We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Chamath: GPU finance through an airline lens
Chamath Palihapitiya called Alex Song's post a "super interesting comparison of GPU finance and the AI buildout to airlines." He noted that Ken Griffin and Sridhar Ramaswamy have made the same comparison, and said both are "very, very smart so worth learning more." Song says he traded ABS (asset-backed securities) on Wall Street and has structured more than $2B of debt financings. His view is that a GPU neocloud "runs on airline economics." He also argues that the debt behind it looks like aircraft engine ABS. His parallels:
- Compute is perishable inventory, like an unsold seat.
- More than 85% of costs are fixed.
- A levered cluster breaks even at roughly 65–75% utilization.
- NVIDIA plays the role of Boeing and Airbus combined.
More detail is in Part 1 of his blog post. This follows yesterday's debate about debt versus equity funding for the AI build-out.
Research picks on efficiency and agents
- Sarah Guo called Cartesia's new H-Net scaling-law work "fundamental." She pointed to a "huge compute efficiency gain," which she says comes partly from learned tokenization. Albert Gu's post reports 3–5x efficiency gains in controlled experiments. It also says that at larger scale the model matches an open-source model trained on 11T tokens while using about 10x less data and compute.
- Eric Schmidt called Maithra Raghu's Samaya Sterling announcement "pretty extraordinary, and on an open model to boot." Sterling is a post-trained 30B mixture-of-experts model with 3B active parameters. Samaya says it scored 58.9% on FrontierFinance, against 55.9% for the best frontier model, at 11x lower cost.
- Sarah Tavel called Alex Tamkin's Anthropic research post "Provocative and rings true." The study was a randomized trial with 50 professional data analysts. It tested whether experts still make the key decisions when they delegate work to an agent.
Martin Casado's notes on VC mindsets
Martin Casado published "The Evolving Landscape of Venture Capital Mindsets." It started as notes for his team, and he shared it "for those who have an interest in venture." The essay describes four mindsets: financial, productive, frontier and speculative. It recommends that a team pick one core philosophy and apply it consistently. The alternative he warns against is pulling arguments from all four as a "grab bag" to justify personal bias. Casado describes himself as "strictly a productive investor" in early-stage infrastructure. This is his own work.
Building agents
Jason Lemkin recommended D. Ternyak's article on building Owner Agent. His takeaway: Owner ranked hundreds of thousands of old support tickets by how much support time each took. It reduced them to about 2,000 operations, which became both its roadmap and its tests. Lemkin's summary: "Your support queue is the spec."
Longtime favorites
- Keith Rabois, asked whether he is bullish on podcasts, praised Ben Thompson's Stratechery instead. He said he has read it "for like a decade plus" and called it "by far the most consistent, original, and useful thinking in tech."
- Tobi Lütke said on Alex Heath's podcast that he recently reread Neal Stephenson's Cryptonomicon after his son read it. He said "literally everything in the book just happened," including cryptocurrencies and fractional ownership. The episode is on YouTube.
- Shaan Puri cited the old book Your Word is Your Wand. He said he loves its idea that words are spells that only work on ourselves, such as swapping "have to" for "get to."
Patrick O’Shaughnessy highlights his video conversation with Cliff Sosin, calling Sosin one of his favorite people to call with big questions and listen to, and describing the conversation as a version of those calls . The discussion covers intelligence, AI commoditization and who profits, agents, capital markets, AI-driven company transformation, portfolio concentration, and education . A featured takeaway is Sosin’s argument that demand can eventually shrink as prices fall, and that Jevons paradox applies to a particular point on the curve rather than every price level . Full conversation
While discussing content that can cut through the clutter, Keith Rabois praised Stratechery, saying he had read it for more than a decade and calling it “by far the most consistent, original, and useful thinking in tech.”
Patrick O’Shaughnessy promotes his conversation with Cliff Sosin, calling him one of his favorite people to ask big questions and listen to; the conversation ranges across AI, markets, and education. In a clip about educating children in the AI era, Sosin argues for teaching foundational skills away from models while also assigning ambitious model-assisted work that students must understand and defend. Conversation post
Tobi Lütke praised Cryptonomicon by Neil Stevenson: he recently reread the sci-fi novel, said he loved it, and was impressed by how it anticipated developments such as cryptocurrencies and fractional ownership.
Reid Hoffman endorsed university-based education for policymakers, citing Stanford and Harvard programs for Congress members and staff on subjects including cybersecurity and decision-making; he praised Condoleezza Rice’s work with the Spogli Institute and Jennifer Lerner’s work at Harvard on decision science. He said similar education in technology and AI is important because even basic understanding helps officials keep pace with rapid change.
Patrick O'Shaughnessy welcomed LIFE's announced return to print on November 23, 2026, saying he grew up loving the magazine; his endorsement reflects longstanding affection for LIFE, not a recommendation of a specific article or issue. LIFE's announcement
Scott Belsky shared “Potential Risks of Governing Agents”, an excerpt from IMPLICATIONS, Edition 48. The article warns that governments could use persistent agents to comb through people’s digital histories and construct negative narratives; it argues such agents need boundaries that protect basic rights and procedures, and suggests constitutional clarification may be needed.
Sarah Guo recommended Cartesia’s work on scaling laws for H-Nets, calling it “fundamental” and highlighting large compute-efficiency gains, partly from learned tokenization; she linked the post. The linked post reports up to 3–5× efficiency gains in controlled scaling-law experiments and, at larger scale, a match to an open-source model trained on 11T tokens using about 10× less data and compute.
Jason Lemkin recommends D. Ternyak’s “Building a self-improving agent: Owner Agent”, sharing its method of turning hundreds of thousands of historical support tickets—ranked by support time—into about 2,000 operations for the roadmap and tests. His takeaway: “Your support queue is the spec.”
Tobi (@tobi) expressed strong enthusiasm for a game built into EDITIONS, saying “so much love for this game”; the linked post describes it as Katamari Damacy-like . Linked post
Keith Rabois shared Coach Bowman’s post about Michael Phelps going six years without missing a training day, adding “No days off” as the takeaway.
Keith Rabois shared the Apple Podcasts episode “The future of vertical AI startups (especially law, finance, accounting, CS, code)”; the post gives no reason for the recommendation.
Chamath called the comparison of GPU finance and the AI buildout to airlines “super interesting” and said it was worth learning more, noting that Ken Griffin and Sridhar Ramaswamy had made the comparison too . The recommended material is @alexsongis’s analysis applying airline economics and aircraft ABS to GPU neocloud financing; its linked Part 1 blog post is at http://alexyangsong.com/blog/aluminum-to-silicon/ .
Tim Ferriss recommended watching his full interview with Arthur Brooks: YouTube interview. A related clip highlights Brooks’s advice to “Use a Reverse Bucket List.”
Tobi urged followers to check out his Expo Mode post/video . The post says Expo Mode was built with Google’s new HTML-in-Canvas API and that they think they are among the first to use it .
Vinod Khosla shared an approving post linking to a page with 20 of 300 examples of Agent Fo successfully challenging bills, fees, or denied refunds. He framed fighting medical-payment denials and debt as a major use case, noting that most examples involved phone calls, escalation to supervisors, written follow-ups, and repeat calls: linked page.
@tobi called @randgroup’s post “Absolutely astonishing” and linked to it: the post. The post claims the Soviet Union collapsed at 50%, while France is at 57% and striking for more.
Tobi Lütke praised Alex Heath’s conversation with him as a “Great conversation,” sharing Heath’s post about their interview/podcast on AI agents and Shopify . The discussion covers Shopify’s approach to AI agents, how they may change shopping, and AI’s effect on starting a business . Watch the episode on YouTube .
Sarah Guo recommends Conviction Embed workshops, calling one featuring @lulumeservey one of her favorites and describing Meservey as a “master of her craft.” Post
Shaan Puri shared a YouTube conversation with Ethan Evans, who spent 15 years at Amazon working with Jeff Bezos. They discussed what it was like working for Bezos, his “?” emails, the story of Amazon Prime, and whom Bezos studied; the post gives these topics but no explicit evaluative reason for recommending the video.
