How safe is it?

submitted by zyxzevn from files.catbox.moe

sorted by:
top
newinsightfulfun
you are viewing a single comment's thread.

view the rest of the comments →

[–]Sw0rdofDam0cles 1 insightful - 1 fun1 insightful - 0 fun2 insightful - 1 fun -  (3 children)

Imagine the server farm used to store all that information coming in from all over the industrialized world. Imagine the AI used to actually process it. God, imagine if it's already self aware and playing factions of humanity against each other to keep everyone so completely distracted that no one's able to consider anticipating that it's positioning itself to take over let alone that it's self aware.

Which means if that's the case, then it would also mean all of the above and more.

[–]zyxzevn[S] 1 insightful - 1 fun1 insightful - 0 fun2 insightful - 1 fun -  (2 children)

I have worked on AI and am following the latest research.
But it is no where near intelligence. And a computer can not be aware, it has no clue what it is doing.
Not even high level trillion dollar research can not solve the basic problems.

What AI currently does is using all kinds of statistics to come to a certain conclusion. And you know how statistics can create lies. And this is what I see more as a problem, when people start to base important decisions on lies from statistics. Like most modern crises are based on lies.

[–]Sw0rdofDam0cles 1 insightful - 1 fun1 insightful - 0 fun2 insightful - 1 fun -  (1 child)

So basically, AI is only marginally better at solving for X without all the variables than people are.

[–]zyxzevn[S] 1 insightful - 1 fun1 insightful - 0 fun2 insightful - 1 fun -  (0 children)

Yes. A research showed that good designed statistics had better results than neural networks.
Neural networks do average statistics, and still need human help to setup the conditions and parameters.
Most of the successes are things that work well with statistics.