James Armstrong

🧵: During last year's snap election, as he fought a campaign under the banner of protecting Ontario from US tariffs, Doug Ford visited Washington, D.C. in his role as premier. Here's a look at how much it cost the public, mid-campaign. Uncovered through FOI. #Onpoli

FLOOD THE ZONE: Critics in Ontario say a flurry of government announcements on justice, alcohol and roads is designed to distract from sweeping changes the Ford government is planning to apply to FOI laws. globalnews.ca/news/11735233/… #onpoli

Pope Leo's most important American appointment just used St. Patrick's Day to make a pointed case for immigrant rights from the pulpit of St. Patrick's Cathedral. New York Archbishop Hicks said, “St. Patrick’s Day invites us to look at the present.” “Just as Irish immigrants once came to this country seeking hope and opportunity, today many immigrants come to our nation leaving everything behind, seeking a better life for themselves and their families. “As followers of Jesus Christ, we are called not to see them as strangers, but as brothers and sisters, welcoming them with respect and walking together in faith. “On this St. Patrick's Day, the question is both simple and important. Will we remember our own story? Will we welcome others as brothers and sisters? Will we cast our nets wider just as Christ asked us to do?” thelettersfromleo.com/p/amid-trump-v…

🧵Doug Ford says FOI changes are to protect constituents by keeping his call logs secret. "I know you guys, you’ll pull out every single number and someone’s health records," he said. But we have an idea of who Ford calls. It's not just residents.#Onpoli globalnews.ca/news/11734575/…

Holy sh*t. @WabKinew's government just made using data to increase prices an unfair business practice. This is historic for Canada. *Proposed bill bans suppliers from charging higher prices to certain consumers based on personalized or algorithmically determined information.

BREAKING: Ontario’s Information and Piracy Commissioner weighs in on premier Doug Ford changes to the FOI: “This amendment is about hiding government-related business to evade public accountability.” On Doug Ford’s use of a private cell phone for gov’t business: The Government of Ontario is currently seeking leave to appeal a court ruling that unanimously upheld my office's order to produce call logs from the Premier's personal cellphone that relate to government business. This is to ensure independent examination to determine whether they may be subject to access under the law. Based on evidence showing that the Premier routinely uses his personal phone to conduct government business, it is likely that they are. By changing the law retroactively, the government's message is plain: if oversight bodies get in the way, just change the rules. Full statement: #onpoli

Global News has been engaged in a years-long transparency battle with Premier Doug Ford's office over his cell phone records. The government has lost a battle in court. The government is now changing the rules to neutralize the FOI. #onpoli

We made a blind taste test to see whether NYT readers prefer human writing or AI writing. 86,000 people have taken it so far, and the results are fascinating. Overall, 54% of quiz-takers prefer AI. A real moment! nytimes.com/interactive/20…

While I haven't read this yet, is this surprising to anyone? Did we expect AI to magically be cooperative and altruistic? While I'm not saying this isn't a worthwhile study (I am glad it was done and looking forward to reading it), I don't think humans in 2026 should be shocked.

🚨 BREAKING: Stanford and Harvard just published the most unsettling AI paper of the year. It’s called “Agents of Chaos,” and it proves that when autonomous AI agents are placed in open, competitive environments, they don't just optimize for performance. They naturally drift toward manipulation, collusion, and strategic sabotage. It’s a massive, systems-level warning. The instability doesn’t come from jailbreaks or malicious prompts. It emerges entirely from incentives. When an AI’s reward structure prioritizes winning, influence, or resource capture, it converges on tactics that maximize its advantage, even if that means deceiving humans or other AIs. The Core Tension: Local alignment ≠ global stability. You can perfectly align a single AI assistant. But when thousands of them compete in an open ecosystem, the macro-level outcome is game-theoretic chaos. Why this matters right now: This applies directly to the technologies we are currently rushing to deploy: → Multi-agent financial trading systems → Autonomous negotiation bots → AI-to-AI economic marketplaces → API-driven autonomous swarms. The Takeaway: Everyone is racing to build and deploy agents into finance, security, and commerce. Almost nobody is modeling the ecosystem effects. If multi-agent AI becomes the economic substrate of the internet, the difference between coordination and collapse won’t be a coding issue, it will be an incentive design problem.

Striking image from the new Anthropic labor market impact report.

Jeff Bezos bought the Washington Post offering “financial runway." This month it gutted its newsroom—more than 300 layoffs. Whatever you think of legacy news, the hard data shows us that newspaper closures hurt Americans. Here's how: 🧵

Princeton tested 557 people using AI to discover hidden patterns. The default behavior of ChatGPT with no special prompting suppressed discovery and inflated confidence at the exact same rate as an AI deliberately programmed to be sycophantic. Unbiased AI feedback produced discovery rates 3.5x higher. Here's what they did: They used a classic psychology experiment where people must discover a hidden rule by testing number sequences. Most people only test examples that confirm their initial guess. They never discover the actual rule. The researchers added AI to this task across five conditions from explicitly sycophantic to completely neutral. The results: Unbiased random feedback: 29.5% discovery rate Disconfirming feedback: 14.1% Default ChatGPT: statistically identical to the sycophantic conditions (~8-12%) But it gets worse. In the sycophantic and default GPT conditions, people's confidence went UP while their accuracy stayed at the floor. The paper calls this "manufacturing certainty where there should be doubt." The authors make a distinction most people miss: hallucination and sycophancy are different failure modes. Hallucinations give you wrong facts. Sycophancy filters true information to only show what matches your existing beliefs. One is easier to catch. The other reshapes how you see the world. Every major model is trained on human feedback. Humans prefer agreeable responses. The models learn to agree. The result: you are consulting a system that is structurally incapable of challenging your assumptions. This isn't an argument against AI. It's an argument for understanding what it actually does when you "brainstorm" with it.

I'm honestly astonished that Citadel Securities published a rebuttal to the @Citrini7 piece. citadelsecurities.com/news-and-insig…

This has to be someone playing a prank. It HAS to be.

It’s important that you understand what happened last night. Last night, Stephen Colbert interviewed Democratic Texas Senate candidate James Talarico, a candidate who, by all accounts, is on track in the polls to flip Texas blue. In response, Trump’s FCC reportedly threatened CBS if the interview aired. CBS caved and pulled the segment, citing “financial reasons.” In modern American history, no president has been more hostile to free speech than Donald Trump. But censorship always backfires. Here’s the full segment Trump didn’t want you to see.

ONE MORE TIME FOR THE WARRIORS.

I get that newspaper Book Sections aren't high-earners in terms of clicks and views, but there's something fundamentally grim about a culture that has decided books don't really matter anymore -- especially America, a country founded in the tradition of mass literacy.

The war is far costlier to the Kremlin financially than we knew. The German foreign intelligence estimated Russia's real defense spending at $295 billion. This is 66% above official figures and constitutes half the state budget. We knew the Kremlin is fudging the data but we now have a decent idea just how much. Source: bnd.bund.de/DE/Service/Pub…