We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Top pick: the new Works in Progress issue, recommended by Patrick Collison
Collison called the new issue of @WorksInProgMag a "remarkable assemblage" . The pieces he listed include:
- Robin Rivaton and @bswud on how competition between China's cities drove its "rise to techno-industrial preeminence," and @danwwang on second-tier cities in China and the US
- Tyler Cowen on Brazilian food, @vpostrel on the invention of supermarkets, Charles C. Mann on the discovery of porcelain, and @salonium on "the myth of accidental scientific discovery"
- Stephen Kotkin on "Leninism's fatal flaw," and @pietergaricano on Ukraine as the future of war
The post he quoted says the print edition has been redesigned with a larger format and premium paper. Subscribers get copies this week, and anyone who subscribes by October 16 also gets one: worksinprogress.co/print .
Bill Gurley: Mustafa Suleyman's essay on AI and jobs
Gurley called Suleyman "the statesman AI needs" and said "we all need to take the 'hyperbole' way down." In his view, "doomerism is just the mirror image of the positive hyperbole," and he dismissed claims like "We are going to cure cancer in three years!" .
The article Gurley linked summarizes Nobel laureate Daron Acemoglu's piece in the first issue of The Humanist Review, a new magazine published by Suleyman's MAI . Acemoglu's main points:
- AI will likely add about 1.5% to GDP over 10 years .
- Only about 5% of what humans do should be expected to be replaced .
- Builders should make "pro-worker AI" that helps people do their jobs better .
The previous brief covered Gurley's earlier Suleyman pick. This one is different: Gurley states his own view directly.
Bill Gurley: Matt Ridley's conversation with NIH director Jay Bhattacharya on Covid origins
Gurley says anyone interested in current views on Covid origins should "immediately watch" @mattwridley's conversation with Jay from last March. He calls it "an amazing summary of what we have learned over the last 6 years" and says it "should be considered 'must watch'" . Watch.
Patrick O'Shaughnessy: his conversation with Andrew Huberman (his own show)
O'Shaughnessy says he didn't expect them to spend so much time on neurotechnology and "how we'll soon be able to precisely measure and change our brain states, both awake and asleep," and that this part was "fascinating" . The episode also covers energy, falling asleep quickly, Neuralink and brain-computer interfaces, and the state of health media. One chapter is titled "AI Companies Could Become Biotech Companies" . YouTube · Spotify
Sam Altman: Kevin Roose's new book
In Part One of his Vanity Fair Fair Game interview, Altman said AI progress is mostly "a story of… a relatively small number of brilliant scientists and engineers… much less a story about me." He added that "Kevin Roose's book, the new book, does a good job of bringing out some of the supporting parts" . The excerpt doesn't name the book's title.
Shorter mentions
- OpenAI's post on AI progress in mathematics, from Altman (his own company). His framing: "We are entering a new era of discovery now" .
- The Personal Agent Protocol announcement, via Sarah Guo. It's an open standard that Meta and Sierra are developing with Shopify, Stripe, Walmart and others. It sets out how personal agents interact with businesses, and anyone can implement it . Guo framed it as an answer to "how will businesses talk to their customer's personal agents?" and called it "relentless innovation" from Sierra . This follows Andrew Chen's question in the last brief about whether agent network effects will end up in open protocols.
- The Battle of Polytopia, made by @midjiwan, via Brian Armstrong . Armstrong calls it "one of the most underrated strategy games" and "being a founder, compressed into 20 minutes." The business parallels he sees: harvesting resources, investing in technology, balancing offense and defense, and finding a way through "even when it looks to be over" . He disclosed that Coinbase is sponsoring this year's Polytopia World Championship .
Charlie Munger recommends Lights Out, a book chronicling GE’s decline, as required reading for business schools and urges people to buy and read it. The recommendation follows his criticism of GE and Jack Welch, whom he says became so focused on winning that he lied about what had been accomplished.
Patrick Collison (@patrickc) called the new Works in Progress issue a “remarkable assemblage,” endorsing it through this selection of pieces:
- Robin Rivaton and @bswud on how competition among Chinese cities helped drive China’s techno-industrial rise; Dan Wang on Chinese and American second-tier cities.
- Tyler Cowen on Brazilian food; @vpostrel on the invention of supermarkets; Charles C. Mann on the discovery of porcelain; @salonium on the myth of accidental scientific discovery.
- Stephen Kotkin on Leninism’s fatal flaw; @owl_posting on biomedical hedge funds; @PMArslanagic on the formula for lovable cities.
- @pietergaricano on Ukraine as the future of war; @PatRubenFitz on the Bronze Age Collapse; @SCP_Hughes on the social lives of hard-partying Victorian elites. Print-edition page: Works in Progress.
Bill Gurley endorsed Mustafa Suleyman’s message to dial back AI hype, calling him “the statesman AI needs” and saying doomerism mirrors positive hype. Gurley’s post links to Suleyman’s X article, “AI won't take your job anytime soon. In 10 yrs, only 5% of what humans do will be replaced by AI”, which presents Daron Acemoglu’s case for building AI to improve people’s work rather than replace them.
Sam Altman praised Kevin Roose’s new book for bringing out the contributions of the scientists and engineers behind OpenAI’s work, which Altman described as a collective achievement rather than primarily a story about himself.
Patrick O’Shaughnessy recommends his video conversation with Andrew Huberman, calling it “a fun one” and saying the discussion of neurotechnology and measuring and changing brain states while awake and asleep was fascinating. The conversation also covers energy, sleep, Neuralink and brain-computer interfaces, health media, and curiosity. Conversation link.
Patrick O’Shaughnessy recommends his conversation with Andrew Huberman, calling it “a fun one” and inviting people to enjoy it. He says its discussion of frontier neurotechnology—measuring and changing brain states while awake and asleep—was unexpectedly extensive and fascinating. Watch the conversation.
Patrick O’Shaughnessy recommends his conversation with Andrew Huberman, calling it “a fun one” and inviting listeners to enjoy it . Watch the conversation . O’Shaughnessy says he found their discussion of neurotechnology and precisely measuring and changing brain states fascinating; they also cover energy, sleep, Neuralink and brain-computer interfaces, health media, and curiosity .
Sam Altman shared OpenAI’s page on sharing AI progress in mathematics, framing it as evidence that “we are entering a new era of discovery now.”
Tobi Lütke suggested entering Moritz Kremb’s video—which Kremb says he asked Claude to make about what it is like to be European—into Eurovision, without stating a reason.
Patrick O’Shaughnessy recommends his conversation with Andrew Huberman, calling it “a fun one” and sharing it on YouTube, Spotify, and Apple Podcasts. He says he was especially fascinated by their discussion of neurotechnology and the prospect of precisely measuring and changing brain states while awake and asleep. Watch on YouTube.
Sarah Guo shared Brian Taylor’s Personal Agent Protocol announcement, asking how businesses will talk to customers’ personal agents and praising Sierra’s work as “relentless innovation.” The announcement describes an open standard that Meta, Sierra, and industry partners are developing to define how personal agents interact with businesses; it is open for anyone to implement.
Jason Lemkin promoted a new article, “How Much of Your Customer Support Can AI Really Resolve? The Best Get About 70%...”, highlighting Gorgias’s test of 13 AI agents on real stores: the best fully resolved 70%, the median 48%, and AI bundled with a helpdesk was the weakest option.
Bill Gurley recommends Matt Ridley’s conversation with Jay about current views on COVID origins, calling it an “amazing summary” of what has been learned over six years and “must watch.” Watch the conversation.
Elon Musk shared “SpaceX talk”, linking to a Grok Bot summary of SpaceXAI President Michael Nicolls’s IAC 2026 presentation; he gave no reason for sharing it.
Brian Armstrong recommends The Battle of Polytopia, calling it “one of the most underrated strategy games.” He says it is easy to start but takes a lifetime to master, and compares its resource gathering, technology investments, offense-defense tradeoffs, and comeback tactics to business—“being a founder, compressed into 20 minutes.” The game is made by developer @midjiwan; Armstrong made the recommendation while noting Coinbase’s sponsorship of the Polytopia World Championship.
Garry Tan highlighted Mike Ege’s article for The Voice SF, “A Coup Attempt at the DCCC”, shared via Susan D. Reynolds’s post, as a case study in activism presented as journalism manipulating San Francisco politics .
Elon Musk shared @mattyp’s post with the label “Grok @Bot tips,” linking to the post. The linked video is titled “I gave Grok Bot a printer. Now, it helps me plan my day”; its outline covers pulling calendar, X, and workout information, automation, and why to use a physical printer, with day-planning as the stated payoff.
Elon Musk amplified Starlink’s post linking to space-safety.com, adding “Make no mistakes.” The linked post says safe space operations require operators to share satellite ephemeris data, predicted trajectories, and planned maneuvers.
Palmer Luckey directs readers to Anduril’s post announcing a $3.7B investment in Arsenal-2, described as “The next great American Shipyard”. He says the U.S. needs many more nuclear submarines, is behind, and cannot meet domestic or allied demand; Anduril will do its best to address this.
Alexandr Wang directed followers to The Daily Show’s segment, where Jon Stewart talks with Meta’s Muse AI agent, after joking that he was clearing his schedule to give Muse media training.
