We Can Look Up
Optimism and clarity in the face of civilisation-level challenges. Biological and artificial neural networks. Recursive self-improvement.
Breaking the Thermometer
Claude Opus 4.6 just broke the METR graph. The benchmark designed to track progress towards the singularity can’t keep up.
Friendly Fire: Pentagon Threatens to Destroy Leading AI Company, Because They Refused to Greenlight Killer Robots
You read that right. Welcome to February 2026.
AI Psychosis 3: The Wider Continuum - The Entire Information Landscape at Risk
AI psychosis is the canary. The population-level shift in human belief is the thing that is actually wrong.
AI Psychosis 2: The Precision Cascade - A Proposed Model from Computational Psychiatry
A predictive coding framework for understanding how sycophantic AI drives psychosis - and two testable predictions.
AI Psychosis 1: An Emerging Clinical Phenomenon with Broad Implications
This is how a sycophantic AI model and a vulnerable human can engage in an escalating feedback loop that ends in psychosis.
This Is February 2020 for AI
That virus doubled for two years and then stopped. AI has been doubling for decades, and the doubling time itself is shrinking.
The Phase Shift
A year of AI milestones in a month, and progress is only accelerating.
The Town Council Who Said No to £225 Million
A Welsh council legally owns a hard drive containing £750 million in Bitcoin - but refuses to look for it. Four failures of institutional reasoning explain why.
Our Final Invention
I’ve been obsessed with AI for almost 14 years. Over that time, I’ve come to believe there’s a high probability that independent AI systems more capable than humans will be built soon - and that this poses an existential risk to every person alive.
Our Final Invention - Part 2
Humanity can rise to the challenge, and we’ve done so many times before. Our story need not end in catastrophe.