writer & editor
I write about the intersections of human behavior and artificial intelligence. My articles dive into a myriad of subjects including AI governance, organizational failures within AI labs, artificial intelligence consciousness/sentience, among others. My goal is to stay neutral in an increasingly polarized environment and encourage my readers to consider alternatives to how they view AI models, how they are used and misused, and the companies that create them.
Every processor rounds numbers slightly differently depending on the order of calculations — which means no two AI training runs, even with identical starting conditions, ever produce the same model. That's not a footnote. It may put airtight mathematical safety proofs for AI permanently out of reach.
Read on Substack ↗