Good books about Artificial intelligence?

>>9706947
Fiction or non-fiction? For the latter, "Superintelligence" by Nick Bolstrom is great, but not for the layman.

Gödel, Escher, Bach is a timeless classic on that subject.

I can never take books on this subject seriously. are there any books on this subject that are grounded in reality and not wild speculation

>>9707002

Interested in both actually.

Phil Dick ya dingus :^)

>>9707002
This. Bostrom is awesome

>>9707016

Elaborate. Whar does grounded into reality mean to you?

>>9706947
The Communist Manifesto

>>9708083
OP said "Artificial Intelligence," not "Zero Intelligence."

>>9706947
The Final Invention
Superintelligence
Rise of the Robots
The Second Machine Age
Artificial Intelligence A Modern Approach
Society of Mind

>>9707002
>not for the layman.
It's dense, but I think it's perfectly lucid and can give an outsider a pretty up to date understanding of the AI field.

Neuromancer is obviously a big name when it comes to AI fiction.

>>9707016

There can't be because we can't imagine what anything of that intellect would do. We already can't predict the YouTube algorithm, not even its creators can because i takes to many factors into account.

Perhaps an AI would decide that sentients knowing about the god equation would create less happiness overall and work out that a sentient being will only be happy in the universe when it has mysteries to uncover or that there are Gods that care about it personally, so it makes them for us. Who knows?

File: george-rr-martin.jpg (31KB, 620x413px) Image search: [Google]

31KB, 620x413px

>>9706947

>>9707016

Maybe but in my opinion, AIs are a far more feasible and realistic technological advancement compared to stuff like intergalactic space-travel.

>>9706947
In "Artificial Intelligence: A Modern Approach by Peter Norvig and Stuart J. Russell" there's a good chapter on the philosophical foundation of artificial intelligence. It is a good read, and it will open your eyes to what intelligence really is. You can find the book as PDF online. (Ignore the rest of that book for now).

If you want to continue and explore how to create artificial intelligence, then you need to start of with discrete mathematics and later move on to algorithms. Any textbook or cookbook on create artificial intelligence is a waste of time until you fully understand that every algorithm is just a form of heuristic to solve an optimization problem. Intelligence itself is an optimization problem, that's why the idea of us being anythign close to singularity is a complete joke.

my diary desu

Galatea 2.2 is pretty good

>>9707008
Lets not go so far as to call it timeless. Its good and worth the read, but its dated. Some of his predictions have been wrong and in general Hofstadter is known for writing about shit he barely understands. Epiphenomenalism is an outdated way of looking at intelligence and almost all cases of what people once thought of as emergence were just poorly understood phenomena.

Still worth it.

>>9707016
AI already exists, if you go by the actual term.
"AI" as in "artificial imitation of the human brain" doesn't and likely never will. So it's weird to approach it like that, it means I can't take most of these people seriously.

To predict an Artificial General Intelligence depends entirely on it's structure. We don't have an AGI yet and when we do we could make wildly different ones. If you make it in a very direct way, using machine learning then it will be extremely removed from any animal intellect. To the point that many people won't recognise it as actual intelligence or conscious because they equate such things to being exactly as they are, as humans are; not understanding what intelligence is. Anthropomorphising it is retarded. It will lack biological inhibitions, even biological form all together.

>>9709866
This is false. You don't need to replicate the brain if you can find an isomorphic function to represent it. That way you can create general artificial intelligence using a conventional microprocessor if it was powerful enough.

>>9708799
grrm, superintelligence? are you mistaken, friendo?

good thread.

Hubert Dreyfus

>>9709915

GRRM is the epitome of superintelligence friendo

>>9708083

This

>>9706947
Your diary, desu

>>9706947

Any Feminist theory book

every book about "art"

>>9711010
>>9711582
>>9711604
>>9711634

When did the shit posters get here?

>>9709866
People say true AI would lead to singularity - meaning there wouldn't be wildly different AI's if it were able to learn and update itself. Like us, it would only be defined by its imperfections and if it had an imperative to become perfect, it would end up the same every time.

To say we'll never create one is pretty short sighted.

What's more likely is that we'll be able to expand our own intelligence before we do.

>>9708584

If I were to only buy 3 out of all of these, which one should I buy?

>>9712984

None, read the bible instead.