Uncapped with Jack Altman

Anduril & Founders Fund’s Trae Stephens on Choosing Good Quests in the Age of AI | Ep. 35

Trae Stephens is a Partner at Founders Fund. He is also Co-founder and Executive Chairman of Anduril, a defense tech company focused on autonomous systems, and Co-founder of Sol, a next generation wearable e-reader. Previously, Trae was an early employee at Palantir Technologies, where he led teams focused on growth in the intelligence/defense space as well as international expansion. He was also an integral part of the product team, leading the design and strategy for new product offerings. Prior to Palantir, Trae worked as a computational linguist building enterprise solutions to Arabic/Persian name matching and data enrichment within the United States Intelligence community. He began his career working in the office of then Congressman Rob Portman and in the Political Affairs Office at the Embassy of Afghanistan in Washington, D.C. immediately following the installation of Hamid Karzai’s transitional government. We covered: - Hard tech and the future of warfare - AI morality and “good quests” - How Anduril scales manufacturing - Founders Fund’s investing philosophy - Contrarianism and concentration - The ethics of autonomy and national defense Timestamps: (0:00) Intro (0:35) Choosing good quests in the AI era (7:35) Ethics behind solving certain problems (11:29) Working with regulators (16:54) What’s left to prove at Anduril (18:32) Anduril, SpaceX, and Tesla at scale (22:30) The future of warfare (24:48) Juggling Anduril and Founders Fund (29:07) A system that rewards going deep (31:21) What’s made Founders Fund great (37:30) The king-making strategy in VC (40:26) Concentrating in the winners (43:55) Where there’s alpha in the market (47:22) Theological revival in traditional faith More on Trae: https://foundersfund.com/ https://x.com/traestephens More on Jack: https://www.altcap.com/ https://x.com/jaltma https://linktr.ee/uncappedpod Email: friends@uncappedpod.com

TSTrae StephensguestJAJack Altmanhost

Dec 3, 2025·52m·Watch on YouTube ↗

📖 CHAPTERS

1. 1
   0:00 – 0:35

   Why AI makes it easy to build “uninteresting” companies—and why that’s a problem

   Jack opens by referencing Trae’s “Choose Good Quests” essay and asks how it applies in the AI boom. Trae argues AI’s distortion isn’t that it enables hard things, but that it makes low-effort, consensus businesses too easy—creating a talent misallocation problem.

2. 2
   0:35 – 7:35

   AI at the edge of human relationships: grief tech, companions, and social consequences

   Jack raises morally ambiguous AI products like recreations of deceased loved ones. Trae empathizes personally but warns that the deeper issue is how technology reshapes interpersonal relationships and social stability, drawing analogies to online dating.

3. 3
   7:35 – 11:29

   A practical ethics framework: the “feels good/bad” vs “is good/bad” matrix

   Trae lays out a two-by-two framework for evaluating controversial technologies and businesses. He places defense and civic “duty” in the “feels bad but is good” quadrant, while highlighting how addictive vices can sit in “feels good but is bad.”

4. 4
   11:29 – 16:54

   Policy and regulation: why guardrails come after reality, not before

   They discuss whether policy is the only mechanism to constrain “feels good but bad” markets and how regulation actually gets made. Trae argues legislators usually react to real-world boundary-crossings, so companies must engage constructively rather than expect Congress to pre-design rules.

5. 5
   16:54 – 18:32

   Working with Washington: defense as a rare bipartisan domain—and autonomy’s long history

   Jack asks how Anduril navigates regulation and procurement. Trae explains defense is unusually bipartisan and that autonomous systems have existed for decades, so legal and ethical frameworks around autonomy started earlier in defense than in consumer tech.

6. 6
   18:32 – 22:30

   The Silicon Valley cultural backlash: Project Maven, recruiting, and “just war” framing

   Trae describes the mid-2010s moment when tech workers resisted defense work, citing Google’s Project Maven controversy. He says Anduril’s main headwinds were recruiting and moral framing, not Washington acceptance, and emphasizes the value of public debate about military ethics.

7. 7
   22:30 – 24:48

   What’s left to prove at Anduril: scaling from thousands to tens of thousands

   Jack asks what remains for Anduril to reach its ambition. Trae says the next frontier is manufacturing scale—building the muscle to produce tens of thousands of units, which requires design-for-manufacturing and an industrial operating model rare in venture-backed startups.

8. 8
   24:48 – 29:07

   Lessons from Tesla and SpaceX: why production and verticalization are brutally hard

   They explore why moving from a prototype to mass production is so difficult. Trae contrasts exquisite early builds with the operational complexity of million-unit production, and notes SpaceX’s need to verticalize due to catastrophic failure risks in aerospace supply chains.

9. 9
   29:07 – 31:21

   Defense-tech reality check: procurement isn’t “Field of Dreams”

   Trae gives a blunt warning to defense startups: building a great product isn’t enough. Government buying is driven by procurement mechanics and primes, so success requires a strategy to get adopted as a non-traditional vendor, making “business of government” know-how a major part of the battle.

10. 10
    31:21 – 37:30

    The future of warfare: low-cost autonomy across domains and the declining carrier era

    Jack asks about space warfare and the broader trajectory of conflict. Trae predicts autonomy will reshape every domain and argues the cost calculus is making traditional force-projection assets like aircraft carriers less viable in great-power conflict, while human combat continues trending downward overall.

11. 11
    37:30 – 40:26

    Operator-investor life: why Trae juggles Anduril and Founders Fund

    Jack pivots to Trae’s dual role as Anduril cofounder and Founders Fund partner. Trae says it wasn’t a planned “slashie” identity—Anduril emerged from a gap he saw in defense primes—and day-to-day execution is a priority-driven context switch supported by a strong chief of staff.

12. 12
    40:26 – 43:55

    Founders Fund’s culture: a system that rewards going deep (without incubating on purpose)

    They discuss how Founders Fund differs from more board-heavy, operational VC models. Trae explains the fund is “for founders,” avoids trying to run companies from the board, and this creates time/permission for partners to go deep—sometimes resulting in partners starting companies as a byproduct.

13. 13
    43:55 – 47:22

    What made Founders Fund great: access, debate-driven evaluation, and concentrated follow-on

    Trae breaks performance into three buckets: access, evaluation, and follow-on investing. He credits early access advantages to the firm’s origin/brand, emphasizes an internal culture of open debate with no sacred cows, and describes a strategy of heavy concentration into top winners rather than many small bets.

14. 14
    47:22 – 52:24

    King-making, kamikaze rounds, and where alpha exists now—plus a revival of traditional faith

    Jack asks about brand-driven “king-making” and today’s venture alpha. Trae critiques oversized, overpriced ‘kamikaze rounds’ that can cripple companies, reiterates Founders Fund’s anti-FOMO stance, and says alpha comes from avoiding hype and backing founders with real origin stories. The conversation closes on a post-COVID revival in interest in traditional faith and how Christian just-war tradition informs moral reasoning about defense.