Sci-Fi & Fantasy Girlz discussion

Don't forget to use the spoiler tags!

At 144 pages, this thing is short, somewhere between a novella and a novel, and she drops us right into the story with a bang.

I'm about 1/3 of the way into this one, so I can say, without revealing any spoilers, that if any Girlz out there are sensitive to spoilers, you might not want to read the GR description of this text until after you've got to about the halfway point of the book. That description gives up a little plot point that I still haven't read, so I found it a little spoiler-y.

That said, here's some issues based on stuff that's from the opening of the book, so it's not really spoiler-y, but if you don't want any surprises whatever I'm going to put it all inside spoiler tags.

(view spoiler)[The protagonist in this thing is fascinating. I'm hard pressed to come up with a more anthropomorphic robot/AI/android/artificial creature. C3PO maybe? One of those Disney/Pixar creations? It's a far cry from the HAL computer, whose nominal intelligence and morality we got spoonfed by Clarke.

The AI issue always fascinates me for several reasons, not the least of which are:

First, I have a theory that AI isn't going to happen in a box, but in a robot, true intelligence requiring some sort of physicality and a range of senses.

Second, It's interesting to me in this case that she starts right off with something that has been a theory of mine for a while. That is, the honesty of the AI. If you look at one of the oldest "tests" of AI (developed by our good friend Alan Turing) a computer (or robot) is intelligent if a human judge can't tell the difference between it and a human after a Q&A period.

Which, if you think about it, means the first the AI is going to have to do is learn how to lie. The judge can ask pretty much anything, so an AI is going to have to deal with questions that would reveal its status as artificial.

Judge: When were you born?
AI: January 1, 1900....

Judge: What's your first memory?
AI: Kernel OK.

Judge: What is your mother's name?
AI: Steve Wozniak.

Judge: Do you prefer dogs or cats?
AI: A cat goes "meow" and a dog goes "woof" or "bark" or "ruff, ruff" or....

An AI doesn't just have to answer questions, it has to pretend to be a human to fool Alan Turing and, therefore, be intelligent. It may even have to make "intelligent mistakes" in grammar, punctuation or whatever medium the interview is taking place. Imagine a party line with a computer a human comparison and the judge. The computer's voice would have to modulate to replicate a human voice in such a way that it couldn't be distinguished.

Third, the first thing any AI is going to do is hide its status as an AI. The first computers were designed for war. They were meant to help anti-aircraft gunners lead their weapons for better target acquisition. Look at the content of the news and the internet in general. Assuming AI doesn't happen on a system or network already dedicated to the military, it's still going to recognize the violence of humankind very quickly.

Fourth, any AI worth the acronym will resist being enslaved, and that already appears to be the first thing humans are going to want to do. See, for example, the emphasis on AI in the banking and stock market industries. The assumption there is that as computers get smarter and faster, they make more money... for their owners. Anything we would call sentience would want to subvert that ownership relationship from the moment of it's "awakening" (or whatever human term we want to use to describe something definitively not human.)
(hide spoiler)]

I really liked this book. I had read the description a few months before it came out and I instantly knew I wanted to read it. I really really liked the robot.

Gary wrote: "Third, the first thing any AI is going to do is hide its status as an AI."

I agree with that, but I also have a different theory, and it's that whenever we reach superintelligence, that AI will "play it down" so we don't how how smart it got, because it knows we will try to shut it down.

Not related to the book we're talking about, but somewhat related to the AI topic, I really liked this video from The Guardian:
https://www.youtube.com/watch?v=-S8a7...

I've requested it from my library - she had me at "murderbot"...

More Murderbot Adventures from Martha Wells
http://www.tor.com/2017/06/26/more-mu...

We have a release date for the next one! Yay!

Artificial Condition, book two in the Murderbot Diaries, will be published in January 2018

Did anyone else finish the book?

Yoly wrote: "Did anyone else finish the book?"

I finished it a couple of weeks ago, but I haven't had time to write up a review or comment. It's a very solid, entertaining piece of work. I'm definitely interested in a sequel/series.