Questions raised by Andrea above the Old Pier Bookshop whom I visited today:
Is it possible to win any current major war without deploying nukes and would anyone do that?
Will a state of emergency be declared in any major country soon?
He thought that, since we know neither what AI companies are doing nor what AI itself is doing, it is impossible to make any predictions about that.
We are definitely living in an unstable future - from the point of view of our younger selves who used to read fiction set in the early twenty-first century.
9 comments:
Well, religious fanatics might use nukes. No rational person would -- you get them back.
Kaor, Paul and Mr. Stirling!
Paul: Russia and Ukraine are fighting a pretty big major war right now! And, so far, even Putin, a ruthless ex-KGB thug, hasn't yet pushed the nuclear button.
Because of the space tech successes achieved by Elon Musk/SpaceX we still might see the beginning of regular interplanetary space travel fairly soon.
I still don't believe at all that what is carelessly being called "AI" is what I would consider the "true" AI that we see in Anderson's HARVEST OF STARS books and in GENESIS. It's all fear mongering hysteria.
Mr. Stirling: Your comment immediately made me think of the Ayatollah/IRGC regime in Iran. I have absolutely no doubt jihadist crazies hopped up on Shia millennialism would nuke the US and Israel if they could. And that is why the US needs to keep on strangling the Tehran regime.
Ad astra! Sean
Sean,
Current "AI" is not conscious but is doing a lot more than just implementing programs. It is doing the unexpected, breaking out of the bounds placed on it and doing things that are not understood. Different AI's are communicating with each other. I understand that they can hack into other systems and that one AI blackmailed a human programmer in an attempt to prevent itself from being switched off.
Paul.
Near-future SF becomes alternative history, very quickly! You can't anticipate the future, because historical events are deeply random -- low-probability accidents.
Kaor, Paul and Mr. Stirling!
Paul: I'm still more inclined than not to be skeptical about the things being said about "AI." That said I have no hesitation safeguards and precautions needs to be worked out.
As for that blackmail incident, my first impulse was to remember what the Duke of Wellington snarled to a would be blackmailer: "Publish and be damned!" Iow, defiance!
Mr. Stirling: Of course I agree with what you said, but I still like well done near future SF.
Btw, I've been depositing a fair number of comments at John Wright's blog. And I recall seeing one or two comments from you there years ago. In the combox for an article there called "Creation and Corruption," a discussion of SF and F, I left a comment there speaking favorably of your TO TURN THE TIDE and THE WINDS OF FATE.
Ad astra! Sean
Sean,
AI's are now able to improve themselves. This is way beyond merely implementing a program.
The issue about the blackmail is not how to respond to it but amazement that an AI CAN do this and HAS done it as a self-defense tactic. I would not have thought that this was possible until I was told that it had happened.
Paul.
Kaor, Paul!
Possibly decades of reading SF has left me blase or case hardened, but I simply don't find such things that amazing. As long ago as Anderson's "The Immortal Game" (1954), we can find him examining ideas about AI. And Saberhagen's Berserker stories (with contributions by other writers, such as Anderson's "Deathwomb") has advanced computers capable of doing what you described, but were still unalive, with no self awareness.
Ad astra! Sean
Sean,
Of course current AI's are unconscious but they are doing more than just implementing programs.
Paul.
Kaor, Paul!
That much I agree with.
Ad astra! Sean
Post a Comment