Skip to content

Cliff Sosin - Owning the Complements to Intelligence - [Invest Like the Best, EP.497]

Invest Like the Best with Patrick O'Shaughnessy

1h 2m
Transcript

Original podcast

Where to listen

Summary

Written by a model from the transcript; not yet reviewed.

This episode features a discussion on the economic and practical limits of artificial intelligence, examining what happens when intelligence becomes cheap and ubiquitous. The guest argues that intelligence is not an unbounded scaler like a tree height, but rather a ratio or tool constrained by complex, dynamic systems that require touching reality. Consequently, as intelligence approaches zero cost, the true economic value shifts to owning its complements, such as data, physical proximity, and attention. The conversation explores how major companies like Meta and consumer services are positioned in this future, the capital cycles of infrastructure build-outs, and how concentrated portfolio strategies hold up against conventional diversification.

  • When intelligence becomes cheap and ubiquitous, investors should focus on owning its complements like data, attention, physical proximity, and scale. 8:20
  • Complex dynamic systems like weather or human interaction have inherent simulation limits, meaning raw intelligence cannot predict everything without touching reality. 2:44
  • Ubiquitous cheap intelligence drives down the cost of software and inputs, which can increase consumer surplus and lower prices across industries like garbage collection. 36:54
  • To effectively direct AI agents and leverage their capabilities, children and professionals still need foundational education and domain knowledge. 57:25

Chapters

  1. 0:00
    Intro and Ad Read

    Patrick introduces the episode and promotes Ramp, WorkOS, and other AI platforms.

  2. 1:54
    Defining Intelligence Limits

    Cliff explains the bounded nature of intelligence and its limits in complex systems.

  3. 4:31
    Complex Adaptive Systems Example

    Cliff uses weather modeling to illustrate how small errors amplify in dynamic systems.

  4. 8:20
    Intelligence as a Complement

    Cliff argues that as intelligence becomes cheap, owning complementary assets like data and attention becomes key.

  5. 9:38
    Meta and the Power of Data

    Cliff analyzes Meta’s business model, showing how data and attention drive value even with cheap AI.

  6. 23:51
    Supply vs. Demand in AI

    Discussion on the tension between building AI infrastructure and meeting consumer demand for AI services.

  7. 32:55
    Future of Agents and Investing

    Cliff speculates on how AI agents will change investing and the role of human judgment.

  8. 42:09
    Future of AI Investment and Consumption

    Discussion on how AI build-out requires future consumption and impacts rates and savings.

  9. 43:01
    Embedding AI in Corporate Culture

    Steps for CEOs to integrate AI through hackathons and read‑only system access.

  10. 44:53
    Timing and Competitive Edge of AI Startups

    Why now is the best time to launch AI‑driven companies and the challenge of sustaining advantage.

  11. 49:44
    AI’s Impact on Income Distribution

    Exploration of how ubiquitous AI could flatten or steepen income inequality.

  12. 52:43
    AI in Portfolio Construction

    Comparison of human versus AI portfolio construction and future investor roles.

Highlights

  • 1:54

    Intelligence as a Ratioinsight

    The guest explains why viewing intelligence as an unbounded scaler to godhood is an intellectual mistake, comparing it instead to a ratio or a roundness of a ball used to interpret complexity.

  • 4:31

    The Butterfly Effect Originstory

    The discussion covers the historical discovery of chaos theory through weather forecasting models, where a tiny rounded decimal in initial conditions created entirely different forecasts within days.

  • 9:32

    Meta as Personalized TVinsight

    Analyzing Meta from a high level, the guest describes its core business as personalized TV, spending its vast resources on intelligence while sitting on massive data and attention complements.

  • 18:00

    Demand Curves and Fiber Opticsdata

    Examining Jevon's paradox through the lens of communication costs, the guest notes that markets eventually shrink when further price decreases in information transfer stop driving new demand.

  • 27:50

    Society Saving for AIinsight

    The comparison is made that current American society resembles a household saving and feeling pinched because massive resources and labor are diverted into building AI infrastructure.

  • 38:05

    Capital Markets Deployment Speeddata

    The conversation highlights the remarkable speed at which American capital markets redeployed approximately three percent of GDP into speculative AI ventures on short notice.

  • 40:21

    The Risk-Taking Theoryinsight

    A pet theory is shared suggesting that American economic dominance stems from an interplay of genetics, cultural evolution, and institutional laws favoring risk-taking.

Quotes

  • “There's this idea that if intelligence starts as an amoeba, and it gets better and better, and then it's like a monkey, and then it's like a person, and then it's like Einstein, and then it goes zoom, and it becomes god.”
    Unnamed speaker · 1:54
  • “People would kill me if I didn't ask about your portfolio a little bit, which is so interesting to me. It's 80% Carvana, vacations, 5% Meta is basically your portfolio.”
    Unnamed speaker · 42:36
  • “I think everything we're talking about, argue down. There's many businesses where they make margin because-”
    Unnamed speaker · 47:59
  • “If you read history, of course, all the people who did really great tended to be very concentrated.”
    Unnamed speaker · 56:07
  • “The more intelligence you have and the more crystallized intelligence you have, which would be said of education, the better you are at communicating with a model and telling a model what you want.”
    Unnamed speaker · 57:29
  • “You don't have to work out the math, but you have to know what it is.”
    Unnamed speaker · 58:28
  • Ramp 0:00–0:27
  • WorkOS 0:28–1:07
  • Rogo 1:08–1:42
  • Vanta 20:31–21:18
  • RidgeLine 21:01–22:34
  • Ramp 46:16–46:33
  • Vanta 46:29–46:43
  • Bridgeline 46:39–46:54
  • Rogo 46:54–47:06
  • WorkOS 47:08–47:20
  • Colossus 59:59–1:00:41
  • Ramp 1:01:43–1:01:57
  • RidgeLine 1:02:02–1:02:16

About this episode

Cliff Sosin is the founder of CAS Investment Partners, which he started in 2012. Roughly 80% of his portfolio is in Carvana. We discuss why Cliff thinks there are hard limits to what even the smartest AI can predict, which businesses benefit as intelligence gets cheap, and what that means for Meta and the frontier labs. We also cover why he avoids investing in the AI build-out, how AI agents will push customers to the lowest-cost provider, and why he is comfortable with 80% of his portfolio in one stock. Please enjoy my conversation with Cliff Sosin. For the full show notes, transcript, and links to mentioned content, check out the episode page here. ----- Become a Colossus member to get our quarterly print magazine and private audio experience, including exclusive profiles and early access to select episodes. Subscribe at ⁠colossus.com/subscribe⁠. ----- ⁠Ramp’s⁠ mission is to help companies manage their spend in a way that reduces expenses and frees up time for teams to work on more valuable projects. Go to⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ ⁠ramp.com/invest⁠⁠ to sign up for free and get a $250 welcome bonus. ----- Trusted by thousands of businesses, ⁠Vanta⁠ continuously monitors your security posture and streamlines audits so you can win enterprise deals and build customer trust without the traditional overhead. Invest Like the Best listeners get a special offer of $1,000 off Vanta when you go to ⁠vanta.com/invest⁠. ----- WorkOS⁠ is the infrastructure B2B and AI-native companies use to sell to enterprise. It covers everything enterprise security requires: SSO, SCIM, RBAC, Audit Logs, AI governance, and more. Trusted by 2,000+ fast-growing companies, including OpenAI, Anthropic, Cursor, and Vercel. ----- Rogo is the AI platform for finance. They're building agents for Wall Street that are trained to understand how bankers and investors actually do work: from diligence and modeling, to turning analysis into deliverables. To learn more, visit rogo.ai/invest. ----- ⁠Ridgeline⁠ has built a complete, real-time, modern operating system for investment managers. It handles trading, portfolio management, compliance, customer reporting, and much more through an all-in-one real-time cloud platform. Visit⁠ ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ridgeline.ai⁠. ----- Editing and post-production work for this episode was provided by The Podcast Consultant. Timestamps: (00:00:00) Why Intelligence Has Limits (00:02:34) Chaos Theory and the Butterfly Effect (00:04:59) A 200 IQ Team in Every Pocket (00:07:33) Meta as Personalized TV (00:11:18) The Frontier Labs as Businesses (00:14:59) Jevons Paradox and the Limits of Demand (00:18:47) Automated Labs and the Speed Limits of Experiments (00:20:39) Why He Stays Away from the AI Build-Out (00:22:28) Why the Best Entrepreneurs Are Building Supply (00:24:22) Lessons from Past Capital Cycles (00:26:05) TSMC and Physical Moats (00:27:03) AI Agents and the Low-Cost Producer (00:29:44) Will AI Replace Investors? (00:32:01) Price as a Signal Wrapped in an Incentive (00:34:16) America’s Underrated Capital Markets (00:37:05) Why America Dominates Risk-Taking (00:38:36) What Higher Rates Are Telling Us (00:39:24) The AI Speech He Gives His Companies (00:41:45) The Best Time to Start a Company (00:43:03) Corporate Margins and Vertical Stacks (00:45:16) AI and the Income Distribution (00:48:13) Letting AI Size the Portfolio (00:50:56) Why He Holds So Much Carvana (00:52:22) Educating Kids in the Age of AI

Transcript

0:00
1:02:32

13,152 words · 3 speakers

Topics

  • Meta
  • Carvana
  • Positive Sum
  • Nvidia
  • Grokbot
  • Metamuse
  • Renaissance Technologies

Page updated · transcript and summary are machine-generated