# AI Coding Caution, Red Plenty, and General Magic *By Recommended Reading from Tech Founders • June 8, 2026* Today's strongest signal is Chamath Palihapitiya's share of a paper questioning how much AI coding translates into shipped software. It sits alongside two durable historical picks: Marc Andreessen's co-sign for Red Plenty and Tony Fadell's recommendation of General Magic. ## Most compelling recommendation ### Unspecified SSRN paper on AI coding tools' productivity impact This is the strongest pick in today's set because it pairs a concrete quantitative claim with a clear operating principle. Chamath Palihapitiya shared a paper presented as showing that AI coding tools pushed commits up 180% while releases rose only 30%, then added his own warning about what happens when teams use AI coding without clear intent [^1][^2]. - **Title:** *Not provided in the notes* — described as an SSRN paper on AI coding tools' productivity impact [^1][^2] - **Content type:** Research paper [^2] - **Author/creator:** Not specified in the provided notes - **Link/URL:** Shared via [this post](https://x.com/rohanpaul_ai/status/2063756891193549168) [^2] - **Who recommended it:** Chamath Palihapitiya [^1] - **Key takeaway:** AI-assisted coding activity can rise much faster than shipped output; in this share, the gap is framed as commits up 180% versus releases up 30% [^1][^2] - **Why it matters:** It is a practical check against mistaking more generated code for proportionate product progress. Chamath's own summary is the reason to save it: lack of upfront intent turns AI coding into "AI slop" [^1][^2] > "Using AI to code, without a clear intent upfront, is just AI slop waiting to happen." [^1] ## Two durable case studies ### *Red Plenty* — Francis Spufford Marc Andreessen's contribution today was a co-sign on another user's recommendation, but the reason attached to it is strong enough to keep. The book is described as a view of the Soviet economic planning system through the people working inside it, from factory managers to mathematicians and "fixers" [^3][^4]. - **Title:** *Red Plenty* [^4] - **Content type:** Book [^4] - **Author/creator:** Francis Spufford [^4] - **Link/URL:** Review link: [chicagoboyz.net/archives/71068.html](https://chicagoboyz.net/archives/71068.html) [^4] - **Who recommended it:** Marc Andreessen, via a co-sign of the linked recommendation [^3] - **Key takeaway:** The book covers the Soviet economic planning system through factory managers, economic planners, mathematicians, computer scientists, and "fixers" [^4] - **Why it matters:** The recommendation points readers to a systems book grounded in front-line perspectives rather than a purely abstract account of planning [^4] ### *General Magic* Tony Fadell's documentary recommendation is the clearest product-timing lesson in today's set. He recommends it specifically as a story about building something impressive far before the market was ready for it [^5]. - **Title:** *General Magic* [^5] - **Content type:** Documentary / movie [^5] - **Author/creator:** Not specified in the provided notes - **Link/URL:** Recommendation context: [Lenny's Podcast interview](https://www.youtube.com/watch?v=RJjl1TwyfWM) [^5] - **Who recommended it:** Tony Fadell [^5] - **Key takeaway:** Fadell frames the story as "the iPhone 15 years too early" — a case where they were making things that were technically exciting but that "nobody needed" yet [^5] - **Why it matters:** It is a compact case study in timing, demand, and the gap between invention and adoption [^5] > "Your viewers should definitely watch the movie General Magic because absolutely we made the iPhone 15 years too early and that was a classic case where we were just making the things that were really cool but nobody needed it." [^5] [![Father of the iPod and iPhone on building taste, judgment, and creativity in the AI era](https://img.youtube.com/vi/RJjl1TwyfWM/hqdefault.jpg)](https://youtube.com/watch?v=RJjl1TwyfWM&t=2822) *Father of the iPod and iPhone on building taste, judgment, and creativity in the AI era (47:02)* ## What connects these picks The throughline today is **fit**. Chamath's paper questions whether AI-generated coding activity maps to real shipped output [^1][^2], Andreessen's co-sign points to a book about how a planning system looks from inside the apparatus [^3][^4], and Fadell's documentary pick shows what happens when a product arrives before demand does [^5]. Together, they make a useful short list for readers who care less about novelty and more about whether work, systems, and products actually connect to reality. --- ### Sources [^1]: [𝕏 post by @chamath](https://x.com/chamath/status/2063847703516451000) [^2]: [𝕏 post by @rohanpaul_ai](https://x.com/rohanpaul_ai/status/2063756891193549168) [^3]: [𝕏 post by @pmarca](https://x.com/pmarca/status/2063776648235716737) [^4]: [𝕏 post by @DavidF1344](https://x.com/DavidF1344/status/2063747856675020874) [^5]: [Father of the iPod and iPhone on building taste, judgment, and creativity in the AI era](https://www.youtube.com/watch?v=RJjl1TwyfWM)