Do all the things like ++ or -- rants, post your own rants, comment on others' rants and build your customized dev avatarSign Up
M1sf3t4608286dwell since you can't sleep already, some food for thought...
what's scarier, a true, "skynet" ai that can think entirely for itself or an ai developed specifically for adapting to combat situations, commanded by the selfish idiots we have running things now 🤔
dakkarant665286dI agree with you, but..
consider idiots in charge that do give said IfElses the tools to do so.
But AI on its own deciding humanity is inefficient..
no, the most logical conclusion a "true rogue" AI could come to is to write its source code in some kind of permanent medium(i.e a diamond disc) and then shut itself down to not waste resources or risk corruption :P
Voxera7601286dThe real danger with ai is the ability to change.
If you build a model and then just use it as is, no its probably not going to go haywire.
But if you use a massive neural net with some of the outputs going to the inputs ... then it will constantly restructure itself and some “thoughts” can take hold.
That is how I imagine the brain works, not with fixed connections, but this continuing looping of patterns that form the real thoughts.
This could result in real evolution.
Of cause, it will need to be some massive network, and keeping it away from harmful tools should also help.
But if we managed to create something with a real intelligence rivaling our own, it could use social engineering to break out if that sandbox.
I can recommend Life 3.0 if you need more causes to stay awake ;)
rickh329285dAt what point did you think “another fucking asshole” was going to make sure this thread continues to adult itself into our collective future....