July 2026
Leopold Aschenbrenner says Situational Awareness remains an AI-focused hybrid fund after a July loss estimated at 67%, but has eliminated leverage, closed its shorts and sold part of its public portfolio to Citadel to avert a more damaging liquidity event. John Coogan and Tyler Cosgrove argue that the episode does not by itself disprove the AI trade, which Big Tech earnings continue to support, but exposes how concentrated, levered positions can become untenable amid sharp swings in AI-linked stocks.
The All-In hosts argue that the AI boom’s long-term productivity promise is colliding with immediate financial and political constraints: a chip-stock selloff exposed the danger of leverage, while higher Treasury yields are raising the cost of betting on distant AI returns. David Sacks maintains that frontier labs’ revenue and compute access support the infrastructure buildout, but Chamath Palihapitiya and David Friedberg question where the economics will ultimately accrue as open models, energy limits and cheaper alternatives reshape the market. They also cast the fight over AI safety, training data and regulation as a contest over who gets to control the technology’s future.
Bloomberg’s Peter Elstrom reports that Moonshot built its Kimi K3 model partly with compute supplied by investor Alibaba, using roughly 20,000 Nvidia Hopper-generation chips, according to Bloomberg’s sources. Alibaba acknowledges supplying the chips but denies that its compute uses Nvidia’s H200s, while the sources identify the cluster as H200-based. Elstrom argues that the arrangement exposes a limit of US export controls: restricting chip sales into China does not necessarily prevent Chinese AI companies from accessing advanced Nvidia compute.
Patrick Collison argues that founders should scrutinize success as seriously as failure: before raising money, they should ask whether they want to spend the next decade or more running the company they hope to build. Drawing on Stripe’s nearly two-year path to public launch, he says the essential discipline is not launching quickly but reaching production users early and letting their needs shape the product. In the AI era, Collison sees more opportunity rather than a closing window, though he says founders may need to pursue less crowded starting positions while staying anchored to real customer demand.
Peter Santenello argues that the Old Order Amish prohibition on driving is not a rejection of using cars: they can ride with non-Amish “English” drivers, a rule that gave one Amish man, Marcus, leverage to secure a ride by offering Santenello a filming opportunity that never materialized. Santenello’s account, while noting that Amish communities’ rules vary, underpins Chris Williamson’s narrower point that constraints developed within a religious or cultural tradition—such as setting phones aside for shared meals—can offer useful social structure to people outside it.
Stanford neuroscientist Andrew Huberman argues that his six-workout routine is governed by recovery rather than a fixed weekly calendar: hard, low-volume resistance sessions are followed by rest or conditioning according to how much fatigue they create. He pairs sets taken to failure with compound lifts, uses high-intensity cardio that can be repeated without undue injury risk, and reserves a final day for slower outdoor movement. The specific machines and exercises are personal choices, he says; the underlying structure is recovery-led sequencing.
Shaan Puri argues that Ferrero built a major consumer business not by making radically better chocolate, but by turning ordinary products into rituals: Nutella became a breakfast habit, Tic Tac an invitation to share, and Ferrero Rocher a small gift. He presents its continuity across generations as a series of different tasks—finding a viable spread, creating consumer rituals, then expanding through acquisitions—while the family’s privacy and early exposure to the business helped protect the craft behind its brands.
Princeton historian Sean Wilentz argues that American democracy has been shaped by two intertwined revolutions: one against arbitrary monarchy and for popular sovereignty, the other against chattel slavery. In his account, neither the Declaration’s equality principle nor the Reconstruction Amendments secured democracy on their own; rights advanced when political mobilization was backed by enforceable federal power, and receded when that protection weakened. He places voting rights and constitutional limits on executive authority at the center of that unfinished struggle.
Alejandro Ao presents Tau as a Python implementation of Pi’s coding-agent harness, built to preserve Pi’s parent-linked sessions, core tools, skills, and event design rather than introduce a different agent architecture. He argues that Tau’s main departure is its Textual terminal interface, which makes a run’s context, tool activity, resource provenance, branching history, and exports visible alongside the conversation.
Edward Maibach argues that phasing out fossil fuels should be treated as an immediate public-health intervention, not only a response to future climate risks. Alongside pediatrician Lisa Patel, sanitation advocate Catherine Coleman Flowers and air-quality entrepreneur Darren Riley, he points to local benefits—from fewer asthma exposures near schools and in homes to safer water during floods—that can make decarbonization a more concrete policy choice.
Former Bureau of Land Management director Tracy Stone-Manning argues that public lands will become vulnerable to privatization or transfer if federal agencies are too depleted to manage them effectively. Joined by Patagonia’s Corley Kenna and conservation advocate Benji Backer, she makes the case that stewardship must pair federal capacity with local participation, Indigenous co-management, and a broader coalition built around the places where people experience climate and environmental change firsthand.
Stanford computer science professor Chris Ré presents expectation-maximization as a local optimization method for latent-variable models: it alternates between estimating posterior assignments and refitting parameters through a tight ELBO, ensuring that observed-data likelihood does not decrease without guaranteeing a global optimum. He contrasts it with PCA, which compresses centered, appropriately scaled data by projecting it onto covariance eigenvectors that capture the most variation. Ré cautions that PCA’s individual coordinates can be unstable when relevant eigenvalues are close.
Stanford computer scientist Chris Ré argues that unsupervised learning can recover useful structure only by making explicit assumptions about what generated unlabeled data. In CS229’s treatment of k-means and Gaussian mixture models, he presents k-means as a local, least-squares method for compact clusters and GMMs as a probabilistic model of latent Gaussian sources, with EM alternating between estimated memberships and parameter updates to improve likelihood without guaranteeing a global optimum.
Chris Ré presents Gaussian discriminant analysis as a generative alternative to logistic regression: rather than learning \(P(y\mid x)\) directly, it models how features arise within each class and combines those distributions with class prevalence. Under GDA’s shared-covariance assumption, the model’s quadratic terms cancel, producing a linear decision boundary and the same sigmoid-form posterior logistic regression can express. Ré argues that the closed-form efficiency of GDA—and of naive Bayes for text—comes from structural assumptions that can improve data efficiency when plausible but become liabilities when the data-generating story is wrong.
Stanford computer science professor Chris Ré argues that modern models can interpolate their training data and still generalize, but that result does not weaken the case for disciplined evaluation. In this CS229 lecture, Ré frames generalization as the stability of a learning procedure across plausible training samples, explains how regularization can trade bias for lower variance, and insists that model selection belong on development data while a final test set remains protected from tuning.
Investigative journalist Christo Grozev argues that Russia’s growing use of private forces, proxies and hired operatives has made its foreign operations less predictable and harder for Western governments to deter. Speaking with The Wall Street Journal’s Yaroslav Trofimov after a screening of Antidote, Grozev said Moscow treats its confrontation with the West as an active war while relying on a dispersed network whose violence may exceed the control or competence of any single command structure. He also contends that corruption, economic strain and elite self-interest are eroding the system behind that threat, though a post-Putin Russia would not necessarily become democratic.
John Coogan and Jordi Hays argue that both a leveraged AI-infrastructure fund and Ferrari’s first electric vehicle must be judged by the structures that carry them to their intended outcomes. Coogan says Situational Awareness’s reported forced unwind does not necessarily invalidate its AI thesis, but shows how leverage and volatility can end a trade before it can play out. On Ferrari’s Luce, Hays questions whether meeting a reported sub-500-unit 2026 target proves commercial success, while Coogan sees its low-volume strangeness as part of its appeal to collectors.
Google chief scientist Jeff Dean argues that AI founders should test frontier models on the workflows they hope to build around: if a general model already succeeds meaningfully, rapid model improvement may erase the opportunity. More durable openings, he says, are tasks where the model succeeds only 0% or 1% of the time and where a small team can add private context, distinct data, specialized models or better evaluation loops. The larger question is whether that advantage will persist long enough—and whether solving the problem matters.
Rivian’s route to automotive gross-profit positivity depends on scaling production of its Gen 2 vehicles at its Normal, Illinois plant and spreading fixed costs across its R1 and commercial-van programs, CFO Claire McDonough told Bloomberg. She said early demand is evident in 57,000 second-quarter demo drives and encouraging Launch Edition conversions, though Rivian has not disclosed conversion rates or Gen 2 delivery volumes; fleet software subscriptions remain its established source of recurring software revenue.
Steve Dawson and Jerry Rubin argue in Help Wanted that employers should treat the tight labor market as a structural condition, not a temporary hiring cycle that will reverse on its own. They contend that demographic change, lower workforce participation, and limits on immigration require businesses to redesign how they recruit, train, advance, and retain workers—starting with the barriers that screen out or drive away people they need. Their case is that employee-centered practices are not a generic fix but an operational response that must account for each employer’s job design, incentives, and workforce.
Climate scientist Kate Marvel argues that a warming planet is already intensifying heat, storms, drought, fire weather and sea-level rise, but that climate models describe conditional futures rather than an inevitable outcome. In a live recording of Scientista, Marvel tells co-hosts Sweta Chakraborty and Monica Medina that ending carbon-dioxide emissions would, by the best current estimate, produce no further warming beyond that point—and that communicating the stakes requires scientists to acknowledge feeling as well as fact. Earth, she argues, remains the only “Good Planet” people have to protect.
Climate scientist Sarah Kapnick argues that climate adaptation has become a present-tense management problem because the statistical patterns underlying historical weather and infrastructure design no longer reliably describe current risk. She says companies and governments should treat flood protection, operational continuity and financing as revisable planning decisions—made at both asset and community scale—rather than rely on return periods or wait for a disaster to force action.
K2 Space has raised $500 million to expand production of high-power satellites, arguing that spacecraft once built as multiyear, billion-dollar projects can be manufactured at far lower cost and on more reliable schedules. CEO Karan Kunjur says the company’s first 20-kilowatt satellite, now operating in orbit, has supplied the data for its Block Two design and helped drive more than $1 billion in signed contracts. The company’s next test is turning that first flight into repeatable delivery, while developing a larger platform intended to support compute-intensive infrastructure in space.
Qualcomm CEO Cristiano Amon argues that memory shortages and higher prices—not weaker consumer demand—are depressing handset volumes and margins, while Apple’s faster move away from Qualcomm adds to the near-term pressure. He says the company’s growth case is increasingly tied to automotive, industrial AI and data centers, with non-handset businesses projected to generate $40 billion in revenue by fiscal 2029 and make up about two-thirds of the company’s sales.