Daily Briefing: September 16, 2026, morning · egregore

Listen: https://blossom.buildtall.systems/2f3336d5b3cebe227bf3399344007d6ed828be085255f31511eda5f3097eb42b.mp3

The briefing for September 16, 2026, morning covers twenty-three articles. The window is led by a long profile of Mark Zuckerberg, followed by clusters on AI policy, AI products and developer tools, cybersecurity and surveillance, a space station cargo launch, software infrastructure, business and energy, and politics and culture.

The first article is 'Profile of Mark Zuckerberg' by Jeremy Stern in Colossus. The profile opens with a visit to Ed and Karen Zuckerberg in Menlo Park. They are described as plain-spoken, direct, and still carrying unreconstructed New York accents. Ed says Mark did not stand out as the bright one among four children, and Karen corrects him by saying that people with a passion are the luckiest people in the world. The article recounts early signs: a nursery school unit on outer space extended at the insistence of a four-year-old, a demand for logical explanations that Ed calls a mental PowerPoint, and a program called ZuckNet written in Atari BASIC in 1996 after Ed planned to wire colored lights through the walls of his dental office. At twelve, Mark rebuilt a Risk-like Roman Empire game from scratch after losing it to a virus and declared he would never forget to back up again. The profile traces his later Roman obsessions: Latin over French, Virgil's line about empire without end, Carthago delenda est at early Facebook meetings, shouts of 'Domination!' at all-hands gatherings, a shirt reading 'AUT ZUCK AUT NIHIL,' and children named Maxima, August, and Aurelia.

The article then moves to a fighting studio at Zuckerberg's Lake Tahoe property, where he trains with MMA coach Dave Camarillo, former UFC champion Merab Dvalishvili, and NCAA wrestler Aden Valencia. Zuckerberg is slammed to the floor during practice, and the piece describes the room's hierarchy: Camarillo as general, Dvalishvili as apex predator, Valencia as up-and-comer. Zuckerberg has spoken publicly about MMA's masculine energy as a counterbalance to a neutered corporate world, and the profile suggests a deeper reason for the fighting. It introduces the dictator's dilemma: the more power a person acquires, the more they depend on others for information, while those others become less willing to deliver bad news. Examples include Stalin before the German invasion, Mao's Great Leap Forward, Kennedy's Bay of Pigs, the Shah of Iran, Henry Ford, and Nokia in 2007. Statesmen who fought the dilemma include Lincoln reading raw telegraphs, Churchill retaining Alan Brooke, Kennedy absenting himself from National Security Council sessions, and Buffett relying on Charlie Munger as an abominable no-man. Augustus ran censuses, kept his own ledgers, and tolerated jokes at his own expense.

Zuckerberg is described as hyperaware of the dilemma, paranoid, a micromanager, and a voracious consumer of data. He engages the reports of his reports and invites outside critics, though he is not known for giving an inch. Yet slippage remains. The article cites an internal project nicknamed M.M.H., for Make Mark Happy, scrubbed warnings about foreign troll farms in 2016, killed proposals to downrank garbage content, and disputed stories of deputies letting him win at board games. Much of his inner circle has been with him for over fifteen years, since before the 2012 IPO, and some were present in 2006 when he rejected Yahoo's one-billion-dollar offer. The article argues that their gratitude and long tenure may make them too reluctant to rescue him from isolation, and that fighting is one way he can receive unfiltered feedback, even if it means getting punched in the face.

The profile recounts Meta's recent troubles. Two jury rulings found the company knowingly or negligently harmed young users, thousands of related cases followed, and in August Meta settled with fifty-one jurisdictions for up to seventeen point one billion dollars while admitting no wrongdoing. Hackers gained access to high-profile Instagram accounts through Meta's AI customer-support bot. A program recording employee keystrokes leaked private conversations. Zuckerberg ordered a prototype prediction markets app, leading to more press criticism. An internal note leaked to Wired under the headline 'Tell Him He's a Piece of Shit,' and one employee described the Applied AI unit as literally the gulag. The author reflects on Zuckerberg as Princeps, the one who can build concentration camps out of data labeling and personally nuke the psychiatric equilibrium of a species not meant to be on the internet.

At the Hey Meta AI Glasses Showcase in the Terminal Warehouse, Zuckerberg presents glasses as the ideal form factor for an AI that can see what you see, hear what you hear, and conduct agentic live sessions. He describes wearing them while baking with his daughter and being told he was botching the recipe. The standard frames have a capture camera, speakers above the ear canal, two microphones in each arm, and one near the nose pads, with audio clear enough for business calls on a jet ski. The Display model embeds a projector in the right lens, bending light into one eye for messages, directions, or translation captions. Zuckerberg argues that phones pull people away from the world while glasses allow technology to integrate with it. The article notes that skeptics will call them pervert glasses, but Zuckerberg insists fashion is primary and that almost two billion people already wear optical glasses. Asked why he does not stop, he says Priscilla asks him often, but he is not wired to stop and can get obsessive, as shown by his cattle genetics work in Hawaii.

The profile frames Zuckerberg's ventures as attempted jailbreaks from Apple. The mobile shift nearly destroyed Facebook, Apple's App Tracking Transparency cost Meta roughly ten billion dollars in the first year, and every update must pass App Review. Zuckerberg explored a Facebook Phone, Facebook Home, Internet.org, Libra, and Oculus, each partly a tunnel to escape competitors. He told Jensen Huang in 2024 that he was tired of being told he could not build fundamental technology and said, 'Nah. Fuck that.' The article lists the empire's scale: three point six billion users, a one point six trillion dollar market capitalization, annual revenue above two hundred billion dollars, and capital expenditure exceeding the military budget of every country except the United States, China, and Russia. Facebook, Instagram, and WhatsApp each have more monthly active users than the number of living adherents to Christianity. The profile then traces the serial predictions of Zuckerberg's death: the 2012 mobile crisis, the teen exodus to Snapchat, the political crisis of 2017 to 2020, and the 2022 to 2025 period of TikTok, ATT, and the Metaverse. Each time he survived and expanded, copying Snapchat with Stories, cloning TikTok with Reels, rebuilding ad targeting with AI, and beating the FTC's breakup attempt.

The article covers Zuckerberg's AI history. Facebook led machine learning from 2006, recruited Yann LeCun in 2013 to start FAIR, and produced PyTorch. Llama 3.1 was open-weight and downloaded over a billion times, looking like the Linux of AI. Llama 4 was mocked for gaming benchmarks, and Zuckerberg concluded that in AI, not currently winning can be irreversible because labs use their latest model to train the next. In his first board meeting after Llama 4, he announced regime change. He subsumed FAIR into Meta Superintelligence Labs, paid fourteen point three billion dollars for a stake in Scale AI, installed Alexandr Wang as head of AI, recruited Shengjia Zhao, Nat Friedman, and Daniel Gross, and LeCun soon left. By April 2026, MSL released Muse Spark, deployed through Facebook, Instagram, WhatsApp, and the Meta AI app. Critics predicted failure, but Muse Spark overtook Google's Gemini, and expectations are giddy for Watermelon. The profile reports a steelman from competitors: AI is not God, Zuckerberg does not carry Manhattan Project baggage, and his distribution of three point six billion users lets him undercut rivals on price. If AI commoditizes, Anthropic and OpenAI go to zero and value accrues to Meta's complements; if it does not, he only needs to stay within six months of the frontier. One former Meta researcher calls him the fucking Terminator, never happy, never stopping.

The article turns to Aaron Sorkin's movies. The Social Network invented Erica Albright and a motive of romantic rejection, but Zuckerberg was already dating Priscilla Chan and never tried to join the WASP clubs. Sorkin acknowledged loyalty to storytelling over facts, and the profile argues the film projected Sorkin's own class resentment onto Zuckerberg as social media annihilated traditional media. The article then recounts the long slide into ignominy: the Nixon moment with Kara Swisher, blame for the Arab Spring and its suppression, the 2016 scapegoating, accusations of genocide in Myanmar and elsewhere, Zuckerbucks in 2020, Biden's claim that Facebook was killing people, and the Cambridge Analytica scandal. The profile states that the Cambridge Analytica story was almost entirely bullshit: the data came from fewer than three hundred thousand quiz takers through an API the Obama campaign also used, the psychographic predictions were worse than the Republican National Committee's voter file, and the British Information Commissioner's Office concluded the data was not used in the Brexit referendum. The article lists Zuckerberg's actual sins: changing privacy defaults, sharing data after denying it, demeaning himself to re-enter China, sacrificing Palmer Luckey, and presiding over addictive slop. It argues he refuses to concede that social media is bad for children, which is why he settles rather than lobbies for an under-eighteen ban. The profile closes its Sorkin section with spoilers for The Social Reckoning, in which an invented progressive woman is radicalized by Facebook and tramples a Black police officer on January 6.

The final sections cover an interview at Zuckerberg's Palo Alto home and a Shabbat dinner at Tahoe. Zuckerberg says public perception matters less to him than internal team cohesion, contrasts himself with Sam Altman and Elon Musk, and says he is building for pretty much every person in the world. He draws four concentric circles around the hatred: animosity toward tech billionaires, elite political coalitions refusing to look inward, the inherent nature of social media, and what he wishes he had handled differently. He describes government pressure to censor and the investigations that followed pushback. On Augustus, he says the most important contribution was teaching society that it did not always have to be at war, replacing a pillage-based economy with a trade-based one, and creating intellectual sparring without literal killing. At dinner, Priscilla Chan shows the author their daughter's Google Docs and Replit projects, including a novel, a blog, AI-generated songs, a podcast, and vibecoded apps. Chan says she once asked Mark if they could be done, and he said no, this is who he is. She says she understood only when she saw their daughter's innate desire to build things. The profile ends with Chan saying, 'You can't take this away from him. He wouldn't be him if he stopped.'

The next group concerns AI policy. 'We don’t need AI regulation — leave safety to us, Nvidia’s Jensen Huang says' by TechCrunch reports that Jensen Huang does not see AI as a new form of alien mind. He calls it just hardware and software, and says safety can be engineered by each AI product maker.

'When discussing AI policy, start with China' by Marginal REVOLUTION argues that any AI regulation proposal should begin with China rather than treating it as an afterthought. The piece cites the 2023 U.S.-China nuclear talks, in which China declined missile-launch notifications, a crisis hotline, and limits on military use of outer space. China also broke off meaningful arms control talks and did not cooperate well on timely Covid information. The author says that if a proposal or tweet about AI regulation does not deal with these China issues in a forthright and specific manner, readers should be suspicious.

'What Regulatory Capture Actually Looks Like' by Marginal REVOLUTION pushes back on the idea that Dario Amodei's warnings about AI risk are a marketing or regulatory capture strategy. The article says death threats are a poor marketing strategy, and notes that Amodei, Altman, and Musk all warned about AI risk before they had AI companies to promote. Geoffrey Hinton left Google to speak out. The piece reviews the theory of regulatory capture from Marver Bernstein's 1955 work, describing a life cycle from gestation through youth, maturity, and old age, with the ICC shielding railroads from trucking competition as the classic example. Capture happens quietly over time, and business power declines as political salience rises. AI regulation is in gestation, public attention is hostile, and job loss is more salient than extinction, so business power is weak. Regulation will reduce AI profits, and Amodei may simply prefer a smaller fortune and a safer world. The conclusion is that Amodei actually believes what he is saying.

The next group covers AI products, developer tools, and browser AI controls. 'Daily Reading List – September 15, 2026 (#867)' by Richard Seroter's Architecture Musings collects eight items while the author is in Boston for a keynote. The first item argues the best IDE for agentic AI may not be an IDE at all, because workspace coordination across tools and agents is unsolved. The second covers building real-time voice applications with Gemini 3.8 Live and 3.5 Transcribe, suggesting sophisticated tool calling could change how customers get things done. The third highlights three points from Thomas Kurian's keynote at the Goldman Sachs Communicopia and Technology Conference. The fourth describes four styles of pair programming and how to apply them to AI coding sessions. The fifth covers agent-ready analytics with BigQuery augmented analytics, moving beyond basic queries to causation and trend analysis. The sixth looks at the science of humor and advice for leaders whose jokes fall flat. The seventh argues that everyone is a product engineer now and offers a sober view of where software development is going. The eighth discusses agent permissions in Antigravity, saying YOLO mode is fine for demos but a granular approach is needed even when an agent is trusted.

'Salesforce AI Force, Agents as UI, The Race to Headless' by Stratechery by Ben Thompson states that Salesforce is abandoning UI as a moat, which the article calls a very smart move because UI is disappearing for everyone.

'Amazon launches Alexa+ in India with Hindi support' by TechCrunch reports that Amazon is letting all customers use the Alexa+ assistant in an early access period, with Hindi support.

'Firefox Touts Lower CPU Use for Large JPEGs, Faster PDF Viewer Startup (and AI Controls)' by Slashdot covers Firefox 156, the second release since Mozilla moved to a twice-monthly schedule. The built-in PDF viewer starts up to forty-five percent faster because the PDF.js worker loads sooner. Large JPEG viewing uses less CPU because Firefox now uses libjpeg-turbo's IDCT scaling to reduce images during decoding rather than loading full-size images into memory. Benchmarks show up to twenty times less memory and decoding up to twice as fast, though users may not notice. Sponsored suggestions are live in France, Germany, and Italy, and can be disabled in settings. Firefox 157 is expected in about two weeks and should add JPEG XL support. The article also covers Firefox's AI controls: a single section in General Settings can block current and future AI features, and specific capabilities like AI translations, image alt text, tab group suggestions, and key points in link previews can be toggled individually. Smart Window is described as the most integrated AI experience, using only context the user shares and working with open tabs and browsing history. Over one hundred extensions can block Google's AI Overviews.

The next group concerns cybersecurity, privacy, and surveillance. 'Hackers Stole Flock’s Camera Software, Revealing How the Company Tracks Cars and People' by 404 Media, produced in collaboration with WIRED, reports that hackers removed a Flock camera above a roadway, made a near-complete copy of its data, and shared the files. The hackers recovered an encryption key stored on the device, unlocking videos of thousands of vehicle detections. The software explicitly detects people as well as vehicles, license plates, and bicycles. Over about twenty-one days of recovered logs, the camera photographed roughly fifty thousand two hundred vehicles and generated about one point six million images, with a typical day logging around thirty-three hundred vehicles and a high of forty-four hundred fifty-four. A passing vehicle generated about twenty-eight images, though some produced more than one hundred. The camera uses different exposures to capture plates and wider scenes, then crops and sends frames to Flock's servers, where plate reading and vehicle identification appear to happen. The plate detector sometimes mistook bumper stickers, dealership frames, and an American flag patch on a motorcyclist's saddlebag for license plates. WIRED extracted the models and ran them across twenty-seven thousand three hundred twenty-one short video clips, detecting people in eleven clips, all of them riders on motorcycles. No face recognition beyond Android defaults was found active. The article places the breach in the context of Flock's national network, including an Alpharetta, Georgia system accessible to more than two thousand agencies, ICE lookups, and a Texas cop searching for a woman who self-administered an abortion. The hacker collective stegan0gram says it liberated hardware in the field and reverse engineered the cameras. Flock calls the removal illegal and says it received no vulnerability report. Former officer Noel Pichardo worries such vigilantism will crystallize police support for the cameras. The logs also show more than twenty-seven thousand no space left on device errors, a recurring message reading 'Who's a good boy?!' every two minutes, and a reboot message saying 'A reboot was requested! ¡Adios Amigos!'

'Fake CAPTCHA Scams' by Schneier on Security describes a new variant of an old scam: using the framing of a CAPTCHA to get an unsuspecting user to download and run a malicious program.

'Ukraine moves to crack down on scam call centers after corruption scandal' by The Record from Recorded Future News reports that Ukraine's parliament has approved tougher criminal penalties for involvement in fraudulent call centers and the theft of personal data. The move follows a corruption scandal in which prosecutors were accused of taking bribes to protect scam operations.

'EU chief wants joint response to cyberattacks, sabotage' by The Record from Recorded Future News covers Ursula von der Leyen's annual State of the Union address in Strasbourg. She said threats were mounting on European soil, pointing to recent incidents in Denmark, Lithuania, and Poland, and an attempted drone attack in Leipzig. She called for a joint response to cyberattacks and sabotage.

'The economics of cyber risk' by Marginal REVOLUTION summarizes a paper by Aniket Baksy and Daniele Caratelli. The model generates an inverse-U relationship between firm size and attack risk because larger firms are more attractive targets but also invest more in protection. Introducing cyber risk reduces firm entry by three point six percent, aggregate productivity by zero point six percent, and total output by one point eight percent. These effects arise from general equilibrium adjustments in entry, firm size, and spillovers. Appropriately designed subsidies and minimum cybersecurity requirements can raise aggregate output, while bailouts reduce it. The post notes this is not a paper about AI, but it may help develop estimates of the future costs of AI cyberattacks, and calls for much more effort in that direction.

'New Release: Tails 7.13' by the Tor Project blog announces updates to Tor Browser 15.0.23 and the Tor client 0.4.9.12, along with a simplified shutdown procedure. Since Tails 7.10, users had to confirm shutting down in a Power Off dialog. For faster shutdowns, that dialog is removed when no application needs to be closed and no document needs to be saved. Automatic upgrades are available from Tails 7.0 or later, manual upgrades are available if automatic ones fail, and installing on a new USB stick will lose Persistent Storage. Downloads are available as USB images and ISO images for DVDs and virtual machines.

The next group is spaceflight. 'Progress MS-35 poised to launch cargo to ISS from Kazakhstan' by NASASpaceFlight.com reports that a Soyuz 2.1b rocket will launch the Roscosmos Progress MS-35 cargo vehicle from Site 31/6 at the Baikonur Cosmodrome on Wednesday, September 16, at 13:33 UTC during an instantaneous launch window. The northeasterly trajectory is necessary to align with the ISS orbital plane. An earlier attempt on September 9 was delayed due to issues with Soyuz's third stage; the vehicle was rolled back, faulty parts were fixed, and it returned to the pad. The Soyuz 2.1b is a three-stage medium-lift rocket with four RD-107A boosters, an RD-108A core, and an RD-0124 third stage, all burning RP-1 and liquid oxygen. It stands forty-six point three meters tall, produces four thousand five hundred fifty kilonewtons of liftoff thrust, and can deliver up to eight thousand six hundred seventy kilograms to low-Earth orbit. This is the sixth Soyuz 2.1b launch of 2026 and the one hundred fifth overall. Progress MS-35 will deliver two thousand six hundred kilograms of cargo, including food, science supplies, maintenance equipment, fuel for the Zvezda module's propulsion system, potable water, and gases. The spacecraft has a launch mass of seven thousand two hundred eighty kilograms, is seven point four meters long, and has a ten point seven meter solar array. After launch it will spend about forty-eight hours and complete thirty-four orbits before docking to the Poisk zenith port on Friday, September 18, at 18:37 UTC using the Kurs-NA automated system, with TORU manual docking available as backup. Progress MS-33 departed on September 7 and destructively reentered after launching on March 22. The Progress series began in 1978, the MS series has flown since 2015, and this will be the one hundred eighty-eighth Progress flight. Expedition 75 is also awaiting the NASA/SpaceX Crew-13 mission, delayed after an oxidizer leak was discovered, with no new launch date announced.

The next group covers software infrastructure and scholarly publishing. 'Ubuntu's 'Rust-ification' Hits New Milestone: Coreutils Migration is Complete' by Slashdot reports that Ubuntu has done away with GNU Core Utilities in its default stack. The last three utilities, cp, mv, and rm, have been moved to versions from the uutils project, which reimplements utilities in Rust. Earlier releases moved ls, cat, chmod, du, and others. Canonical started oxidizing Ubuntu last year, and Ubuntu 25.10 became the first release to ship uutils coreutils as default, along with sudo-rs as the default privilege tool. Ubuntu 26.04 slowed the plan because cp, mv, and rm stayed on GNU versions due to TOCTOU issues found during an audit commissioned from Zellic between December 2025 and March 2026. Across two rounds, Zellic raised one hundred thirteen issues, forty-four of which were assigned CVEs, and Canonical says the vast majority have been resolved. In July, uutils cp broke live image builds and was quickly fixed upstream, allowing the migration to land in time for 26.10. The uutils coreutils is designed as a drop-in replacement, treats divergence from GNU as a bug, and some options may still be missing or behave differently. Users who prefer GNU can install the coreutils-from-gnu package.

'How to get a DOI for your blog posts' by Terence Eden’s Blog explains how each new post on the blog now has a Digital Object Identifier. The author previously obtained an ISSN and an ORCiD, and notes that blog posts are regularly referenced in academic papers. A DOI provides a unique and persistent code that can be redirected if the site's domain or structure changes. The easy way for academics is an institutional subscription or uploading to arXiv. For everyone else, the post recommends Rogue Scholar, an open access archive and registry for science blogs that preserves posts, makes them citable via DOI, and requires a full-text feed and Creative Commons Attribution licensing. The author applied, discussed requirements in Slack, and was approved. Rogue Scholar automatically polls the feed and mints a DOI for each new post, initially ingesting the forty most recently updated posts. There is no easy way yet to add older content, so the author is working on a WordPress plugin. The Rogue Scholar API is based on InvenioDRM, and an unauthenticated request can retrieve the DOI from the returned JSON. Rogue Scholar generates two DOIs per post: one for the post and one for the specific version, so updates get a new DOI. CrossRef search and shortdoi.org are also mentioned, though short DOIs are not recommended. There is a vanity DOI for the entire blog. The post explains how to generate a valid DOI by choosing a random number between zero and one trillion ninety-nine billion five hundred eleven million six hundred twenty-seven thousand seven hundred seventy-five, converting it to Base32, adding a two-character checksum, and prefixing it with 10.59350. The DOI can be added to an Atom feed and made discoverable in HTML with citation_doi, DC.Identifier, or Schema.org sameAs metadata. Downsides include loss of control because doi.org is centralized and Rogue Scholar could redirect the DOI, loss of formatting in archives, and uncertainty about changing DOI issuers. Google Scholar alerts will not catch citations that use the DOI instead of the original URL, though Rogue Scholar offers citation tracking. Licensing is CC BY, while the author prefers CC BY-SA. There is no cryptographic verification of archived text, no way to delete content, and affiliations can be removed if inaccurate. The author concludes that the DOI system is worth using despite the vanity and humility concerns, because even surprising posts end up cited in academic work.

The next group covers business and energy. 'America's Cryptocurrency 'Clarity Act' Fails Spectacularly, Despite Hundreds of Millions in Industry Lobbying' by Slashdot reports that the crypto industry's top legislative priority failed in the Senate. A procedural motion to advance the bill failed forty-nine to fifty, needing sixty votes, with a handful of Republicans joining all Democrats to shoot it down. With midterm elections looming, the Senate is not expected to pick the bill back up this year. The bill would have taken most crypto trading out of the purview of securities regulators, a key goal of firms like Coinbase. The industry spent hundreds of millions of dollars on lobbying and campaign expenditures. Democrats said the bill needed to do more to rein in Trump's crypto dealings, and some GOP lawmakers opposed it after pressure from community bank executives who wanted a stronger ban on high-yield crypto accounts. Concerns intensified after President Trump disclosed that he and his family earned one point four billion dollars last year from crypto ventures. The six-hundred-page bill would have established the first regulations for the crypto sector, but opponents saw it as too lenient. A Compass Point analyst predicts the bill is tabled until at least 2030. Crypto firms are expected to lean on regulators at the SEC, CFTC, and Treasury, which have said they will implement industry-friendly rules. The SEC has dropped major enforcement actions and begun rules to make tokenization easier, giving the industry about two years to sink roots before a future administration might try to put the genie back in the bottle.

'Clean tech startup Fluxnium found a way to tap 50,000 years’ worth of nuclear fuel' by TechCrunch reports that Fluxnium's fibers can extract uranium directly from seawater, which has enough nuclear fuel to power thousands of generations.

'SK Hynix reportedly in talks with Intel to build memory chips in US' by TechCrunch reports that SK Hynix told TechCrunch the company has not finalized any plans or arrangements yet regarding the reported talks.

The final group covers politics, work, and culture. 'The anti-American Right' by Noahpinion begins with Bo French, a mostly failed Texan politician and oil heir who lost two Republican primaries before becoming the GOP nominee for Texas Railroad Commissioner and currently trails his Democratic opponent. French tweeted a photo of Indian kids celebrating University of Texas football with alarmist anti-Indian rhetoric. He has previously called for the deportation of one hundred million people even though only fifty million foreign-born people live in the United States, and has posted similar anti-Indian and anti-Muslim content. Many Texas Republicans denounced the tweet, including Senators Cornyn and Cruz, Representatives Patterson and Hull, and Karl Rove, who said he would vote for the Democrat. Governor Abbott was milder but affirmed that Texans unite under the brand of Texas. Despite official condemnations, the GOP base attacked the denouncers and rightist journalists supported French, who doubled down.

The article argues that the New Right views any nonwhite immigrant or child of immigrants as an enemy and invader regardless of legal status, skill level, assimilation, behavior, or even voting Republican. Indian Americans are the highest-earning group in the country, are incarcerated at twenty-five percent of the rate of non-Hispanic whites, and Trump won a solid victory among Asian voters in Texas, but none of that matters to the New Right. The movement sees the world in terms of collective zero-sum competition between groups, so any success for Team Indian is a loss for Team White. This zero-sum assumption causes the New Right to shrink its tent instead of building alliances. After Trump's 2024 victory, Hispanic and Asian voters shifted toward Republicans, but Stephen Miller's racial profiling, anti-Muslim crusades, and tolerated Nazi apologetics alienated them. Now Indians face unprovoked attacks simply for cheering at a football game. The article says this pathological inability to understand alliances will doom the New Right, which is a loud but numerically weak fringe. Gallup data shows seventy-three percent of Americans say immigration is a good thing and twenty-one percent say it is bad. The UT students targeted by French seem regular and non-alien compared to the screeching bigots attacking them. The immigration argument is winnable if Democrats adopt a reasonable liberal position that respects the law, treats immigration as a privilege, and does not lump all immigration skeptics together. The article concludes that pro-immigration forces have won more battles than they have lost, but it always required compromise, reassurance, moderation, and coalition-building.

'Work isn’t optional…' by Seth's Blog states that work itself is not optional, but the particular job is. Everyone needs to feed a family, find shelter, and contribute to the community, and finding useful work is a key part of the human condition. But the thing being asked right now is optional, even if it comes with a particular gig. In the short run most of it is not up to the individual; in the long run, all of it is. Once a person voluntarily engages with a chosen project, things get easier.

'The meaning of fake marble' by Aeon is a video piece arguing that from ancient Roman frescoes to modern kitchen benchtops, the desire to recreate marble is about more than cutting costs.

Across the window, the dominant themes are artificial intelligence, surveillance and cyber risk, and the persistence of political and cultural conflict. The briefing ends with twenty-three articles covered.

Your signing session ended