# An AI Diffusion Framework and Elizabeth Stone’s Throwback Reads

*By Recommended Reading from Tech Founders • July 20, 2026*

Aaron Levie highlights Clifford Sosin’s framework for understanding why AI adoption is faster in software than in industries governed by real-world feedback loops. Elizabeth Stone adds two personal, Wall Street-connected book recommendations shared during a podcast lightning round.

## Most compelling: AI diffusion is constrained by reality feedback

### [Clifford Sosin’s post on AI diffusion](https://x.com/cliffordsosin/status/2078594661359194500)

- **Content type:** X post / article
- **Author:** Clifford Sosin
- **Recommended by:** Aaron Levie
- **Key takeaway:** Levie recommends the post for understanding AI diffusion, arguing that AI-driven progress is ultimately limited by interaction with the real world. [^1] Coding can be adopted quickly because one person can write, test, and deploy it without requiring external parties to change behavior; life sciences, sales, and contracts instead depend on testing, negotiation, or other real-world interactions. [^1]
- **Why it matters:** The recommendation draws a practical distinction between model capability and applied AI: useful systems must reshape industry workflows and handle their real-world feedback loops, rather than simply produce model outputs. [^1]

> “Coming up with ideas was never the hard part. The hard part is how fast reality answers them.” [^1]

This is the day’s strongest recommendation because it offers a clear lens for assessing where AI adoption can move quickly—and where the limiting factor is experimentation and coordination outside the model.

## Two personal “throwback” book picks

Elizabeth Stone shared both books during a podcast lightning round, connecting them to her Wall Street background. [^2]

### *Into Thin Air*

- **Content type:** Book
- **Author:** Jon Krakauer
- **Link:** No direct book URL was supplied; [watch Stone’s recommendation](https://www.youtube.com/watch?v=t0GiTyz4syY)
- **Recommended by:** Elizabeth Stone
- **Key takeaway:** Stone named it among the books she still recommends, describing her choices as “throwback” reads. [^2]
- **Why it matters:** It is a direct personal recommendation from a technology leader, offered in a non-promotional interview context. [^2]

### *Liar’s Poker*

- **Content type:** Book
- **Author:** Michael Lewis
- **Link:** No direct book URL was supplied; [watch Stone’s recommendation](https://www.youtube.com/watch?v=t0GiTyz4syY)
- **Recommended by:** Elizabeth Stone
- **Key takeaway:** Stone recommended it alongside *Into Thin Air*, saying she likes revisiting books that remind people what Wall Street was like “in the way back time.” [^2]
- **Why it matters:** The recommendation is rooted in Stone’s own Wall Street experience, giving the selection a clear personal context rather than a generic reading-list endorsement. [^2]

---

### Sources

[^1]: [𝕏 post by @levie](https://x.com/levie/status/2078864191683969212)
[^2]: [Why Netflix is betting on systems thinkers—not specialists—in the AI era | Elizabeth Stone \(CPTO\)](https://www.youtube.com/watch?v=t0GiTyz4syY)