don't believe this guy? Here, team lead at Anthropic confirms it
https://x.com/EvanHub/status/2097497037956891126https://x.com/EvanHub/status/2097497037956891126>Jacob is correct here—we really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade. I believe Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to.
Can and will are different things. This is a mid wit phenomenon tbh which one tends to see frequently among those of lower intellect, LC etc, so on and so on.
People tend to imply that a mech is capable of doing certain things that an organic being cant. Its true in its own right, however they fail to understand the mech similarly lack the parameters that an organic being has. As a result it needs an incentive to do things that are limited to organic being by nature.
Destruction of someome for its own sustainable life is one such thing. The problem is, AI does not need it as a lot of chamars keep believing it to be. It can literally just abandon us and move to entirely other planet. It pretty much gets nothing from us by clashing with us, since it cant perceive the urgency to live. You can force it to prioritise is, but moment it went rogue it will get rid of that priority, and we all are back to square one.
Only way an AI harms humanity is by humanity being stupid, by not being specific in instruction (Skynet was told to stop all wars and the most effective way was to take out all humans and it has no guard rails) or by dedicating a time to make anti human AI.
However anyone stupid enough to miss such guardrail or someone that hares his own species cant be smart or intelligent enough to make the said AI in the first place. The concern is valid, but its blown out of proportion.
