August 2026
Bloomberg’s Ed Ludlow reports that Broadcom is discussing an AI-infrastructure financing structure that could total $100 billion across senior and junior debt, rather than a single $100 billion corporate borrowing. Under the proposed arrangement, an SPV financed by lenders and investors would buy Broadcom equipment and lease it to an end tenant, while Broadcom backstops the $60 billion to $70 billion senior-debt portion. Ludlow argues that the structure leaves debt central to the transaction but distributes ownership, borrowing and equipment-use obligations among several parties.
Scott W. Atlas argues that American healthcare’s high costs and uneven access stem less from insufficient coverage than from a system that shields patients from prices and limits competition among providers. He favors consumer-directed insurance and health savings accounts, transparent pricing for scheduled care, and fewer restrictions on physician training, licensing, and nonphysician practice. Atlas also contends that Medicaid enrollment does not ensure access, while drug-price controls risk weakening the incentives for pharmaceutical innovation.
Scott W. Atlas argues that U.S. health policy should be judged by whether patients can obtain effective care in time, not by insurance coverage, spending levels, or stated intentions. He contends that single-payer systems achieve lower costs partly by restricting treatment, technology, and specialist access, while the United States delivers stronger outcomes through greater capacity and innovation. Atlas also links the post-COVID decline in institutional trust to what he calls the false claims and overlooked harms behind pandemic restrictions.
Scott Atlas argues that public health can rebuild trust only by showing greater discipline under uncertainty: testing evidence, admitting limits, and permitting dissent rather than enforcing consensus. He says COVID-era officials and physicians too often substituted conformity and fear for judgment, then failed to revisit the assumptions behind prolonged restrictions. Atlas extends that critique to prevention policy, arguing that government should inform people about health risks but not use aid programs or emergency authority to direct private choices.
Chris Williamson argues that opposite-sex friendships can be behaviorally platonic while still masking an asymmetry in attraction and perceived intent. Citing research and polling, he says men are more likely to report attraction to female friends and to assume it is reciprocated, while women are more likely to view such friendships as genuinely non-romantic. His narrower case is not that platonic friendship is impossible, but that friends may be describing different internal realities.
Consumer researcher Lance Katigbak argues that GDP and other economic measures can show growth without revealing whether citizens believe they are moving toward the lives they want. Drawing on a nationwide survey of more than 1,500 Filipinos, he says governments should track ambitions, barriers and perceived progress alongside conventional indicators—starting with the widespread desire for financial security against health shocks. Katigbak’s proposed “Department of Dreams” would use that data to test whether health, labor, education and industry policy aligns with citizens’ goals.
Nellie Liang of the Brookings Institution argues that stablecoins could widen global access to dollar-denominated value and add demand for Treasury bills, but only if their growth expands the dollar market rather than shifts money from existing dollar instruments. In her paper with Brent Neiman, Liang contends that the potential gains in payments and dollar use carry policy costs: regulators must address illicit-finance risks and possible losses of bank lending, while Treasury must account for reserve demand concentrated in short-term debt.
Germany’s industrial model has become a drag on growth, Bloomberg’s Marilen Martin argues, as the advantages that sustained it—cheap Russian gas, open global trade and leadership in high-value manufacturing—have eroded. Economist Martin Ademmer says the decline in output since 2018 increasingly reflects a lasting loss of economic potential rather than a cyclical downturn. Higher energy costs, Chinese competition and a more protectionist trading environment now leave Berlin trying to defend industry with limited fiscal and political room to maneuver.
Astrocade CEO Amir Sadeghian argues that generative AI can lower the cost of making interactive experiences, but cannot replace the human judgment required to make a game worth playing. His company pairs natural-language generation with an editor for tuning mechanics, pacing and other precise choices, on the premise that “taste is the new programming language.” As more people publish AI-made games and other interactive content, Sadeghian says discovery and recommendation will determine whether creators can find an audience.
Bloomberg’s Brody Ford reports that Meta has become one of Microsoft’s largest AI customers, spending hundreds of millions of dollars annually on Azure and consuming trillions of tokens a week, apparently largely for coding assistance. Ford argues that the relationship underscores how Microsoft’s AI revenue remains concentrated among a small group of large technology buyers, despite its broader customer base. Meta’s interest in building an API business could eventually turn a major Azure customer into a more direct supplier of AI services.
Castelion’s central challenge is not building a hypersonic weapon but manufacturing thousands of them at low cost, according to co-founder and CFO Andrew Kreitz. The defense startup, valued at $13 billion after raising $1 billion, plans to begin producing its Blackbeard strike system next year while advancing a larger long-range weapon and interceptors. Kreitz says his experience in SpaceX’s finance organization shaped the company’s view that defense-system costs reflect supply-chain choices and customer incentives, not an unavoidable constraint.
Economic pressure on Iran cannot reopen the Strait of Hormuz or produce a durable settlement unless it is tied to clear objectives, sustained enforcement and credible military backing, John Cochrane, Niall Ferguson and H.R. McMaster argue. Cochrane doubts Washington will bear those costs; Ferguson says earlier military choices have narrowed its options, while McMaster contends that Iran can still be weakened through an integrated strategy extending beyond economic measures alone.
Adam Semel argues that cities should treat energy storage as civic infrastructure rather than a remote grid asset. His proposed Gravity Bank system would package pumped-hydro storage into towers that, he says, could store electricity, cool data centers and industrial facilities, and hold water reserves near urban demand. The case rests on the view that rising power needs require durable, locally integrated systems built to manage electricity, heat and water over generations.
Retired Navy SEAL officer Jocko Willink argues that military effectiveness depends less on raw toughness or aggression than on an institution’s ability to select, direct and restrain it. Drawing on SEAL selection, hazing, difficult personnel and working dogs, he says physical capability and willingness to fight become useful only when they are tied to a defined mission, accountable leadership and clear limits.
ElevenLabs says its generally available Eleven v3 Conversational model is built for realtime voice applications where delivery is part of the response, not merely a rendering of text. The company pitches audio tags for fine-grained expressive control, consistent streaming quality, and support for more than 70 languages, positioning the model for support agents, assistants and interactive characters that need to convey reassurance, urgency or tension as an exchange unfolds. ElevenLabs offers it through ElevenAgents and its API from $0.05 per 1,000 characters, with lower rates at scale.
Science.xyz co-founder and CEO Max Hodak argues that neural interfaces should be judged less by the BCI label than by whether they can restore the sensory and motor pathways through which the brain receives and acts on the world. He presents Science’s PRIMA retinal implant, which has European marketing approval, as an early proof: it bypasses damaged photoreceptors to give some blind patients form vision rather than ordinary sight. Hodak’s broader case is that treating the brain as a computational system could lead to devices that repair lost capabilities while offering a new way to study intelligence.
Evergrande founder Hui Ka Yan has been sentenced to life in prison after pleading guilty to charges including illegally taking public deposits and fundraising fraud, Bloomberg reports, citing Chinese media. Bloomberg’s Trista Xinyi Luo says 56 others, including Hui’s two sons, were also sentenced and Hui’s personal assets will be confiscated. But she argues the ruling does not resolve creditors’ recovery prospects, which depend on whether offshore liquidators can pursue Hui’s global assets and potential claims against auditor PwC.
Bloomberg Investigates reporter Matthew Campbell argues that Myanmar’s civil war has become an income system as well as a struggle for territory and political control. With the military and its allies controlling an estimated 40% to 45% of the country, competing armed groups can monetize mining access, border crossings, illicit trade and scam compounds—revenues Campbell says can make continued war more profitable than peace. The result, the reporting shows, is borne by civilians displaced, impoverished and separated by a conflict with no decisive end in sight.
ElevenLabs argues that localizing YouTube thumbnails is a design problem as much as a translation task: foreign-language headlines must fit the original composition without altering logos, faces, backgrounds or visual hierarchy. Its ElevenCreative Flows tutorial shows creators how to generate language-specific versions from one master image, with each language branch verified by an LLM that checks the headline’s translation and meaning. The company presents the resulting template as a repeatable alternative to redesigning each thumbnail manually.
Economist Valerie Ramey argues that the 2008 U.S. rebate checks—and comparable payments in Singapore—produced visible jumps in disposable income and saving, but little aggregate increase in consumption. That pattern, she says, is inconsistent with household-level estimates suggesting recipients quickly spent 50 to 90 cents of every payment dollar; her re-estimates and aggregate evidence put the 2008 transfer multiplier below 0.2. Ramey distinguishes such temporary transfers from direct government purchases, which create demand without relying on households’ spending decisions.
Valerie Ramey argues that fiscal stimulus should be assessed not only by its short-term lift to demand but by the debt and interest costs that remain after a crisis. While research showing high household spending from temporary transfers strengthened the case for emergency support during the financial crisis and COVID, she says it did not establish that the resulting borrowing would pay for itself. In her estimate, those two episodes account for roughly 27% of the current US debt-to-GDP ratio.
Valerie Ramey argues that fiscal stimulus should be judged by the economic gain it produces relative to the debt it adds, not by the amount government spends. She says broad COVID-era payments were defensible as emergency income insurance in spring 2020, when targeted aid was impractical, but later rounds were harder to justify as the economy reopened. For short-run stabilization, she places greater weight on government purchases that can occur quickly and cautions that tax changes and deficit reduction carry their own, often larger, economic effects.
Two Minute Papers host Károly Zsolnai-Fehér argues that DeepSeek V4 Pro 0813 makes open model weights a material alternative to dependence on a single AI provider: although DeepSeek has raised API prices, developers can retain the MIT-licensed model and run it through competing hosts or their own infrastructure. He says the release delivers its largest benchmark gains in software engineering and harder data-science tasks, while still trailing the cited closed-model baseline in full-stack and tool-use tests. Zsolnai-Fehér attributes the improvement to post-training specialist models, multi-teacher distillation and DSpark inference techniques rather than a new base architecture.
Fuse CEO JC Btaiche says the company’s $100 million financing will fund faster, cheaper production of fusion generators while supporting a radiation-testing service already operating for customers. He argues that commercialization will depend less on scientific milestones than on repeatedly building and operating hardware, with private companies playing a central role in helping the US counter China’s investment and construction pace in fusion.