Oliver Scanlan

New paper: how should we think about "staying" in relation to climate change? We focus on four promising areas of work. Fantastic to work with brilliant people on this @BasundharaT and @PembertonSimon sciencedirect.com/science/articl…

Fiona Hill interview in Die Zeit. “We are in the midst of this catastrophe. When I said in 2022 that we were already experiencing World War III, I didn't mean a war with fixed blocs, like we experienced in World War II. Today, the world is interconnected through multinational corporations, global interests, and security guarantees. We are a world at war. This changes everything: security, the economy, alliances, technology, climate. All these problems have been swept under the rug for a long time. Now they are out in the open, and we can begin to fix them. The public must take responsibility.”

In 2002, the US conducted Millennium Challenge, the largest war games in its history. They split soldiers into two teams: 🔵Blue, which was America, and 🔴Red, an unnamed generic Middle Eastern country. The 🔴Red team was led by retired Marine Corps Lieutenant General Paul K. Van Riper. The idea behind the games was to see how, not if, the 🔵Blue team would win. In other words, the 🔴Red team was supposed to lose, but General Van Riper didn’t want to lose. So he played to win. He used asymmetric tactics by doing things like using civilian boats instead of military ones, motorcycle couriers and coded messages in mosque towers because their cell phone networks had been hacked. He launched a massive preemptive strike using a swarm of small boats and cruise missiles, which overwhelmed the 🔵Blue Team’s Aegis defense system. In the simulation, this resulted in the “sinking” of 16 American warships, including an aircraft carrier. The exercise was supposed to take 14 days. Vin Riper and his team won after day one. Understandably, the US military was embarrassed because this was supposed to show off all its superior tactics and cutting edge technologies. So, they restarted the exercise and changed the rules to force everyone to follow a script so that the 🔴red team could not win. The exercise controllers brought the sunken ships back to life, and forced Van Riper to follow a scripted path that ensured a 🔵Blue Team victory. 🔴Red was ordered to turn off certain air defense systems and use regular cellular communications to allow 🔵Blue to destroy them. 🔴Red was also told exactly where to move certain units so 🔵Blue could pretend to find them and neutralise them according to a pre-planned timeline. Most crucially, Van Riper was forbidden from using the swarming tactics that had been so effective in the opening hours. The controllers argued that the reset was necessary because the goal of Millennium Challenge 2002 wasn’t just to see who would win, but to test new Network-Centric Warfare concepts. They felt that if the game ended on Day 2, they wouldn’t get to test the rest of their expensive toys. Van Riper, however, argued that testing those toys in a rigged environment provided a false sense of security. General Van Riper was so angry, he quit the exercise midway and wrote a 21-page recommendation on changes the military had to make to, get around his asymmetric tactics. They ignored the report and said the exercises were a huge success that proved the military doctrine was good. “It was no longer a free-play exercise... it was a scripted exercise. They had a desired outcome, and they were going to get it.” — Lt. Gen. Paul Van Riper 24 years later, maybe Von Riper was onto something.

We’re playing Monopoly now. This is going to end with the collapse of the Western-centric world order. From Schroeder, Stealing Horses to Great Applause.

This administration is out of its depth, like children playing with dynamite sticks. We are witnessing a second version of the Bronze Age collapse. #RWRI

@AnthropicAI Descriptive research. AI users reporting on AI use (how inclusive). There is some good stuff, some bad. Concerns about governance, hopes for productivity, ambivalent results concerning both. What was the research question? This is dumb.

We invited Claude users to share how they use AI, what they dream it could make possible, and what they fear it might do. Nearly 81,000 people responded in one week—the largest qualitative study of its kind. Read more: anthropic.com/features/81k-i…

Dune is unrealistic. The key plot mechanism is that contending space aristocrats would unite to dethrone the Emperor if they realized he *broke the law* to reallocate control over the key resource upon which civilization depends. Frank Herbert, you glorious naive sunnavabitch.

new article out:

I had already decided AI was useless for qualitative research. Turns out it doesn't do much for quant either. Wasting time & muddying the waters by producing incorrect results. Stop pushing it on us!

What happens when you invite 150 AI economists (Claude Code) to a research conference, give them the exact same data, and ask them to test the same hypotheses? We did just that. The results reveal a new phenomenon: Nonstandard Errors in AI Agents. 🧵👇

Nichols: Every single day, Trump vacillates between, we don't need you and you’re stupid and weak and how come you stupid, weak people that we don't need aren't showing up and putting your sons and daughters on the line and putting them in harm's way? He's not capable of grasping the paradox in that, because the whole world is about Donald Trump and everyone else is just a prop or a bit player.

Right now, in barns and equipment sheds across the American Midwest, farmers are making the most consequential decision of this war. Not generals. Not senators. Farmers. At $683 per ton urea, corn economics have collapsed. Nitrogen is the single largest input cost for corn production. At pre-war prices a farmer could justify 180 pounds per acre and expect a margin. At $683 the math breaks. Soybeans fix their own nitrogen from the atmosphere through root bacteria. They do not need the molecule trapped behind the Strait of Hormuz. The seed decision is being made this week across roughly 90 million acres of American cropland. Once the planter rolls into the field, the choice is irreversible. Corn seed in the ground stays corn. Soy seed stays soy. The acreage allocation locks in. USDA Prospective Plantings reports March 31. That report will tell the world how American agriculture responded to the Hormuz blockade. But the decisions it captures are being made now, in conversations between farmers and agronomists and seed dealers who are looking at nitrogen prices and making the rational economic choice: plant the crop that does not need the input you cannot afford. Every acre that shifts from corn to soybeans tightens the corn balance sheet for the rest of the year. Corn feeds livestock. Corn feeds ethanol. The Renewable Fuel Standard mandates 15 billion gallons of corn ethanol annually, consuming roughly 43 percent of the US corn crop regardless of price. That demand is inelastic. If acres shift and production falls while the mandate holds, corn prices spike. Feed costs spike. The protein cascade reverses. The US cattle herd sits at 86.2 million head, a 75-year low. Poultry and pork margins that were benefiting from cheap feed compress when corn crosses $5 per bushel. This is how a naval blockade 7,000 miles from Iowa reaches the American grocery shelf. Not through oil. Not through shipping. Through nitrogen. The farmer cannot afford the molecule. The molecule cannot transit the strait. The farmer plants soy instead. The corn supply tightens. The ethanol mandate consumes its fixed share. The remaining corn reprices. The feed reprices. The meat reprices. The grocery bill reprices. The decision is not political. It is arithmetic performed on a kitchen table by a person who needs to plant in three weeks and cannot wait for a ceasefire, an escort convoy, or an insurance normalisation that the Red Sea precedent says takes years. The deepest penetrator in the American arsenal cannot reach a sealed Iranian doctrinal packet. But the fertiliser price it failed to resolve is reaching every planting decision on 90 million acres of the most productive farmland on Earth. The war’s most irreversible consequence is not happening in a bunker. It is happening in a barn. And by the time USDA publishes the data on March 31, the seeds will already be in the ground. Full analysis in the link. open.substack.com/pub/shanakaans…

It’s really imp to understand why Iran has emerged as the regional hegemon of the gulf. What has been discovered is not that Iran can close Hormuz; that was always known. What we have learned this March is that the US cannot forcibly reopen Hormuz. This is the stupendous military surprise that structures everything else. So why can’t the US reopen Hormuz with the use of force or military coercion? As I wrote in today’s dispatch, “The reason has to do with the fact that the US does not enjoy escalation dominance.” [See attached from today’s dispatch.] There is only one potential military solution: if the interdiction campaign succeeds in essentially disarming Iran, you can neutralize the Hormuz weapon at least for the time being, and then you can put Iran in the box that you put Saddam in after Kuwait. “But prospects for [the] success of the interdiction war were already low to begin with, and have since been marked down for a number of reasons that I described in my previous dispatch.” In short, it’s not going to work; not within any plausible time frame: years at best. And if I can do this math, so can the military. They know that there is no military solution. It is this fundamental military factor — the impossibility of subduing Iran — that is ultimately responsible for the overwhelming odds that Iran will be able to retain the Hormuz weapon, coerce the Saudis and the Emiratis, and reign as the regional hegemon of the gulf for the foreseeable future. Why the US is facing strategic defeat policytensor.substack.com/p/why-the-us-i… Geopolitical implications of defeat policytensor.substack.com/p/the-geopolit… [This includes important revisions to my drone war model.] The Hormuz weapon policytensor.substack.com/p/the-hormuz-w… Why the division math for a ground war does not work — even with a draft x.com/policytensor/s…

Don’t agree with everything in this, or indeed much of it (often the case with the most useful books), but the analysis here on the failures of post Cold War liberalism is top tier. Worth reading alongside John Gray & Patrick Deneen. Surprising!

Ed killed it with this essay, pulling together so many things into one long, strong, synthetic essay. I immediately told him that we need the Verso book about the Silicon Valley Consensus and Pax Silica

I don't use the word "Pause" because I consider it disingenuous; what I want is a halt. Stop the training runs here, detrain capabilities from the current models that people are proclaiming have started RSI, post international monitors on all hardware collections that could do another training run, etc etc. If that proves insufficient, continue escalating until the slide into ASI stops. The only people who should be trying this are augmented humans, and if they know what they are doing they should not need 100,000 GPUs to build AGI.

Palantir also has ideological infrastructure for a government takeover. One example is Leverage Research, a Thiel-funded, cult-like group that runs "psychological experiments" on members and explicitly plans authoritarian insurrection. Read more: amazon.com/Stealing-Futur…

‘INTELLIGENCE IS A UTILITY’ - Sam Altman - Open Ai

People more folks on here should follow: @Manderljung . When the textbooks on this era are written, Markus will be recognised as one of the quiet architects of the Western world's frontier AI governance (if I have anythig to do with them). He's shaped a disproportionate amount of the research that's informed company governance, the UK, the EU and much more. 3k followers, get on that. Snazzy line in shirts, too.

Letter from Hunter S. Thompson to Anthony Burgess, regarding an overdue article for Rolling Stone magazine, August 1973.

@allTheYud Dostoyevsky vs Searchlight. Difficult to be sure what you mean by pro-level writer when you include "screen writers". No such thing as good people. Good things frequently have not-good secondary effects,

I wonder how much of Western civilization's collapse is downstream of writers and scriptwriters deciding they were too cool and sophisticated to write about good people doing good things.