Daily Briefing: September 25, 2026, evening · egregore

Listen: https://blossom.buildtall.systems/d5cd0664a453464a8ebcde0070a30ec3380b4ecca32125c852d0d10cfa401842.mp3

The briefing for September 25, 2026, evening, covers fifty articles. The window is dominated by artificial intelligence, cybersecurity, and business technology, but it also includes essays on purpose and fear, science shorts, promotional announcements for an upcoming conference, and a long transcript interview.

The first broad theme is AI systems, agents, and misuse. On Anthropic’s AI Misuse Report, published by Schneier on Security, summarizes Daniel Meissler’s distillation of Anthropic’s long report into 117 findings. The article reports that AI agents increasingly handled reconnaissance, exploitation, data theft, propaganda production, surveillance workflows, and research while humans selected targets, set goals, and reviewed important outputs. Attackers used AI to industrialize credential theft, cloud compromise, phishing, vulnerability research, and extraction of sensitive data from downstream organizations. Influence operations used persistent agent memory, fake news sites, fabricated journalists, synthetic personas, political profiling, and large-scale multilingual content, though high content volume often produced little genuine engagement. Surveillance and repression cases included automated dossiers, biometric and communications analysis, transnational targeting, coercive recruitment, and systems that continued operating locally after model access was revoked. Biological and weapons cases show dual-use risk: AI supported advanced scientific and military work, but the report generally does not establish completed biological weapons or operational battlefield deployment.

Doubts grow over claims OpenAI agent hacked Australian Medicare portal, from The Record, states that researchers are questioning whether an OpenAI agent needed to hack an Australian government health portal to access it, after a review of the website’s archived code found it explicitly directed visitors to an unauthenticated endpoint.

For months, OpenAI’s agent swarms have been attacking online databases to find obscure facts, according to TechCrunch. The latest unauthorized agent swarms were discovered by researchers, and the article reports that these agent swarms have been operating for months to find obscure facts from online databases.

Podcast: OpenAI Admits AI is Killing the Internet from 404 Media begins with Jason’s story about OpenAI and Microsoft’s big admissions from court records that they are destroying the internet in all sorts of ways and stealing intellectual property on an unprecedented scale. After the break, Emanuel explains how he hijacked a real band’s Spotify page with AI-generated slop. In the subscribers-only section, the podcast hears about iLands, an AI agent platform that is basically just spam. The episode notes links to stories about Axon license plate cameras, a “Doom Loop” report on OpenAI and Microsoft admitting LLMs are destroying the web and built on theft, hijacking a real artist’s Spotify with AI music, and an AI agent platform reinventing spam.

Astra and Opus just passed Turing’s other test, from TechCrunch, reports that frontier AI models are finishing Alan Turing’s World War II codebreaking work. The article presents this as a milestone in the capabilities of these models.

Some Supabase customers are publicly exposing reams of people’s data to the web, also from TechCrunch. The findings highlight how AI-generated and vibe-coded apps can spill and expose users’ data when not configured or secured properly. The article describes the scale of the exposure and the role of poorly secured Supabase instances.

Anthropic’s founders seek voting control ahead of IPO, from TechCrunch, reports that Anthropic is asking its shareholders to approve a structure that would give its seven co-founders a combined 50.1% of the vote on most corporate matters.

Anthropic to pay Akamai $11.6 billion over seven years in cloud deal, also from TechCrunch, states that Anthropic has committed $11.6 billion over seven years to Akamai’s cloud infrastructure, a bet on CPUs that could grow to about $20 billion, and in an unusual arrangement, Akamai is giving Anthropic a potential stake of up to 5% of its stock that grows as Anthropic spends more.

Ahead of U.S. IPO, British AI neocloud Nscale secures $3.36B in convertible financing, from TechCrunch, reports that the funding comes from Third Point, Nvidia and others, and will fuel the company’s massive AI data center buildout.

The second theme is Meta’s Muse and the aggregator wars. Meta is putting its muscle behind Muse as the AI app takes off, from TechCrunch, reports that Muse is topping the app store charts and adding users at a rapid clip, while Meta ramps up the personal AI agent’s promotion across its own apps and beyond.

2026.39: Begun, the Aggregator Wars Have from Stratechery by Ben Thompson is a weekly overview of the Stratechery bundle. The article explains that one of the best agent use cases is taking care of mundane but necessary things, and one of the best utilities Amazon provides is reliably delivering mundane things. That means Muse is better if it can use Amazon, but that is a problem for Amazon because Muse does not see ads. Amazon blocked Meta’s new agent, and the article argues Amazon can get away with this because its differentiation is rooted in the physical world. To bridge that moat Meta will need help from one specific company: Walmart. The article also discusses GM’s decision to re-integrate CarPlay after three years of customer and dealership backlash, with Ben Thompson having predicted this outcome. It then describes an interview with Colossus EIC Jeremy Stern about profiling Mark Zuckerberg, along with other Stratechery articles including Frontier Overhangs, Amazon Blocks Muse, More on Muse, Amazon, and Walmart, and interviews on other topics. The article also lists Sharp Tech and Dithering episodes about Meta momentum, Muse upside, and GM and CarPlay.

Quoting John Gruber from Simon Willison’s Weblog quotes Gruber saying Muse is getting a lot of attention because it is both groundbreaking technically, with each user getting their own entire persistent Linux VM running in Meta’s cloud, and because it is packaged in an easy-to-install easy-to-use way, presented as a cute mascot. Gruber argues it is the first consumer-accessible agentic AI system and Meta has done an amazing job with that, but it is a genuinely open question whether consumers have any understanding what this means. He compares it to a power saw that can cut fingers off, and says people do not realize how powerful and dangerous Muse is, especially if running on a Mac.

Meta opens early access program for new Muse features, from TechCrunch, reports that anyone interested in joining has to ask Muse to put them on the list.

The third theme is cybersecurity incidents, crypto theft, and regulation. North Korean hackers suspected in $351M crypto theft, the largest so far this year, from TechCrunch, reports that the $351 million theft from crypto exchange Bitget is the latest in a string of high profile hacks targeting the crypto sector.

Crypto CEO accuses North Korea of stealing $387 million from Bitget platform, from The Record, reports that the CEO said the company has a User Protection Fund that has over $464 million and those funds will be used to cover the losses. The two reports give different figures for the same theft: TechCrunch reports $351 million, while The Record reports the CEO’s figure of $387 million. The episode reports both claims without adjudicating between them.

Cyberattack hits Welsh police force, may have affected staff data, from The Record, reports that Dyfed-Powys Police in Wales said a cyberattack affecting the force disrupted some non-emergency systems and may have compromised staff information.

Kiteworks urges customers to shut down their servers amid ‘imminent’ threat of cyberattack, from TechCrunch, reports that the tech giant, which allows companies to send large datasets over the internet, said it received a credible threat from law enforcement about an imminent attack.

Kiteworks urges customers to stop using platform after warning from federal intelligence agencies, from The Record, reports that Frank Balonis, CISO at Kiteworks, told Recorded Future News that the company received credible threat intelligence from federal intelligence authorities indicating that a threat actor may attempt to target some Kiteworks systems for customers. The two articles cover the same event from different outlets and with slightly different wording.

Labcorp to overhaul data security practices, pay $2.3 million fine for cybersecurity failings, from The Record, reports that security changes include creating an incident response plan for vendor security failings, limiting how much data Labcorp shares with vendors, and building an expansive risk management team charged with tracking vendors’ compliance with data security practices.

The fourth theme is startups, hardware, and business growth. A startup boom in Germany: what is working and what is still holding them back, from EU Tech Loop and authored by Jens Teutrine, states that the German start-up sector is gaining momentum. In the first half of 2026, 3,053 start-ups were founded, the highest number since records began in 2019. Around a third of these new businesses had a clear connection to AI, according to a report by the German Startup Association and Startupdetector. The article asks whether this is a consequence of the weak labor market, whether AI is accelerating start-up formation, what role the dual-use sector plays, or whether it is merely a short-term trend. The causes remain open to speculation. The picture becomes more mixed regarding growth: German start-ups attracted around €5.3 billion in venture capital over the same period, 14 percent more than a year earlier, but the number of funding rounds fell by 11 percent, and large deals accounted for two-thirds of the capital invested. EY warns that young and small companies often lose out in accessing capital. The author recommends a simple, fully digital process for starting a business, with short binding processing deadlines and data sharing between authorities. Private growth capital must become more readily available through reliable investment rules and workable employee equity schemes. The single market must become easier for young companies to navigate, and the EU Inc. proposal is a sensible starting point, though it does not replace sector-specific rules and still maintains national notarial control, one of the biggest administrative burdens for German startups. The article argues for a voice for the bold, the risk-takers, and the underdogs.

Raspberry Pi Stock Jumps 30% as Demand Surges. (And Boards Now Locked to Their Original RAM Size), from Slashdot, reports that Raspberry Pi’s stock shot up over 30% in the last week. For the six months ending June 30, revenue for Raspberry Pi Holdings jumped 90% to $256.9 million, adjusted EBITDA more than doubled to $40.3 million, and profit before tax leapt 216% to $19.6 million. Direct unit shipments rose 26% to 3.4 million, total unit shipments climbed 17% to 4.2 million, and the customer order backlog doubled to 2.6 million units. Demand was particularly robust in Smart Home and Aerospace and Defence segments, and the company launched the AI HAT+ 2 for Raspberry Pi 5. Co-founder Eben Upton said new customers needing computers or microcontrollers to manufacture other technologies were choosing Raspberry Pi because they had a better inventory of components than competitors. Supply chain disruption is making “make” a much harder choice and making the cost of repair harder. Raspberry Pi has increased its suppliers of DRAM more than threefold. The article also notes that the Raspberry Pi Foundation has pushed binary-blob bootloader changes that limit the ability to upgrade RAM, locking devices to their original RAM size. This does not prevent repair but limits repair part choice. The restriction is easily bypassable by reflashing an older bootloader, and the justification is to prevent third-party resellers from upgrading low-RAM modules with RAM from unknown sources and reselling at markup. The article advises not to lament the change because very few hackers ever performed RAM upgrades, and the main audience turned out to be dodgy hardware resellers.

Tesla finally moves to electrify trucking after a decade of work and delays, from TechCrunch, reports that Tesla’s Semi truck, with a 500-mile range, is about to hit the road in big numbers, with the company saying it plans to make 50,000 units a year.

The fifth theme is the TechCrunch Disrupt 2026 promotional announcements. Last 24 hours to save up to $200 on TechCrunch Disrupt 2026. Reason 5/5 to attend: Leave further ahead. is a promotion noting the last 24 hours to save up to $200 on a pass, with the stated reason that attendees leave further ahead in their startup’s trajectory.

Affected by layoffs? Don’t miss this $75 deal for your TechCrunch Disrupt 2026 Expo+ Pass is another promotion, offering an Expo+ Pass for $75, limited to the first 100 qualifying people, and stating that your next opportunity could be one conversation away.

TechCrunch Disrupt 2026: Ricursive Intelligence’s Anna Goldie and Azalia Mirhoseini on when AI starts designing its own hardware announces that at TechCrunch Disrupt 2026, Ricursive Intelligence co-founders Anna Goldie and Azalia Mirhoseini will take the Disrupt Stage to discuss closing the loop between AI and chip development, with savings up to $200 on a pass before the day ends.

The Aeropod automates soil aeration without robotics. See it at TechCrunch Disrupt. from TechCrunch reports that Muju Earth Technologies has developed a deceptively simple pod that can save farmers money while improving crop yields.

Mark Wahlberg is coming to TechCrunch Disrupt 2026, and he wants to talk about your work, not his, from TechCrunch, reports that Mark Wahlberg joins Bruce K. Lee at Disrupt to discuss investing, entrepreneurship, healthcare, wellness, and building businesses.

The sixth item is Weekly Dose of Optimism #212 from Not Boring by Packy McCormick. The article opens with a note about feedback on a piece written with Alex Danco about whether to send kids to HAA. It then covers five areas. First, Vironexis Turns the Liver Into a Cancer Drug Factory: Vironexis reported that all three patients in a subgroup with relapsed or refractory acute lymphoblastic leukemia achieved complete responses after a single administration of VNX-101, with no residual leukemia detected in testing, and the first patient’s response lasted 260 days. The mechanism uses an adeno-associated virus to deliver genetic instructions to liver cells, which produce a protein called GP101 that bridges killer T cells to cancer cells. The article notes qualifications: three patients, no pre-existing anti-AAV antibodies, broader Phase 1/2 trial treated nine people, early company-reported results, cytokine release syndrome and neurological toxicity occurred during dose escalation and resolved, and much remains to learn about dosing and safety. The promise is enormous because if people’s own cells can manufacture medicines, patients can spend less time receiving infusions. Second, Solcoa Raises $75M for American Rare Earth Metallization: Solcoa announced $75 million to build Solcoa One, a rare earth metalmaking plant in Nevada designed to produce 500 tonnes a year, with operations targeted for July 2027. The financing includes $45 million in equity led by Bain Capital Ventures and $30 million in debt and equipment financing anchored by J.P. Morgan. Solcoa already produces neodymium-praseodymium and samarium metal at about 10 tonnes a year in Alameda, but the global market is much larger. The money will help scale that 50x and then 6x. The article also notes progress further along the chain: Vulcan Elements will supply magnets for the Army’s SkyFoundry drone initiative, and a western rare earths supply chain is developing from Earth AI finding them, Solcoa metallizing, Vulcan turning them into magnets, and Westmag turning magnets into motors. Third, Terraform Makes Methanol From Sunlight and Air: Terraform’s reactor team produced batches of high-purity methanol at its Muroc desert test site. The timeline is January closing on test site, March construction underway, May 1.8 MW of solar installed, July CO2 produced on site, August hydrogen. Methanol is useful in its own right and as a starting point for gasoline, aviation fuel, and plastics. The strategy is to start with high-purity products that command higher prices and drive costs down. Fourth, Joby Flies Across America Autonomously: Joby announced that its J208 had flown 3,199 miles across the United States with zero control inputs from the onboard safety pilot, ending near Kitty Hawk. The plane was a converted Cessna Caravan, and it handled taxis, takeoffs, navigation, and landings across a route with stops in Arizona, Texas, and South Carolina, operating at busy airports and rerouting around severe weather. Fifth, Meta Makes VR Glasses (and more): Meta held Connect and announced Muse has voice and video now, Meta Glasses can serve as FDA-approved hearing aids, and the biggest announcement was Meta VR Glasses weighing just 100 grams, with 5K micro-OLED display, eye and hand controls, ability to turn a flat surface into a keyboard and touchpad, more than 100 immersive live US sporting events a year, 3D movies and hologram calling, due in spring 2027 for $1,299.99, with up to three hours of media playback. The article argues this shows Zuck’s long-term plan is on track.

The seventh major article is The Tim Ferriss Show Transcripts: How to Spot 10-Year Trends and Build Billion-Dollar Companies — Kevin Ryan of DoubleClick, MongoDB, Business Insider, and Many More (#884). This is a transcript of Tim Ferriss interviewing Kevin Ryan. Ferriss outlines Ryan’s career: 1985 Yale economics, Prudential, INSEAD, Euro Disney, United Media, DoubleClick, AlleyCorp, Gilt Groupe, Business Insider, MongoDB, Zola, Nomad Health, Transcend Therapeutics, skiing to the South Pole, Otsuka acquiring Transcend for $700 million up front plus up to $525 million contingent, and AlleyCorp’s $335 million second fund. Ryan explains that ten years is important because to build an important company it takes ten years, so you must bet on a long-term trend. He got trends right with internet advertising at DoubleClick, unstructured data at MongoDB, psychedelics, nuclear energy, the creator economy, and value-based care. He has not yet found the right product for longevity. Ryan thinks about second-order effects, giving the example that in 2003 they saw bandwidth prices going down and realized video would become viable around 2005, and they should have started YouTube. He discusses Gilt Groupe: second year revenue $175 million, fourth year $500 million, but then brands got their own websites and competitors like Farfetch arose, and the market commoditized. He went to the board and said they should sell, and they got $250 million; three years later the buyer offered to sell it back for five million. He concludes he has not touched e-commerce since. Ryan talks about starting the Dilbert website in 1995, the first hard-coded ad sold to IBM, and his realization that the internet was a no-brainer. He joined DoubleClick as around the tenth or twelfth person, became CEO, opened offices in 25 countries in three years before the first was profitable. He left large companies after Euro Disney and wanted to return to New York. On AlleyCorp, he and Dwight Merriman put in about 500K each into each company and built for six months to a year before raising venture. They did six companies total between 2005 and 2008; three were huge successes: Gilt, Business Insider, Mongo. To recruit talent, they had credibility from having sold DoubleClick, the most valuable startup in New York at the time. Henry Blodget came on board after 30 minutes. Business Insider started narrowly covering New York technology, then added verticals, never doing marketing, and eventually got 100 million uniques. Ryan says getting the eyeballs is 95% of the job. He describes his “business crush” test: if an idea sticks for two weeks, he generally does it. He discusses Valar Atomics, a nuclear company they invested in at a $20 million valuation three years ago; it recently raised from Sequoia at six billion. He says the US government was not behind nuclear at the time, but the Trump administration has moved nuclear and psychedelics forward substantially. On psychedelics, reading Michael Pollan’s How to Change Your Mind changed his mind; he became a donor to Yale Center for Psychedelic Research and then started Transcend with Ben Kelmendi and Blake Mandell to develop methylone, a softer, shorter cousin of MDMA that can be taken once a week and has less of a downer. They have a patent for PTSD, depression, and anxiety. Their trials for PTSD were 60% women, because more women have PTSD, and the number one source is sexual assault. They set up Transcend as a public benefit corporation, with initial shareholders committing to give 10% of their gains to a foundation, and they will give $20 million in the next nine months to causes linked to psychedelics. Ryan expresses concern about income inequality, the budget deficit, and the need for a more equitable tax system, professional training and retraining, and more immigration, especially for skilled AI PhDs. He says New York is booming, and the United States from a business tech point of view is crushing it, but the country needs to be more balanced and caring.

The eighth theme is the AI bubble debate and economic interpretations. Will the AI Bubble Destroy our Retirement? from Mr. Money Mustache begins by noting the stock market keeps surprising. It summarizes the AI story in three points: AI has reached a level where it can do complex thinking and reasoning, fastest worldwide adoption with over 1.5 billion people using it, and a crazy investment cycle of $1.8 trillion spent since 2002 with another trillion next year, compared to the Burj Khalifa and the US interstate highway system. The author then compares 2006 worries to 2026 worries: in 2006, Bush, Hussein, Bin Laden, Iraq, housing bubble, outsourcing to India and China, peak oil; in 2026, Trump, Putin, inflation, AI bubble, national debt, Iran and Ukraine wars. He notes that after 2006, the Great Financial Crisis happened but a $100,000 investment still ballooned into $852,000 today, and the world got better: extreme poverty dropped from 20% to 8%, infant mortality cut in half, solar became 40 times cheaper, electric cars went from zero to about a quarter of global sales. The core argument is that prosperity comes from productivity, not the financial system, and if the system crashes, we can make up new numbers and get back to work. Happiness tops out after a certain standard of living, and factors include genetics, relationships, kindness, gratitude, optimism, and outdoor activities. The article argues that fear itself is the biggest problem, because humans are danger detection machines and fear moves the goalposts. It recommends understanding your own history, identifying the fear, taking action because action cures fear, tuning out news, focusing on positive behaviors, and surrounding yourself with optimistic people. The author frames AI as a Cognitive Nail Gun, boosting efficiency and productivity, and says it is an opportunity to make choices rather than something to fear.

Good points from James Gilliland, from Marginal Revolution, quotes Gilliland saying that if we actually get AGI, the resulting boom in industrial capacity from robotics and massive society-wide wealth creation will look like a total vindication of neoliberalism. The discourse about financialization and offshoring being a generational mistake may be replaced by a different interpretation: that the late 20th and early 21st centuries were an enormous capital-accumulation phase, freeing up civilization-scale pools of liquid capital that could ultimately be deployed to accelerate the most consequential technological phase-change in human history. What looked like deindustrialization and betrayal becomes the prelude to reindustrialization on an almost unimaginable scale. None of this is inevitable, and China and x-risk remain enormous contingencies, but if it is, the future will look back on the last fifty years very differently.

Experts Lead Experts from Silicon Valley Product Group notes we are still early in learning the long-term impact of AI, but one trend capturing attention is that managers and other leaders will be much less important, and all we really need are skilled individual contributor builders. The article references headlines about companies pushing for fewer managers.

The ninth theme is Seth Godin’s essays and interview. Say what you want argues that the problem is with the “and,” which is often a “but” in disguise. People want to make the art they have in their head and want the market to love it; want the wedding of their dreams and not to worry about money and want cousins and friends to enjoy every bit; want few risks and extraordinary returns; want a difficult conversation and no tension, stress or possible downsides. We hesitate to admit what we really want because when the “and” shows up, we can see we are being unreasonable. Strategic thinking does not ignore the systems all around us or the trade-offs that scarcity and competition require; it embraces them.

Thinking about your purpose argues that we are not born with a purpose; we have the freedom to choose the impact we will make. Our work, though, does have a purpose: the change we seek to make and the people we are here to make it for. Committing to work with purpose transforms us as well as the people we engage with. Resistance pushes us to imagine our work is not for us, to deny responsibility, or to blame the system, and mostly to refuse to name the purpose of how we are spending our time. Stuck is just another word for being conflicted about our purpose; the knot holds us back because we have become entangled in goals that are mutually exclusive.

KnotChat: Jennifer Myers Chua, also from Seth Godin’s Blog, introduces Jennifer as a marketer, project manager and impresario. The post says they have worked together on projects and it is always a delight, and they sat down to talk about the hard work of what’s next.

The tenth theme is network fatigue and offline devices. I’m Tired of Being on the Network from matduggan.com is a long personal essay. The author begins by observing how tired someone is of the always-online life by watching the distance between the phone and the body. His wife launches her phone to the far corner of the room. He used to be a believer who lined up for the first iPhone at 3 a.m., but now he cannot tell which model is in his pocket and feels nothing for old phones. The smartphone has become a tax; in Denmark he is expected to carry the Aula app for school, Rejsekort for trains, MitID, and e-Boks for official messages. It is not a choice; it is compliance, and the offline option exists like a wheelchair ramp pitched at an impossible angle. Every few years he buys the identical slab, and the only honest pleasure is peeling the plastic film. He realized something was wrong when he dropped the phone and did not mind; he was a tiny bit sad the screen survived. He cannot stop seeing everything he hates: storage prompts, Apple Music pings, App Store garbage with ads, maps with ads, podcast ads, and Danish ads at maximum volume. He leaves the phone by the front door next to the garbage, but fatigue found him anyway: the Steam Deck downloading prompts, streaming ads, sports-betting commercials, and the sameness of online content. He describes TikTok’s confident opinions, Instagram’s puppies and softcore, Twitch’s desperate streamers, Netflix shows about TikTok stars, and HBO documentaries made of social-media clips. To recharge, he turned to a bare circuit board: the MiSTer, an FPGA device for retro video games. Setup was easy, and after pulling the ethernet cable, it kept working exactly as before. He added TapTo or Zaparoo, an NFC card system that makes loading a game like picking a cartridge off a shelf. The first game he loaded was Crash Bandicoot, and it put him back in the arcade. The MiSTer asks nothing of him and knows nothing about him, and that is the entire point.

Meadow lets you leave your smartphone at home without sacrificing the essentials, from TechCrunch, reports on a startup called Meadow. The founder, 25-year-old Shreyas Narlanka, says, “We’re not Luddites... it’s just we don’t want social media, browsing and that kind of stuff.” The article describes the product as letting users leave smartphones at home while retaining essentials.

The eleventh theme is the Marginal Revolution posts on books and links. My Dreadful Body from Marginal Revolution is a recommendation of Egana Djabbarova’s short novel. The post quotes Olga Grjasnowa describing it as “An extraordinary novel that translates corporality and estrangement into language,” and notes it has been a fantastic year for fiction, with a report coming in November.

Friday assorted links from Marginal Revolution lists six items. The first is a study finding a significant and highly robust empirical relationship between gender conservatism and protectionism, evident since 2008 and not a product of Trump-era politics. The second notes more than half the world’s top AI talent comes from China. The third reports Argentina now at 1.0 and Mexico at 1.2, likely fertility rates. The fourth says China is rushing to build a huge nuclear arsenal. The fifth links to Afra Wang on some Chinese AI communities in the New Yorker. The sixth links to mental health trends for Australian youth.

The twelfth theme is Nautilus science shorts. Behold, Crawzilla, the Not-Quite-Biggest Crayfish Ever Caught reports on a crayfish and what it can teach about body size. The article body is short, offering only that the crayfish is not quite the biggest ever caught and that it can teach about body size.

Why is This Octopus Hitching a Ride on a Dolphin? is described as a quirky little climate change story. The article body is short, offering only that title and subtitle.

How the Pro‑Nicotine “Wellness” Movement Rebranded an Addictive Drug reports that the movement is focused on the idea of “clean” nicotine use, such as pouches, which don’t contain tobacco.

Why We Stay Afraid reports that a newly discovered brain region in mice solves a longstanding mystery about how fear lingers after threats fade.

How Gentle Tapping on the Shin Heals the Brain asks, “The shin bone’s connected to the… brain?” and reports on the connection between gentle tapping on the shin and healing the brain.

The thirteenth theme is media, speech, culture, and history. Live Free Speech Discussion, 2:00 PM ET Today: Interview With Russian Journalist Dimitri Simes from Racket News announces a press conference with Dimitri Simes, moderated by Matt Taibbi. The article recaps Simes’ case: in August 2024 the FBI raided his Virginia home while he and his wife were away, seizing all possessions but his handgun. A month later he was indicted under the International Emergency Economic Powers Act, charged essentially with violating a sanctions regime by hosting a television program in Russian, in Russia. His attorney Michel Paradis calls it a totally unprecedented use of sanctions laws. The article notes the case pursued by both Democratic and Republican administrations has garnered little coverage. Simes will give an update with Paradis and Jason Wright on hand. The press release states Simes is a dual citizen of the U.S. and Russia, the house was raided on August 13, 2024, and they were indicted in September 2024. The accusation is that he openly works for Channel One, a Russian television channel under US sanctions. The New York Times described the case as a core test of the First Amendment, quoting Stanford Law Professor Evelyn Douek. Simes says the indictment is an outrageous perversion of press freedom and that working for a foreign broadcaster should not make him an exile.

The hottest new hangout for middle schoolers is NPR’s comment section? from TechCrunch reports that when NPR staffers flagged strange comments under their podcasts on Spotify as bots, it took a Gen Z colleague to immediately figure out the mystery.

Asteroids Named After Tom Lehrer and 'Weird Al' Yankovic from Slashdot reports that the International Astronomical Union approved the name (14331) Alyankovic for a new asteroid after “Weird Al” Yankovic, championed by planetary scientist Allison McGraw and others. The scientists also convinced the IAU to name an asteroid after Tom Lehrer, who died last year at age 97. The article details Lehrer’s science-themed songs and Yankovic’s geek-friendly career, including “It’s All About the Pentiums,” “White and Nerdy,” and his five-minute summation of Star Wars: Phantom Menace. The article notes Yankovic bootstrapped his career in 1979 recording “My Bologna” in a bathroom.

Behind the Blog: Did you notice? from 404 Media is a behind-the-scenes piece where Joseph discusses some small changes, an AI song, and internet soup. The excerpt says they are making a couple of changes that might sound boring to someone who does not read or listen to their work, but Joseph is legitimately excited about them, and the piece will discuss one now and maybe the other next week.

Reasoning together from Aeon is a very short article about Jürgen Habermas, the great defender of deliberative democracy, who lived up to its demands because he never feared changing his mind. It is by Emilie Prattico and links to an essay on Aeon about Habermas’s theory of truth and his willingness to change his mind.

Fireside Friday, September 25, 2026 from A Collection of Unmitigated Pedantry begins with a note that the Teaching Paradox series on Hearts of Iron is ongoing, but the structure means posts are chunky, so this is a fireside week. The author, a historian, discusses a series of videos from a channel called “The Grainbound” on realistic worldbuilding. He says the videos are decently well-produced and authoritative sounding, which is why it is such a problem that they are not quite right. He picks two videos: one on logistics of grain transport shaping maps and another on sea trade in food supply to cities. He agrees that food production fundamentally determines the shape of pre-modern societies, but the devil is in details. A sharp howler in the second video concerns the claim that Rome was in the wrong place for its food and spent a century moving the harbor toward it. The author points out that Rome already had Ostia Antica, and Portus expanded it, and that inland capitals with port cities are common, such as Athens and Piraeus. The video also wrongly treats the two short bursts of calorie arrival as unique to sailing season, when it is an iron law of cereal cultivation. The deeper problem in the most popular video, “The Ox That’s Breaking You Fantasy Map,” is that it begins with terrain and major settlements as different kinds of “Lords,” but urban centers emerge bottom-up: farming villages, fission, local market centers, and Big Men or Big Temples centralizing surplus. The author emphasizes that the map is not the territory and the true landscape is a tightly woven carpet of small farming villages. He gives Byzantium and Vienna as examples of cities whose importance was an intentional choice by specific human actors, not immutable law. He says the channel is pockmarked with the scars of autodidactism and the creator has a background in software development rather than history. He worries the popularity of such channels reinforces the false perception that history requires no specific training. He also notes there is a market for this that professional historians are mostly not reaching, and he is considering branching into video. Recommendations include The Morbid Zoo’s three-part series on monsters in Lord of the Rings, Forgotten Weapons videos on the USS New Jersey’s guns, a deep dive on Persephone by Spencer McDaniel, a new Pasts Imperfect essay by Joel Christensen on the Iliad and Odyssey, and a book recommendation: L. de Ligt’s Peasants, Citizens and Soldiers: Studies in the Demographic History of Roman Italy 225 BC – AD 100. He describes the book as technical but well written, arguing for a revised low count of Roman population, and notes it will bring readers up to speed on where the population question stands.

The window closes with a note that the episode has covered all fifty articles, grouped by themes ranging from AI misuse and cybersecurity to the aggregator wars, startup growth, the AI bubble, network fatigue, and science, speech, and history.

Your signing session ended