# Zero to One, The Bitter Lesson, and an Austin Housing Case Study

*By Recommended Reading from Tech Founders • April 15, 2026*

Anj Midha credits Peter Thiel, Richard Feynman, and Rich Sutton with shaping how he thinks about competition, teaching, and compute-driven discovery. Bill Gurley adds a Pew article on Austin housing supply as an empirical read for affordability debates.

## High-signal recommendations

Today's authentic signal is concentrated: Anj Midha surfaced three durable resources on competition, teaching, and compute-driven discovery, while Bill Gurley highlighted one empirical article on housing affordability [^1][^2].

## Most compelling recommendation

### *Zero to One*
- **Content type:** Book [^1]
- **Author/creator:** Peter Thiel [^1]
- **Link/URL:** Not provided in source material
- **Who recommended it:** Anj Midha [^1]
- **Key takeaway:** Midha credits Thiel's Stanford class, later turned into *Zero to One*, with shaping his business thinking. He updates the familiar line "competition is for losers" into a frontier-AI view: neither commoditized overcompetition nor monopoly is healthy; the best structure is "optimal competition" with three or four top teams in each frontier [^1]
- **Why it matters:** This is the day's strongest pick because the recommendation comes with an applied framework readers can use to think about startup positioning and market structure right now [^1]

> "competition is for losers" [^1]

## Also worth reading from the same conversation

### *The Bitter Lesson*
- **Content type:** Essay [^1]
- **Author/creator:** Rich Sutton [^1]
- **Link/URL:** Not provided in source material
- **Who recommended it:** Anj Midha [^1]
- **Key takeaway:** Midha says the essay still holds in unsaturated domains. He contrasts saturated areas like coding with materials science, where he says more compute is still generating super-exponential gains per iteration [^1]
- **Why it matters:** This is a precise recommendation for readers trying to understand where scaling still appears most powerful, rather than treating the debate as one-size-fits-all [^1]


[![The Early Days of Anthropic & How 21 of 22 VCs Rejected It | The Four Bottlenecks in AI | Anj Midha](https://img.youtube.com/vi/a1ymdW-h33E/hqdefault.jpg)](https://youtube.com/watch?v=a1ymdW-h33E&t=104)
*The Early Days of Anthropic & How 21 of 22 VCs Rejected It | The Four Bottlenecks in AI | Anj Midha (1:44)*


### *The Feynman Lectures on Physics*
- **Content type:** Lectures / book series [^1]
- **Author/creator:** Richard Feynman [^1]
- **Link/URL:** Not provided in source material
- **Who recommended it:** Anj Midha [^1]
- **Key takeaway:** Midha says the lectures influence how he teaches because they combine technical education with life lessons [^1]
- **Why it matters:** It stands out as a recommendation about explanatory style as much as subject matter: how to teach hard things without stripping away the human element [^1]

## One empirical policy read

### *Austin’s Surge of New Housing Construction Drove Down Rents*
- **Content type:** Research article [^2]
- **Author/creator:** Pew [^2]
- **Link/URL:** [https://www.pew.org/en/research-and-analysis/articles/2026/03/18/austins-surge-of-new-housing-construction-drove-down-rents?utm_campaign=pewtrusts&utm_source=twitter&utm_medium=social](https://www.pew.org/en/research-and-analysis/articles/2026/03/18/austins-surge-of-new-housing-construction-drove-down-rents?utm_campaign=pewtrusts&utm_source=twitter&utm_medium=social) [^2]
- **Who recommended it:** Bill Gurley [^3]
- **Key takeaway:** Gurley points to Austin as a case where a surge in new housing construction drove down rents even as population grew [^3][^2]
- **Why it matters:** He frames it as a seriousness test for leaders who say they care about affordability, which makes this a useful evidence read rather than a generic policy opinion [^3]

> "If your local leaders says they 'care' about housing affordability and can't share all they have learned studying Austin, they aren't serious. They are performative." [^3]

## Bottom line

The clearest pattern today is that the best recommendations came with explicit models attached. Midha's picks offer frameworks for competition, scaling, and teaching [^1]. Gurley's Austin article adds an empirical case study with immediate policy relevance [^2][^3].

---

### Sources

[^1]: [The Early Days of Anthropic & How 21 of 22 VCs Rejected It | The Four Bottlenecks in AI | Anj Midha](https://www.youtube.com/watch?v=a1ymdW-h33E)
[^2]: [𝕏 post by @scottlincicome](https://x.com/scottlincicome/status/2044055641912680507)
[^3]: [𝕏 post by @bgurley](https://x.com/bgurley/status/2044207861610229900)