We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
The standout recommendation is Packy McCormick’s pick: he read an early draft, explains why its planetary-modeling problem resembles biology, and directs readers to spend time with the finished essay and its interactive graphics.
Start here
Natural General Intelligence
- Type / creator: Essay by Ryan Orbuch for Lowercarbon. Link:Read the essay.
- Recommended by: Packy McCormick, whose endorsement is unusually substantive: he describes reading an early draft, gives a biology-based rationale, and calls the goal worthy.
- Key takeaway: The essay proposes a “Natural General Intelligence” foundation model grounded in the planet’s state and dynamics—connecting the biosphere, atmosphere, oceans, land, ice, and subsurface so it can anticipate responses to intervention and update from observation. Its proposed loop is practical rather than mystical: ingest observations, simulate possible futures, direct new measurements toward key uncertainties, and use the Earth’s response as evidence.
- Why it matters: This is a concrete agenda for making AI useful in the physical world. It also keeps authority in the right place: the model would improve the empirical feedback loop, while choices about interventions, risks, and governance remain human responsibilities.
Three strong supporting picks
Searching for Feynman
- Type / creator: All-In Summit presentation by Bill Gurley. Links:Watch the presentation post · Resource links. The source labels it “Bill’s Presentation: Searching for Feynman” and supplies the companion resource page.
- Recommended by: Chamath Palihapitiya, who called it “incredibly important” and “so well done,” then said: “Please watch.”
- Key takeaway: The companion guide makes the recommendation more than a talk: it assembles Feynman’s “Personal Observations on the Reliability of the Shuttle,” his own account of the Rogers Commission, the commission’s full report, and the televised ice-water demonstration. Its governing lesson is that “Reality must take precedence over public relations,” paired with Feynman’s emphasis on independence from organizational interests.
- Why it matters: Read it as a primary-source path to institutional truth under pressure—not merely as a story about a famous scientist’s presentation.
The Game Industry Is Broken—Can We Fix It?
- Type / creator: Edge Magazine long-form interview / cover story built around nine industry experts. Link:Read the full interview.
- Recommended by: Tim Sweeney, who says that many quotes have circulated without context and tells readers to “read the full thing.”
- Key takeaway: The feature frames the downturn as multi-causal: macroeconomic pressure and internal dysfunction combine with escalating AAA budgets. Sweeney contrasts a $12 million Gears of War with a roughly $100 million sequel budget while the revenue projection remained about $120 million because the audience was not growing. He argues that the industry is approaching a hard audience cap and that companies must either build an ecosystem or operate inside one; the proposed direction includes cross-game social features and a more interoperable economy.
- Why it matters: It is a concrete case study in what happens when cost growth outruns audience growth—and in the strategic tradeoff between owning a platform and reaching users through someone else’s.
A long-form X essay on effective altruism and longtermism
- Type / creator: Long-form X post by @banteg. Link:Read the post. Recommended by: Ben Thompson, who places it in the context of accounting, finance, and especially discount rates, then calls it “Long but excellent.”
- Key takeaway: The post’s arc moves from measuring “people helped” to measuring the “fraction of cosmic endowment preserved,” then uses expected-value calculations built on enormous future stakes and highly uncertain probabilities. It ends by asking whether the hypothetical constituency invoked by those calculations can answer questions or falsify the assumptions behind them.
- Why it matters: This is not a neutral primer on effective altruism; it is a pointed stress test for long-horizon reasoning. Thompson’s framing makes it especially useful alongside finance basics: inspect the discount rate, the probability estimate, and the sign of the intervention before accepting a very large expected-value claim.
Direct answer: The supplied bundle identifies the essay as Natural General Intelligence; its body introduces “a new framework: Natural General Intelligence” whose core is a general-purpose nature model. The author is not named in the supplied text. The essay uses first-person authorship (“I wrote this essay”) and provides ryan+agents@lowercarbon.com for proposed contributions, but that contact clue is not a verified byline.
- Central thesis of the essay: Natural resources are produced by Earth systems rather than human intelligence, so applying AI’s “automation” frame to nature misses the opportunity; the appropriate goal is stewardship—using greater intelligence to improve what Earth systems provide while sustaining their conditions.
- Proposed mechanism: NGI would model the connected biosphere, atmosphere, oceans, land, ice, and subsurface; anticipate how the planet responds to interventions; and update from direct observations. The intended system is a closed loop in which the model guides observation, humans decide how to act, and the Earth’s response becomes new evidence.
- Important qualification: The essay says NGI would not itself solve alignment or decide which interventions society should pursue; it would provide an empirical feedback loop and make causal consequences more legible, while judgment and governance remain human responsibilities.
- Packy McCormick’s endorsement: The specific reason Packy recommends the essay cannot be verified from this bundle because it contains no Packy quotation, recommendation, or endorsement rationale. The essay’s own stated purpose is to show readers “how we could build NGI and why we should”; that is the author’s rationale, not evidence of Packy’s separate reason for recommending it.
Assessment
Direct answer: The bundle supports treating “Searching for Feynman” as a curated resource guide about Richard P. Feynman’s work on the Rogers Commission’s Challenger investigation, especially technical truth, institutional independence, and public relations—not as sufficient evidence about a standalone presentation’s creator or production format.
Substantive topic: The core material covers Feynman’s “Personal Observations on the Reliability of the Shuttle,” his own account of the Rogers Commission, the commission’s full report, and his televised ice-water demonstration. The linked book is described as Feynman’s own account of the commission, including the demonstration, commission dynamics, and his fight to include Appendix F.
Creator: The supplied page does not identify a presenter or creator for a work specifically titled “Searching for Feynman.” It instead attributes the central source material to Richard P. Feynman and identifies Ralph Leighton as editor of “Mr. Feynman Goes to Washington.”
Format: The artifact itself is explicitly a “Resource Links” page organized under “Primary Sources,” “Context,” “Book,” and “Documentary,” linking to an archive/PDF, YouTube footage, a book, and a Netflix documentary. The original presentation’s medium, length, venue, or slide/video format are not established by this source.
Useful takeaway beyond the recommendation: Feynman’s stated principle is that “Reality must take precedence over public relations, for nature cannot be fooled.” The resource also highlights his view that being unaffiliated with an organization meant he had “no weakness from that point of view,” making independence from institutional interests a central lesson.
Recommendation authenticity: This is more substantive than bare event promotion: it provides multiple topic-specific primary and secondary sources, including Feynman’s own account, the official commission materials, the televised demonstration, and a documentary. However, it remains a curated recommendation/resource list; the bundle supplies no presentation description, speaker identity, date, venue, transcript, or direct evidence that the recommender personally viewed the presentation.
Assessment: Treat this as a substantive external resource, but do not label it “recommended by Tim Sweeney” on the supplied evidence. It is identified as an Edge 428 cover story involving nine industry experts and includes a dedicated “Interview: Tim Sweeney, Epic Games” section; Sweeney is presented as an interview subject, not as someone endorsing the article.
Resource record
- Title:The game industry is broken – can we fix it?Original publisher:Edge magazine, issue 428.
- Creator/byline: No individual creator or byline is identified in the supplied text; the piece is framed as an Edge cover story built from interviews with industry figures.
- Useful links: The supplied web article is
https://knowledge.edge-online.com/p/the-game-industry-is-broken-can-we-fix-it-tim-sweeney-edge-crash; the source also provides an issue-purchase link at magazinesdirect.com/uk/single-issues/edge.
Key themes
- The feature diagnoses a severe, multi-causal industry crash involving layoffs, uncontrolled budgets and schedules, studio closures, and rising console prices; Sweeney calls it the worst videogame crash or disruption since the 1980s and attributes it to both macroeconomic pressures and internal industry dysfunction.
- Its central economic argument is that AAA development costs have escalated while the core audience is no longer growing; Sweeney argues that developers must either build an ecosystem with a massive audience or operate within an existing ecosystem.
- Sweeney’s proposed direction is an interconnected ecosystem model: Unreal Engine/UEFN integration, cross-game social features, and potentially portable cosmetics or revenue-sharing across participating games. The article also notes substantial technical and commercial obstacles to interoperability.
- The broader discussion considers practical alternatives to the AAA cost spiral, including smaller games, lower hardware requirements, leaner teams, residual-style royalties, mid-tier publishing, and avoiding growth pressures associated with public ownership.
- The interview gives a qualified view of single-player games: Sweeney says they retain strong creative potential, but their audience is shrinking and the economics have become difficult for independent studios.
- Sweeney also describes a hardware supply shock linked, in his account, to AI-system and data-centre investment, which he says is driving sharp increases in RAM and storage prices and could produce a multiyear gaming-hardware supply crisis. This should be presented as his forecast, not independently verified fact.
Briefing use: Strong as a substantive source for industry-crisis analysis and Sweeney’s own views on ecosystems, single-player economics, optimisation, and hardware constraints; weak as evidence that Sweeney personally recommends the article, because no such endorsement appears in the supplied text.
- Video — “Jerry Seinfeld How Warren Beatty changed his mind on Kids”: Elon Musk replied “Well said” to the post containing the video and linked to it, signaling an endorsement. Watch the linked post. The posts provide no stated rationale, lesson, or learning takeaway.
- @banteg — long X post on effective altruism and longtermism (link) : Ben Thompson called the linked post “Long but excellent” and framed the recommendation alongside his view that understanding finance—especially discount rates—and accounting is underrated. The post explores longtermist reasoning through protecting the far future, expected-value calculations, existential risk, and uncertainty around intervention probabilities.
- Article — “Is AI really going to kill us all?” by Clair Lehman: Bill Gurley amplified Steven Pinker’s post by calling him “A man of reason.” Pinker’s post links to Lehman’s article, which argues that apocalyptic AI-extinction scenarios are preposterous and that fatalism, panic, and distraction can draw attention away from more realistic safety challenges. URL: https://www.theaustralian.com.au/inquirer/superintelligence-is-a-fantasy-leading-scientists-reject-apocalyptic-rogue-ai-extinction-warnings/news-story/618eef10b92d4357e51e4b459eaf9fa9?giftid=JrfvmQcvVC
- Video/presentation — “Searching for Feynman” by Bill Gurley: Chamath Palihapitiya explicitly recommends watching the All-In Summit presentation, calling it “incredibly important” and “so well done.” Watch the All-In Summit post.
- The talk’s post includes a resource page with additional materials mentioned in the presentation and used in its research.
- Natural General Intelligence — Essay by Ryan Orbuch, a Lowercarbon partner, arguing that humanity should use new tools to gather much more planetary data and build a model of Earth. Packy McCormick recommends it after reading an early draft, connecting the idea to biology, where models help interpret enormous, changing systems too complex for humans to understand holistically; he also highlights the essay’s interactive graphics.
Alexandr Wang endorsed Trevin’s Travel Time Buddy AI prompt as a Muse prompt that works well, and shared it for use with Muse, Instinct, or Grok. The prompt protects calendar travel time by scanning 14 days ahead, adding live-ETA buffers around in-person meetings, and retaining a ledger of user-edited or dismissed buffers. Direct prompt
- The Game Industry Is Broken. Can We Fix It? — Edge Magazine long-form interview featuring many developers discussing the state of the game industry. Epic Games CEO Tim Sweeney recommends reading the full interview because individual quotes have circulated without the broader context. Read the interview.