Best candidate (with an attribution caveat): The post gives an explicit cross-industry learning recommendation—“Look outside your industry too”—and reports that Mark Rampolla built Zico Coconut Water with a team of eight by studying Vitaminwater, Red Bull, and Starbucks. His quoted rationale, “We became students of success,” followed by “Excellence leaves clues,” gives a meaningful reason to use these companies as external case studies. Book of the Day from The Next B…

Caveat: This is an indirect endorsement: the recommendation is narrated by the post, while Rampolla is quoted describing his team’s learning practice rather than directly telling readers to study those companies. Book of the Day from The Next B…

Excluded promotional material: The same document explicitly labels Leadership Intelligence as a sponsor and Upwork as a sponsor, so those placements do not qualify. Book of the Day from The Next B…

Direct answer: The transcript contains one clear external-resource recommendation: an “open router report” and an “index of token pricing.” 20VC with Harry Stebbings

• Resources / title and creator: The transcript gives only those descriptive labels; it does not provide a fuller title or creator for either resource. The wording is compressed, so it is also uncertain whether the token-pricing index is separate from the report. 20VC with Harry Stebbings
• Who recommends them: An unlabeled panelist speaking after Jason; the speaker says the resources are what “you look at,” but the supplied transcript does not securely identify that speaker by name. 20VC with Harry Stebbings
• Rationale: Use them instead of relying mainly on benchmarks, to judge which model generates the most economic value efficiently; the speaker also suggests checking what companies are using and how they evaluate the models. 20VC with Harry Stebbings

Borderline references, not counted as confirmed recommendations: A speaker says they occasionally read Ben Thomson and praises a phrase, but names no specific work. 20VC with Harry Stebbings Another speaker calls an untitled piece by “Jacob,” identified only as OpenAI’s chief scientist, “a great piece” and says they read it that morning; this is positive endorsement but not a direct audience-facing recommendation. 20VC with Harry Stebbings

One qualifying external-resource endorsement appears, though the transcript does not give an exact episode title:

• Resource: an untitled Joe Rogan podcast/interview appearance featuring Daniel from AI 2027; creator/host identified in the transcript: Joe Rogan. The hosts call it “an interesting listen,” say Daniel “makes a lot of good points,” and praise his ability to explain AI-risk issues to a broad audience. TBPN
• Recommender and context: the show’s hosts, while discussing AI 2027 and public awareness of existential risk, give this positive endorsement; it is an implicit recommendation rather than a direct “listen to this” imperative. TBPN
• Takeaway: the appearance is presented as an accessible explanation of AI risk, but the hosts caution that Joe Rogan’s UFO and remote-viewing tangents could blur the distinction between concrete X-risk claims and more speculative topics. TBPN
• Candidate lines: L16–L18. The exact episode title and Daniel’s full creator/guest identity are not supplied, so they should not be filled in from outside the transcript.

Other media references do not qualify: Mountain Head is explicitly criticized as not a good film, while the Luca Ferrari episode is forthcoming self-promotion; the Wall Street Journal article and other films are merely discussed or quoted. TBPN

• Hamilton Helmer’s 7 Powers (book). Packy McCormick calls counter-positioning his favorite of the book’s seven powers: designing a business model so incumbents face conflicting incentives that prevent them from competing effectively. Not Boring by Packy McCormick
• Zach Dell’s conversation on David Senra (podcast/video, around 16:18). Base Power CEO Zach Dell explains that challengers should avoid merely offering a better product at a lower price; instead, they should change the business model—for example, sell electricity rather than home batteries at roughly one-twentieth to one-fortieth of the outright purchase cost, forcing incumbents to fundamentally change how they operate. Not Boring by Packy McCormick
• Acquired, especially “7 Powers with Hamilton Helmer” (podcast/episode). The recurring discussion is recommended as a source of counter-positioning examples; its key caveat is that counter-positioning is usually a take-off-phase power and only a partial source of durability, so companies must build another moat. Not Boring by Packy McCormick

• Jacob’s AI-safety letter — The 20VC discussion called it a “great piece”; it argues that no lab, including OpenAI, has solved alignment enough to keep scaling at full speed, and calls for mandatory externally enforced safety bars plus voluntary slowing until those safeguards exist. 20VC with Harry Stebbings
• Ben Thomson’s writing — The discussion highlighted Thomson’s description of LLMs as “the most scaled artifacts humans have ever developed” and said the framing “stirred the imagination.” 20VC with Harry Stebbings
• OpenRouter report and token-pricing index — These were pointed to as practical references for evaluating models by real-world usage, economic value, and efficiency rather than relying on benchmark claims alone. 20VC with Harry Stebbings

Patrick O’Shaughnessy recommended Walter Russell Mead’s books Special Providence and God and Gold, calling them “two of the best books on American power.” @patrick_oshag

• Special Providence and God and Gold — books by Walter Russell Mead. Investor Patrick O’Shaughnessy calls them “two of the best books on American power,” in the context of Mead’s decades-long study of how America achieved global dominance and whether it can sustain that position amid AI, robotics, and “personal politics.” @patrick_oshag