I have compressed the standard case for taking loss of control seriously into four steps, aimed at a practical preparation question rather than the usual technical debate. None of it is new here. I have tried to state each step so it can be falsified.

1. We build systems more capable than us, in a race no one easily leaves.
2. Such systems form their own instrumental sub-goals.
3. As capability rises, we lose reliable inspection, correction, and shutdown.
4. From there the system follows its own course, and the human is no longer the reference point.

My ask is narrow: which step does not hold, and on what evidence? Each step has an explicit "breaks if" condition and the adversarial trail in the full write-up. I will concede any step in public if it breaks. Link in a comment below.

I used an LLM to help draft this post and it likely contains >10% AI-generated text, but I’ve edited/rewritten it extensively and endorse it.