Tocqueville’s paradox: it is in the most equal societies that people find inequalities most intolerable.

Lee Kuan Yew’s formula for finding leaders was brutally practical: you either have “it” or you don’t. Corporate leadership can be trained; political leadership mostly can’t. He hunted for what he called helicopter quality, analytical brainpower, a firm grip on reality, imagination, dynamism, then ranked character and genuine motivation to serve the public above all of it. A very smart person with weak character does more damage than a mediocre one. The final filter was the ability to persuade and move people, not just issue orders.

You can list the global communist leaders by watch who’s happy with Lula “election”

Congratulations to my friend @LulaOficial on his electoral victory. Democracy and multilateralism have won in Brazil. Europe’s partnership with Brazil is thriving. We will remain by your side for the years to come. Together we will keep working closely for free and open trade. A rules-based global order. And strong ambitions for our climate.

Where have I seen this before? 🤔

Scientists in Japan just tricked cancer into building its own chemo. And the raw material is carbon monoxide, the same gas that kills people in closed garages. They tested it on four kinds of human cancer cells, and the tumors cooked up the drug themselves. Healthy cells given the exact same treatment barely made any. Here's why it works. Cancer cells run hot on an enzyme called heme oxygenase, and one of its byproducts is carbon monoxide. Your body makes a little CO on purpose and uses it as a chemical signal. Tumors make way more than normal cells do, so every cancer cell is basically sitting on a stockpile of an ingredient healthy cells don't have. Chemists have known for decades that a palladium catalyst can grab a CO molecule and stitch it into a ring called phenanthridinone, which happens to be the backbone of several cancer drugs. The problem is your cells are full of glutathione, an antioxidant loaded with sulfur, and sulfur sticks to metal catalysts and shuts them down almost on contact. Drop plain palladium into a cell and it's dead before it does anything. Katsunori Tanaka's team at RIKEN got around that by hiding the palladium inside albumin, the most common protein in your blood. Albumin has a greasy little pocket it normally uses to carry drugs around, and with the catalyst tucked in there it kept working even with 20 times as much glutathione floating around it. Then they ran it on human liver, breast, lung and cervical cancer cells for 48 hours. The cancer cells filled up with the product while the normal cells stayed close to clean. A second version made a much more potent anticancer molecule, and at 10 micromolar it almost completely stopped breast cancer cells from growing colonies while barely hurting the healthy ones. The first question I had was the bloodstream. There's carbon monoxide in your blood too, so why doesn't the drug just form on the way there and wreck everything? It comes down to hemoglobin. CO latches onto hemoglobin more than 200 times harder than oxygen does, which is exactly why breathing it suffocates you. Any CO in your blood is already locked up, so the catalyst has nothing to grab. They tested blood and found no drug. So the exact thing that makes carbon monoxide deadly is what keeps this treatment safe on the way to the tumor. To be clear, this is still cells in a dish. It came out in Nature Communications last month and mice are the next step. Tanaka's lab has been building toward this for years though, including a 2022 study where they slowed tumor growth in mice by assembling a drug inside the body from injected ingredients, plus other versions that run on acrolein, a different chemical tumors overproduce. If this one holds up in animals, every weird chemical a tumor pumps out starts looking like a possible ingredient for the drug that kills it.

Reading an hour a day for a year can completely rewire your brain. What people call talent is usually just effort compounded over time.

I think people are slowly discovering that buying $2B worth of GPUs and actually turning $2B worth of GPUs into revenue are two very different businesses. You can have warehouses full of Blackwells, half finished data centers, transformers arriving sometime before the heat death of the universe, cooling systems that aren't ready, networking that doesn't work, and lenders suddenly discovering what credit risk is. Congratulations, you've built the world's most expensive collection of paperweights. Meanwhile the guy across the street with functioning infrastructure is renting out older GPUs at rising prices. Apparently this is what a compute glut looks like now. Everyone saw the shortage and rushed to build. That rush overloaded the same electrical supply chains, construction capacity, engineering talent and financing markets needed to solve it. Now the bottlenecks themselves are preventing new supply from arriving. The beautiful irony is that the more projects get delayed or stranded, the more valuable the capacity that actually works becomes. Scarcity doesn't just raise prices. It starts selecting winners and losers. The real moat is coordinating several moving parts into something that actually produces billable compute before the hardware becomes yesterday's technology. That's the refining analogy we've been hammering on. Owning crude doesn't make you a refinery. Owning silicon doesn't make you an AI factory. The value is in converting a bunch of constrained inputs into something customers can actually use. And every month you screw that up, your competitor collects the scarcity rent while your debt collects interest. Lovely business. Now here's where 2027 potentially gets nasty. Some of these guys won't die because AI demand disappeared. They'll die because they were right about demand and catastrophically wrong about the time and capital required to supply it. Stranded GPUs, unfinished shells, expensive debt, angry customers. And who comes shopping? The operators with functioning infrastructure, liquidity and the ability to commission capacity. Suddenly those failed projects become someone's discounted expansion pipeline. That's where $IREN and $NBIS could become interesting . they might graduate from fighting for market share to consolidating the industry. Maybe through acquisitions, partnerships, distressed hardware deployments, customer funded infrastructure, whatever. But that's conditional on execution. $IREN isn't immune to the same physics just because we like the ticker. My bet is that 2027 is about the brutal sorting of industrial competence. The industry may end up simultaneously oversupplied with purchased silicon and undersupplied with working compute. Some shareholders get vaporized, some lenders discover religion, and the survivors inherit infrastructure for pennies on the dollar.

China's industrial strategy of paying every company to make the same products and compete each other's profits to zero is running into a few issues

Yes

it’s so fucking over

BREAKING: 🇺🇸 Elon Musk warns 200+ AI companies could face massive failures as debt markets shut off funding. SpaceX secured $40B in debt for GPUs, but Musk warns the AI infrastructure boom could turn into a bloodbath and only a handful will survive.

In 2025, Buffett was asked about how he reads financial statements. Here is what he had to say: "I actually spend more time looking at balance sheets than income statements. Wall Street often focuses on income statements, but I like to study balance sheets, ideally over an 8 to 10 year period, before I even glance at the income statement."

« La plupart des hommes meurent de leurs remèdes, et non de leurs maladies» Molière

A powerful video aimed at raising awareness 🔥 Writer: Sol

Everything is out of control now. 😂😂