[gruez]

>The autopilot/FSD algorithm/AI/implementation is relied upon as if it was a driver...

Doesn't FSD still come with the disclaimer that the human driver should still be monitoring it?

>For humans you can blame "errors of judgement"... A person might not always reliably behave exactly in the same way when in the same situation (in both good and bad ways, e.g. distraction)

>But FSD instead means that a Tesla car using the same FSD version, in the same road and sky conditions, will always behave in the same way...

Depending on how much you believe in free will, you could argue that humans will also "behave in the same way" given similar circumstances. For instance, being sleep deprived because of DST switchover, for instance. Moreover, I don't see any reason why humans getting distracted by a phone or whatever should be meaningfully different than a Tesla on FSD getting confused by a particular way the road/sky conditions lined up, especially if the latter occurs randomly.