs to zero... competition will return value back to consumers.” —Sebastian (154:15)

Critical Segment Timestamps

• Anthropic/AI scaling: 00:00–13:04
• Nvidia v. Google TPUs: 13:04–21:59
• Genesis Mission (White House): 22:07–44:00, 119:44–145:44
• Trait-based embryo selection, deep-dive + interviews: 44:00–87:39
• Klarna stablecoin: 147:29–157:55
• Nvidia debt, market shakeups (Joe Wiesenthal): 87:44–112:18
• 1Password, AI agent security: 158:22–168:37
• Zipline, US/Africa drone expansion: 169:19–187:53
• Clearspace, screen time discipline: 188:27–199:17

Takeaways for Non-Listeners

• Embryo selection technology is moving faster than the regulatory/societal consensus, creating intense competition—and controversy—over scientific rigor, transparency, and ethics.
  • Nucleus, the startup in the spotlight, faces criticism for potentially overstated science, hasty marketing, and industry-standard shortcuts—while competitors urge for more transparency and cross-validation.
• The US government is reasserting a role in enabling scientific and AI breakthroughs beyond consumer applications—creating public-private partnerships aimed at “moonshot” progress.
• Nvidia, long the unchallenged king of AI hardware, is now occasionally on defense amid rising questions about its accounting, its dominance, and challenger platforms like Google’s TPUs.
• Fintechs like Klarna are quietly starting to experiment with stablecoins/crypto as core infrastructure for both consumer and internal operations, indicating mainstream finance is finally integrating crypto rails.
• Drones (like Zipline) are shifting from sci-fi novelty to core infrastructure in health and food delivery, with real implications for both US jobs and African healthcare.
• Trust, transparency, and being able to critically evaluate claims remain central themes across all aspects of emerging tech discussed—from biotech to AI to financial infrastructure.

Tone: Candid, lively, richly detailed, occasionally irreverent, but always focused on substance over hype.
Recommended for: Listeners who value first-principles analysis, spirited debate with strong technical context, and want to keep pace with the most current threads in technology, policy, and business.