Admiral Bunter
In which the topic of ethics recurs.
And in this chapter we get a direct reference to the events of “Wise Child”, described in sufficient detail to assist anyone who hasn’t read it already. I wonder if that portends that Disian is going to be showing up in this story. It might not, it might just be mentioned for the light it sheds on the present circumstance, but I’d like to see Disian again and learn how she’s getting on.
It appears that Admiral Bunter was intending to do away with Tolly, but was stopped by Tolly having reset his Ethics module to the agreed value. And Tolly’s explanation of what he did to the Ethics module, and why, seems to have gone some way toward rebuilding the Admiral’s trust in him.
It’s interesting that Tolly was expecting to be done away with but didn’t do anything to stop it, because it would still be a better result than falling back into the hands of the Institute. (And also because there wasn’t anything he could have done to stop it without proving to be the kind of person Admiral Bunter suspected him of being.)
One of those little confluences: two chapters in a row, the viewpoint character finds themself locked in a room by the AI controlling the ship. One tries opening the door to be certain, the other makes a point of not even trying. Different circumstances call for different responses.
I don’t believe we’ve heard of Wez diaGrazia before. This doesn’t surprise me; as far as I recall, no director of the Lyre Institute has so far survived their first appearance in the series.
I find it surprising that Tolly apparently doesn’t know the effects of ethics setting seven. Admiral Bunter says that it requires a proof of threat which isn’t present for him to kill Tolly, but Tolly, was nonetheless expecting to have been killed.
Maybe Tolly was sufficiently tired that he hadn’t reasoned out that he was a potential threat but not a proven threat to the Admiral.
It’s also possible that the ethics setting interacts with other elements of the AI’s personality in complicated ways, so that its difficult to predict exactly what degree of threat will allow an AI to kill.
The question of what constitutes a proven threat is also complicated in this case by the fact that a mentor could act against an AI without leaving any proof for the AI to find. Tolly knows it, and in Admiral Bunter’s place he wouldn’t wait for evidence, and he knows Admiral Bunter knows it, because it’s already happened to Admiral Bunter once. Apparently the Ethics module isn’t convinced.
PA: Tolly was expecting to be done away with.
I’m not quite sure that he was strongly expecting it so much as he did not know what to expect and found it easier to prepare for the worst. He knew that a self-aware independent logic would be tempted to do away with the last mentor who had been inside his core. But he also know that he had reset the Ethics module up to 7, which gave the Admiral flexibility in making the decision.