Sci-fi and Heroic Fantasy discussion

This is our discussion of the novelette:


"Swift as a Dream and Fleeting as a Sigh" by John Barnes

This story is part of the Edge of Infinity anthology discussion.

★★★★1/2

Synopsis:  Two humans, Laura and Tyward, get couple therapy by an A.I. psychologist. Ty works as a mining robot analyst who tries to optimize evasion strategies of those robot ants. Some of them built A.I. models to please humans, one of them manipulating Ty's emotions.  Laura doesn't like Ty's sensitivity, his lack of manliness.

Review:  This is certainly not the first SF work employing an A.I. as a psychologist, e.g. Gateway by Pohl comes to mind. But it tells the story from the point of view of this A.I. which was refreshing new for me, as it explains the situation of A.Is when talking to humans: eons of times between human words, i.e. some milliseconds which have to be filled with tons of analysis, watching thousands of movies. In short: Humans are so slow.

In parts, I found the background a bit difficult to digest, like in some Rajaniemi story. But in general, the story flowed nicely, and the motivations of the diverse forms of A.I. (psychologist and mining ants) were played out very good.

I found the story thought-provoking, though it didn't bring up any new impulses concerning post-singularity besides the point of view narration. The end wasn't motivated and came out of a sudden.

Boy, there was so much great stuff in "Swift as a Dream and Fleeting as a Sigh," but such a flat and disappointing ending.

I really liked the exploration of the AI. (AHAI — did the story say what that stood for? Or am I supposed to guess?) Interesting choice of PoV, an pretty well executed. I especially liked the analogy of a human "conversing" with an oak tree.

We have two different varieties. The AI as psychologist, who has, together with his brother AI's, been trying to guide the human race (which has been slowly dying out, it seems.)

We also have some fascinating, adaptive AI's in small worker robots. They've been set up to innovate and evolve, in the hopes that they occasionally hit on improvements it can be passed on to their silicon brethren. And one of them has developed a fascinating strategy of sucking up to the boss. (Which apparently really ticks off the boss when he discovers the AI is manipulating him. Strangely, that discovery of AI manipulation doesn't change their trust for the psychotherapist AI.)

I read this eagerly, because it seemed what a fascinating story. I was wondering if the psychotherapist AI would either take the side of the worker bot AI's, or yearly conspiracy or something.

Instead the ending was abrupt, flat and as Andreas says, unmotivated. The unfortunate conclusion to a story that really had me intrigued.

I'm only taking off a few points for the ending, because I enjoyed the trip....

4 Stars ****

One of the characteristics of the robotic AI "ants" in this story is that they record what they are doing so if something goes wrong, improvements can be shared with the other units. So I was amused reading Why You Shouldn't Worry About Self-Driving Car Accidents in IEEE's Spectrum that one of the pluses of Google's self-driving cars is that if they have accidents they can update all other cars to improve them.

So SkyNet begins....