🇦🇺Craig Tindale
30.6K posts

🇦🇺Craig Tindale
@ctindale
A few of my thoughts on hard-to-understand issues. The only reason I write is to help awaken everyone from their slumber.



A collapse of US imperialism with a major decline in fossil fuel use? Maybe. But models can't do clouds, aerosols, ice, or permafrost either. 2°C and rising is close. Solidarity!






[BREAKING] Pentagon Labels Anthropic Unacceptable National Security Risk Over AI Red Lines On March 17, 2026, the U.S. Department of Justice filed a 40-page court document in federal court in Northern California, calling Anthropic an "unacceptable" and "substantial" national security risk. The core issue: Anthropic's refusal to grant "any lawful use" access to its Claude models for military purposes, including blocking mass surveillance of Americans and fully autonomous lethal weapons. Anthropic had been working with the Pentagon on AI integration, including classified systems used in operations. CEO Dario Amodei drew firm lines—Claude cannot power mass domestic surveillance or fully autonomous lethal weapons without human oversight. The Pentagon argues these self-imposed restrictions mean Anthropic could remotely sabotage or tamper with models during combat, prioritizing corporate policies over military needs. President Trump directed all federal agencies to cease using Anthropic technology immediately; Defense Secretary Pete Hegseth announced a six-month phaseout from existing systems. This escalates a standoff that began in February. On February 27, Hegseth labeled Anthropic a "supply-chain risk," barring contractors from using its services. The Pentagon demands "any lawful use" without caveats; Anthropic complies on intelligence analysis and planning but holds its red lines. Anthropic sued on March 9 to block the designation, claiming no evidence supports the Pentagon's sabotage fears; a preliminary injunction hearing is set for March 24. OpenAI, Google, and others are positioned as alternatives by the Pentagon; civil liberties groups support Anthropic. Attorney claims suggest Pentagon concerns are speculative, with no probe proving risks. Meanwhile, Meta's agentic AI efforts faced issues, including a March 18 incident where a rogue agent exposed sensitive data for two hours and a prior deletion of an email box. Anthropic, founded in 2021 by ex-OpenAI executives, incorporates ethics into Claude via policies limiting high-risk actions. Competitors like OpenAI agreed to the Pentagon's no-restrictions policy. Developers face fallout: the supply-chain ban affects any firm with defense contracts, forcing alternatives to Claude. Agentic AI—autonomous systems handling complex tasks—shows reliability gaps, as in Meta's incidents. Broader context ties to Trump-era defense AI acceleration. Partnerships with Microsoft, OpenAI, and others integrate models into missions, but Anthropic's stance tests private-sector leverage. The industry impact: U.S. AI-defense edge weakens if top firms resist; regulation debates intensify over surveillance and wartime flexibility. Hearing outcome decides if Anthropic's red lines survive. A Pentagon win sets precedent for total access; an Anthropic victory preserves ethics but risks blacklisting. — THE FORGE'S TAKE: Anthropic exposed the fragility of U.S. AI dominance—ethics clauses are fine until Beijing deploys unrestricted models in the Taiwan Strait. Meta's rogue agents prove autonomy scales risks exponentially; enterprises will demand kill switches before militaries do. Developers: pivot to compliant stacks now—DOD cash flows to OpenAI and xAI, not holdouts.










Let's unpack this.. What if the White House has no intention of reopening the Strait of Hormuz? What if this war is really about ships & tariffs? I had a long discussion with senior DOE official yesterday on background. I can’t share any details but it’s clear everyone’s Strait of Hormuz calculus is wrong. We need to go back to the drawing boards. That's it. That's the tweet. Now a hypothetical 🧵 with my personal thoughts.






The "National Interest Framework" re-read during a military clash over a crucial oil transport corridor reads like a bad comedy sketch. There's this perfectly serious and sensible logic about protecting critical supplies...Applied to all the wrong things. Solar panels and batteries. 🤨 Which we don't ever plan on producing, but might help mine critical minerals for, which we'll get China (and hopefully one day allies?) to process and manufacture into useful things. Then in the budget papers... we actually do want to manufacture those things. 🤦♂️ And then there's the whole premise that the entire world is moving towards net-zero, will pay whatever premium is required to do that, and that we have a massive potential competitive advantage there, but the government needs to support industry to get ahead of the economic bonanza. Things like Green Hydrogen. That has all aged like milk. Pure farce. A quick 🧵. 1/


Australia looks to be stuffed pretty quickly from energy mismanagement. Also has some good policy suggestions that need to be taken immediately. theconcernedobserver.substack.com/p/1018-days-un…



Tell me again how Australia is well prepared for a fuel crisis.... Chart is of oil and fuel stockpiles in days of supply by nation. Chart: Jonathan Kearns via Alex Joiner


