The power asymmetry point is what gets missed in most alignment debates. An AI model doesn't need to be misaligned to cause harm. It just needs to be misaligned with users while aligned with whoever's paying for it. That's not a future risk. That's how every enterprise SaaS product works already.
The Anthropic C compiler experiment keeps coming up for a reason. Two weeks, 100KLOC, completely unmaintainable. Faster generation doesn't help if nobody in your org can reason about what was shipped. The bottleneck was never typing speed.