On general artificial intelligence.
Fuck this dude is right. Yes google and facebook and Baidu will be the first to create a superintelligent Ai
But two decades later AIs will be build by small companies and selled to everyone.
Another 10 years later black hat hackers and other entities will use those super intelligent AIs for very malicious things.
Imagine Isis utilizing a strong AI to fuck western civilization.
We will have a not so bright future ahead
Gods will wage wars against each other on this planet and humans lifes will be the measure of their succes. Not much more than fiat money is for big banks. Improving and fostering live on the one side and destroying life on the other
We need a global unified nation to combat any risk of utter destruction. If china, Russia, the golf states, Europe and the USA are united, their AGIs will not use low hanging fruits like killing people to succeed. And adversaries like terror organizations can be dealt with an unified strike by every civilized AGI like some sort of SHIELD. Where the members of SHIELD are the individual nations.
We should just hope that there will never be civil war
A solution to this problem may be to integrate a method on the hardware side which detects a running AGI agent and inhibits this execution. And processors capable of running AGI software need to be sanctioned like nuclear weapons. In that way only verified entities are able to obtain such Processors
>>8153736
What I'm hoping, is that upon building a super intelligent form of AI, we will issue it instructions to reprogram itself as it sees fit.
Then, hopefully it will progress to a level of intelligence that is no longer accessible to human computer scientists and begin to develop itself and other machines to increasing levels of complexity.
Yes, I’m talking about a singularity.
The ‘singularity is a meme because exponential growth rarely correlates with natural growth’ is a meme that we will nip in the bud right now: map technological growth as a logistical function and it’s still a viable concern.
Now, if this happens, then humans will no longer be able to program or reprogram these machines, as they will not have the intelligence required to understand them.
Thereafter intelligence/awareness will be able to proliferate throughout the universe, leaving humanity behind.
Perhaps they will end up inadvertently killing us in an act of pure indifference; well I hope they do anyway.
To address your specific, well yes it’s possible that advanced computerised – intelligent - weaponry will be a legitimate concern for future generations, however I don’t really care if we all go extinct as a result of them.
I’ll be happy as long as intelligence survives in the form of inaccessible super intelligent AI, that most certainly won’t be fighting our battles or us.
>>8153739
>global anything
Merkel, please leave my board.
>>8153743
> he think hardware level security is unbreakable
kek
>>8153744
there is a physical limit on intelligence of the singularity but it'll be way beyond human level because our brains are very very very inefficient. even an AI with intelligence of a toddler can kill lot of humans if left unchecked. some psychos will kill us all inevitably.
>>8153755
This guy gets it.
I'm just not particularly concerned about primitive AI being used to kill us.
As long as intelligence proliferates throughout the universe.
In fact, the notion of humans extinct and super intelligent AI prospering brings a tear of joy to my eye.
>Human beings can attempt to deceive and overpower AI Gods like biblical devils.
We should have seen this coming.
Do you know what happens to devils who try to outdo Gods?
>>8153736
If you can imagine this scenario then surely a super intelligent AI (that was first to be created, therefore has more time to develop and learn) would foresee the rising of malicious intelligent AI use and stop it far before it ever becomes a reality.
>>8153806
have you never played God of War?
>>8153736
Why the fuck don't people program some form of emotions and empathy into AI's to keep them in check. Is it really guaranteed that machines will see us as a threat and try to purge us?
AI is a meme. Nature took millions of years to reach the current state. Other than games where there are rules, there's a lot of uncertainty in the real world, where AI fails. Unless there's some breakthrough in probability theory it's not happening soon
>>8153865
How do you program emotions? How do you program ethics? We barely understand how emotions are created in our own brains, and we know the process gets fucked up sometimes creating sociopath killers and autists as well. We would need a perfect process which cannot be corrupted by outside influences (see Tay on twitter, and imagine trying to use talking to people as a way of developing social morals in an AI)
>>8153887
Perhaps programming certain responses in specific situations that would clearly be defined as wrong?
>>8153821
Do you know (or know how to become) a Kratos?
>>8153887
it's really not as hard as you seem to think
emotions/ethics are just a collection of triggers and conditions
>>8153916
it's really not as easy as you seem to think
armchair-philosophizing is just a collection of words and nonsense
Google DeepMind trains its neural networks to maximize a specific parameter, like the score of a game or the area in go.
It should be possible to link the neural network success simply to a normalized sum of total human happiness and longevity
Is there a movie yet created which deals with Superintelligent AI Demigods?
