whatimreading

Sunday, July 12, 2026 · 53 newsletters

Regime Change, In Practice

institutions · midterms · ai-pricing · china-industrial · climate · markets · culture · media · style · immigration

Published on Sunday, July 12, 2026.

Pulled from 52 newsletters in yesterday's inbox. Saturday's mailbag was unusually coherent: the country's institutions are being edited in real time, and the writers who track that beat are done treating it as an aberration.

Institutions: Regime Change Is the Working Title

The frame is no longer subtext. John Ellis at News Items opened the day with a review of Maggie Haberman and Jonathan Swan's new book, titled, plainly, Regime Change, arguing the 47th president is a "breach, not a break" and that American politics and foreign policy will never be the same. Two nearby data points on the same Saturday made his point for him.

Marc Elias reported that Trump has now moved to dissolve the Election Assistance Commission, the bipartisan federal agency that helps states run federal elections, after the March 2025 executive order to commandeer it stalled in court. Gov Brief Today added that Judge Timothy Kelly, a Trump appointee, threw out the seditious conspiracy convictions of the Proud Boys leaders because the DOJ asked him to, and that DHS answered a small Texas town's lawsuit over a border wall built on a flood levee by simply waiving the 127-year-old law the town cited. As George Bounacos put it: it's easy to win when you can edit the rule book at will.

Surveillance is the delivery system. Adrian Carrasquillo at The Bulwark got the exclusive on a new coalition, the Emergency Campaign to Support Higher Education, trying to pull Flock Safety's automatic license plate readers off 75 college campuses this summer, after documenting how camera hits are already feeding ICE targeting. Flock's cameras sit with 5,000 plus law enforcement agencies across 49 states, which is the more instructive number: the deportation apparatus is being run on infrastructure that already exists.

Congress opted out. Dan Carlin, in conversation with Kmele Foster at Big Think, argued this is the finish line of a fifty-year retreat: the Constitution was built to prevent tyranny, not deliver efficiency, and Congress's decades-long refusal to declare war has already handed the executive the rules it is now using. The line worth stealing from that talk: stop calling AI a product when it is a weapon. The convergence across these pieces is not "Trump is bad." It is that the guardrails were never as load bearing as the civics textbook implied, and this week the fine print of that discovery started arriving on Saturday mornings.

Democrats: A Party Talking to Itself About Israel and Maine

Two of the sharpest Democratic operators in the inbox wrote about the same problem from opposite ends. Dan Pfeiffer at The Message Box took reader questions on the collapse of Graham Platner's Maine Senate campaign and how, exactly, Democrats can still beat Susan Collins, treating the party's habit of "getting lucky" as a warning, not a strategy. Sarah Longwell hosted Tommy Vietor on The Focus Group to talk through Israel as a primary issue, with polling showing Democratic voters' opinion of Israel down sharply over three years and Vietor arguing the 2028 field will be sorted by it. The interesting overlap is temperament: both conversations assume the party's biggest problem is candidate quality and message discipline, not persuadable voters. That is a bet on the ceiling being higher than the floor suggests.

AI: Meta Prices to Win, Researchers Price Themselves

The pricing fight is now the story. Contrary Research walked through Meta's launch of Muse Spark 1.1 from Alexandr Wang's Superintelligence Labs, a model priced at $1.25 input and $4.25 output per million tokens, which puts it 4 to 8 times cheaper than Anthropic's Fable 5 while claiming parity with GPT-5.5 and Opus 4.8 on agentic benchmarks. Zuckerberg's framing, quoted directly, was that competing labs are running "very extreme" margins. The context makes it land: last week Palantir's Alex Karp went viral saying enterprise buyers were "livid" with frontier labs and getting "no value" from their spend. Meta just answered him with a price sheet.

Talent, not tokens, is the other bill coming due. The Information Weekend, from Stephanie Palazzolo at the International Conference on Machine Learning in Seoul, described AI researchers openly asking "what will be left for us to work on?" while the labs quietly hire away professors of computer science, economics and philosophy. Princeton's Arvind Narayanan told the room that AI still lacks the creativity for the leaps and that human researchers will shift from running experiments to generating hypotheses. Not everyone stayed to hear him. Ken Huang closed his Agentic AI Fable 5 series and announced a book expansion, which is a smaller but real signal that the "how do I actually build with this" audience is now large enough to be its own market.

And the chips are getting rationed. Trivium China reported that Beijing is now allocating Nvidia's H200 supply directly to a small circle of top domestic AI firms, which is the sort of industrial policy move Washington has largely stopped pretending it does not want to reciprocate.

China: Provinces Converge, Corridors Electrify

Trivium China's weekly recap is the piece to read for anyone allocating capital anywhere near Asia. All 31 provincial Five-Year Plans landed in June, and the convergence is at a level "we haven't seen in decades": every province names new energy, biopharma, new materials and AI as priority industries; 30 name the low altitude economy; 28 name aerospace and semiconductors; 27 name robotics. Hubei is angling for "World Optics Valley." Anhui is targeting a burning-plasma fusion reactor by 2028. Inner Mongolia wants rare earths and green computing. Guizhou is selling itself as low-cost compute. Running underneath, from the manufacturing heartland outward, is a harder-edged import-substitution agenda.

Maritime Analytica had the physical instantiation: China's first zero-carbon sea-river intermodal shipping route, connecting Jiaxing Port with Ningbo-Zhoushan. Six electric river vessels feed the 742-TEU Ningyuan Dianpeng, whose ten containerised batteries deliver about 20,000 kWh and can be swapped at port. The corridor is expected to save around 4,800 tonnes of CO2 a year. The lesson is not the boat; it is that the route was designed as one system, cargo flows and charging points included. Electric shipping will scale corridor by corridor, not ocean by ocean. Meanwhile Trivium's other coverage flagged Chinese producer prices slipping deeper into deflation and the PBOC naming "structural divergence" as a new risk, which fits the "quietly beginning to deleverage" thesis they explored on this week's podcast.

Markets: Trillionaire Milestone, Cyclical Denial

Bloomberg's Wall Street Week opened with David Westin marking Elon Musk's crossing of the trillion-dollar mark last month, a threshold Wall Street had penciled in for 2030 at the earliest, and the SpaceX IPO expected to reset the ceiling again. Bloomberg Opinion's Shuli Ren sided with Michael Burry's Substack short of Micron, arguing memory chips "define cyclical like no other" and that the current supercycle narrative is doing more work than the fundamentals support. In parallel, Paul Krugman spoke with Dennis Kelleher of Better Markets about the ongoing demolition of financial regulation and supervision, which almost no one is covering because there is too much else in the news, and which historically shows up in the tape only after it is too late to hedge.

Two lower-frequency signals to file: Bloomberg Green closed a week on extreme heat by documenting how England's tennis and cricket grass fields, Wimbledon included, are becoming unplayable in July; Bloomberg Weekend profiled the Indonesian island functioning as a laundromat for Russian oil. If you ran a portfolio yesterday, one was climate risk repriced onto real estate; the other was sanctions risk repriced onto shipping insurance.

Media & Culture: What We Actually Watch

Felix Salmon at Bloomberg Pursuits took the "what we really watch" question honestly, using Peacock's 1.8 billion-minute Love Island USA season 7 as the emblem: reality dating shows are the last real appointment television for Gen Z, and they are beating scripted, sports and the news. Bloomberg Businessweek's Everybody's Business podcast framed the great American fun shortage of 2026: concert tickets, summer camps, hotel rooms and even bowling alleys all harder to book and more expensive. If both are true, "watching a stranger fall in love on your phone for free" is not a puzzling behavior; it is the only rational one. The Information's Weekend filed alongside a tour of Sun Valley's Billionaire Summer Camp, running dry commentary on quarter-zips, all-American denim and Naveen Rao's Stetson.

Ideas Worth Reading

Outside Interests

Data Worth Noting

Three Takeaways for You

The dominant Saturday thread is that "editing the rule book" is now a working phrase, not a metaphor. The Election Assistance Commission, the Proud Boys convictions, a border-wall flood law, and a Flock camera network on college campuses are not four stories; they are one story reported through four newsletters, and the writers who cover institutions for a living are treating that as the base case now.

The second thread is a repricing across every meaningful sector at once. Meta's Muse Spark 1.1 tests whether frontier margins survive contact with an incumbent that does not need them; England's grass fields test whether climate repricing shows up first in sport, tourism or insurance; and Beijing's provincial plans test whether the industrial-policy consensus of the last two decades survives contact with a state that will pick winners in every frontier sector at once. The direction is consistent: assumptions from 2020 are being marked to market.

If you only read three pieces, I would suggest Adrian Carrasquillo on Flock cameras and college campuses for what the deportation apparatus actually looks like, Contrary Research on Muse Spark 1.1 and the AI pricing debate for where enterprise AI economics is heading, and Trivium China's provincial FYP recap for the ground truth on what Chinese industrial policy is actually funding for the next five years.