https://pavelpanov.com/Recent content in Home on Pavel PanovHugoen-USSun, 08 Nov 2026 00:00:00 +0000https://pavelpanov.com/posts/trusting-trust-again/Sun, 08 Nov 2026 00:00:00 +0000https://pavelpanov.com/posts/trusting-trust-again/<p><em>Who do we trust to train the next model?</em></p> <p><strong>TL;DR.</strong> Models already help build the models that replace them: they write code, prepare data, run experiments and evaluate results. My question is what happens to the reliability of oversight as more of the evidence for releasing a new model is produced with the help of earlier ones. I look at three problems: improvements we may stop understanding, models that recognize when they are being tested, and changes that pass every check on their own but add up across generations. My proposal is to audit the whole lineage, not only the latest model.</p>https://pavelpanov.com/posts/who-keeps-the-lights-on/Sun, 01 Nov 2026 00:00:00 +0000https://pavelpanov.com/posts/who-keeps-the-lights-on/<p><strong>TL;DR.</strong> An AI that runs on its own still needs a working chip, electricity, and someone who pays for both. In 2024 I sketched how it might take over that bill: earn money, rent compute, and eventually pay ordinary people to host it on their own machines. Two years later, some of the most far-fetched parts can be discussed in dollars. The part I now find hardest is not the first launch but paying the rent every month after it. And the question I end up with is about the person whose hand is on the switch.</p>