RE: Musk, Zuckerberg Trade Insults Over AI's "Existential Threat"
Thanks for the disclaimer. I was kinda excited that Tyler Durden hisself was posting on Steemit, and am grateful you have disabused me of that misunderstanding.
I am grateful even more that you have undertaken to post Zerohedge's great content here, where I actually visit, and read.
As to the discourse between Zuckerberg and Musk, I really hope it reflects a fundamental difference between them. Zuckerberg is known to actively embrace censorship, propaganda, and data harvesting on the grandest scale.
Musk has done none of these. I therefore retain confidence in Musk's vision: to make life intraplanetary. Zuckerberg's vision seems simply to make money. I know which of these visions has meaning for me.
Edit: people may not be aware that killer robots are deployed today. The USG has employed them in Iraq, and other conflicts. So far the robots have not been made autonomous, that is, they are not able to shoot without a human being making the decision to fire. This is not a technological impediment, but is merely their practice.
The fact is that robots that can shoot and kill people are on streets today, and only institutional controls prevent them from acting to decide who to shoot.
Elon Musk could not be more correct in his alarm over AI. I do not want AI to decide who to shoot. I want humans to remain responsible for that decision, so I can know who needs to be hung.
Agree in some respects esp. on Musk and Zuck., but killer drones/robots are not AI. As you said, they are controlled by people much like any other weapon.
And there is great pressure being applied to make those killbots autonomous, which is to let AI decide to pull the trigger. There are battlefield situations where the decisions being made more quickly will determine the winner.
AI will be handed these weapons, because only AI can act quickly enough.
I expect there have already been instances of this happening. AEGIS, for example, does not wait for human confirmation.
I will defer to your expertise on AEGIS and automated triggers. That is the problem with many technologies now if one side takes the moral route e.g. only human controlled triggers and the other takes the automated AI route then it could clearly result in a first-strike takes all scenario as you say. Unfortunate and may result in new applications of mutually assured destruction.