Invest Like the Best with Patrick O'Shaughnessy

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

1h 3min

Cliff Sosin challenges the prevailing narrative that intelligence is an unbounded scalar that will unlock infinite value. He argues that intelligence is primarily useful for interpolation and prediction, but many real-world systems are complex dynamic systems where prediction is fundamentally limited. Consequently, as AI becomes cheap and ubiquitous, the most valuable assets will be the complements to intelligence, such as data, attention, physical scale, and regulatory capture.

Summarized by Podsumo

🎧 Listen 🎙️ Ask about this episode

✨ Key Takeaways

1

Intelligence is like the roundness of a ball, not the height of a tree; it has fundamental limits in complex dynamic systems like weather or biology.
2

The key investment thesis is to focus on the complements to intelligence—such as data, attention, physical proximity, and regulatory capture—which become more valuable as intelligence becomes cheap.
3

Meta is cited as the archetype of a company with massive data and attention that will benefit from cheaper intelligence, while Grindr is given as an example of a company constrained by limited ability to deploy intelligence that could win big.
4

The investor’s job may bifurcate into picking assets (human) and portfolio construction (AI), with the latter already outperforming human judgment, especially at shorter timescales.
5

Education should be bimodal: teaching foundational skills (without AI) so kids can direct models effectively, and giving them ambitious, AI-assisted projects to own and justify.

💬 Notable Quotes

"What you should really think about intelligence as is more like a ratio, like the roundness of a ball. Intelligence is a gadget and we use it to make predictions about the universe."
"If intelligence becomes cheap and ubiquitous, what you want to own are the compliments to intelligence. The complements to intelligence are data, attention, love, affection, physical proximity, and scale."
"The models will be able to judge management. They'll be able to interview them, they'll be able to figure it out. And then it gets really interesting."
"I started out by saying you need to touch reality to answer questions. And so if you're gonna build an automatic biology testing laboratory and connect it to a computer, that's awesome. But if that experiment requires breeding mice and mice need a year to grow up and die, that experiment is gonna take time."