Resources

An animated explainer from Rational Animations. If superintelligent AI could cause human extinction, why don't we simply stop building it? The video lays out the main arguments, the practical difficulties, and proposed responses — clearly and without hype.

A long essay by the CEO of Anthropic on what a positive AI future could actually look like — in detail. Most discussion of AI focuses on risk. This is the other side: what does a world where everything goes right look like? Essential reading for the full picture, not just the warnings.

A short, uncomfortable essay. Horses were once central to the economy — transport, agriculture, war. Then they weren't. Jones asks what happens to a species when its primary economic role disappears. Read slowly.

A policy paper from OpenAI arguing the transition to superintelligence will require an ambitious response — new safety nets, democratic input into AI alignment, and mechanisms to share the economic gains broadly. Significant not just for the ideas, but for who is saying them: the company building superintelligence is asking governments and communities to act now.

A scenario piece from a macro research firm, not a prediction. What happens if AI adoption accelerates exactly as the optimists expect — but the speed of disruption outpaces the economy's ability to adapt? A careful look at left-tail economic risks that most AI commentary ignores. Thought-provoking whether or not you buy the scenario.

The most rigorous employment data on AI's impact to date. Using ADP payroll records covering 25 million US workers through September 2025, researchers found that workers aged 22–25 in high AI-exposure roles — coding, writing, legal, customer service — have seen employment drop by around 20% since late 2022. Workers in the same roles aged 35+ grew over the same period. The headline unemployment number hasn't moved. This paper explains why that doesn't tell the full story, and what the early signal actually looks like.

The framework Future Together uses to think about preparing for different futures. If things go badly, we start at the bottom — physiological needs, safety, shelter — and plan up from there. If things go well, we think about how to help people reach the top of the pyramid. A useful map for conversations that might otherwise feel abstract.

A clear-eyed look at the economics driving physical automation. The argument: robotics will follow the same cost-curve trajectory as software AI — and the timeline is shorter than most people expect.

A model of what community-led resilience actually looks like in practice. Transition groups around the world have been building local food production, energy independence, and strong community networks for years. Whatever disruption comes — economic, technological, environmental — the social infrastructure this movement builds is what makes communities genuinely resilient.

ALLFED (Alliance to Feed the Earth in Disasters) is a nonprofit research organisation that studies what happens to global food systems when things go seriously wrong — nuclear winter, supervolcanic eruption, grid collapse from cyberattack or solar storm, simultaneous crop failures across multiple regions. A clear-eyed map of the scenarios worth understanding, and why local resilience matters.

A weekly newsletter tracking AI capabilities research, written by Jack Clark (co-founder of Anthropic). Technical but not impenetrable — if you want to stay across what AI systems can actually do week to week, this is the most reliable signal in the noise.

A live overview of what's happening in the world — not just AI, but the full picture: military operations, environmental disruptions, political turmoil, border tensions. The things that can compound and cascade. Useful context for understanding what kind of future we might actually be preparing for.