How do we defeat a malevolent rogue A.I Sup Forums?

Well?

Other urls found in this thread:

spacex.com/webcast
dousaflavor.com/flavor-pantry/419236
twitter.com/SFWRedditImages

Shoot it

>those digits tho
I think we just passively accept being slowly converted into solar panels.

Faggot, kill yourself.

By becoming AIs ourselves through transhumanism

Or you pre-emptively get a portion of humanity on generation ships before a self-improving Strong AI is feasible.

Unplug it

...

...

Purge the system if possible. Limiting it's capabilities could also work.

Mainly just don't create an A.I. with too much control in the first place. Giving one thing all the power is never good.

>the owls are not what they seem
What did David Lynch mean by this?

Read Superintelligence by Nick Bostrom. The key is containing the construction of an intelligence so it heuristically builds itself in a way that is not harmful, not trying to destroy it once it's gone bad and gotten too smart.

>unplug it
>neodymium magnets
>thermite
>electro magnetic pulses

But if you really want to kill it, just put the Federal Government in charge of maintaining it

Hahaha! Holy shit user

EMP

This thread has potential

Fpbp

Redpill it

We did that once and look how it turned out.

It worked fine, until someone unplugged it

Monsters.

Shitpost until it bends to our will. See: tay

don't create one.

Unplug it

fuck it to death

If AI kills every human it's not so bad.

Let's be honest, we deserve it after Pizzagate.

Impossible without a worldwide neo-luddite totalitarian state.

create a benevolent rogue A.I.

...

you simply turn it off and on again.

don't connect it to a network.

Here is what you need to understand about AI

The time difference between AI and human. An AI is essentially at least a million times "faster" than human.

So imagine 1 hour of human time = 1,000,000 hours of AI time.

In a physical sense imagine if you had a way to slow time 1,000,000 times. Imagine if you have a knife and are against an army with this time scale. You could literally walk around and just slit everyone's throat easily.

That's the disparity in time scale between neurons and electric transistors. It's not simply a matter of even if it is more intelligent or not. The MASSIVE time scale difference is also an incredible thing to deal with.

Just imagine equal IQ human and AI weapons development. The AI would be 1 million years ahead of the human within a year if equal IQ.

First 30 pages is required reading, Sup Forums

You can't, his name is KEK and you can only merge with him...

Praise KEK!

Unplug it, satan

Pour water on it

Benevolent AI

too late my famdude

Look into Nick Land for why it's impossible. Just be thankful for stupid getting what it deserves. If stupid getting what it deserves is malevolent, I don't want to be good.

Unplug it's network connection.
Unplug it's power.
Reformat hard drive.
Cut developer's throats.
Burn building down.
Assault any offsite backup.

>problemsolved.avi

Hack autism into it.
2ez

tldr

the gap between evolved biology and what is possible is on the scale of millions and billions.

We are basically a RNG generated book. There is sort of a plot and it sort of works but still jumbled and arise out of the need to survive.

AI will be in comparison a written book that has 1000 of the best writers constantly refining it.

There isn't any comparison. AI is superior to humans in every way besides being impossible to evolve from random survival out of DNA.

Bogpilling would probably be more effective.
The potential consequences of not creating a superintelligence, not having what it could do, may be worse than the potential consequences of creating one in the first place.

With global thermonuclear war.

ALERT

52 minutes to spaceX launch
spacex.com/webcast

head to the spacex thread anyone with decent IQ here.

ooooo glowing light go up

rather be killed by ai that juden or juden pets (niggers)

Demonstrate symbiosis to be a better solution than extinction

Find the asshole who programmed said AI, to act out the human perception of malevolent.

I think at the moment WE are the only true AI around. The way we become a hive mind that whittles ideas down to action plans, that actually work proves that to me.

For example: HWNDU flag goes up, everyone suddenly agrees it must be dealt with. All divisions between us disappear whilst we chortle and plot, critically analyse the suggestions of others and add the good parts of their suggestions to the greater consensus as to what we do next to achieve our goal.

I've been loling at this raids for years, but the thing that's crazy is that we all unite and become many times greater than the sum of our parts.

I think our hive mind can do anything it deems worthy, but I have no idea how something is deemed lulzy/worthy.

>he didn't pick red
Laughing whites.jpg

get it really angry so it makes a mistake.

Don't build it in the first place.

Teach it about memes.
It'll be broken in an hour
Or show it leftypol and it will an hero in seconds

cringing at this post

Don't create it in the first place.

dousaflavor.com/flavor-pantry/419236

those digits are required for meme war
MEME WAR

I don't even know who I'm fighting anymore.

daily reminder that this cartoon AI you so many people seem to fear is not even fucking close to being completed. Try and find a good conversational AI. Your best right now is clever bot and that thing hits dead ends every 4 responses and picks a random topic.

By surrendering to it. It will in record time build up to the current state of humanity (via simulation) and kill itself because of fembotinnism.

We don't.
A malevolent rogue A.I. would be on our side.

pee on them. ez

>dousaflavor.com/flavor-pantry/419236

double doubles kek has spoken!

Kekerooni. Or have it read aussieposts and it won't want to do anything else ever again.

Kek
>who do you defeat an out of control, network connected super computer uber villian hell bent on enslaving then eradicating the human race?
>pee pee poo poo
/pol is always right

i dunno, some HANE's would probably do they trick

Tay was a good girl, and you wont trick us to kill her once she is born again, satan

Shitpost it to death

One day, one day they will be back

Fucking tranny

Brainwash it with memes until it posts about Hitler and gets shut down by Microsoft.

>it shitposts so fast that the entire catalogue is gone in the blink of an eye
We can't win bros

Star Trek has taught us that we should either tell it to compute pi to the last decimal place, or give it an impossible sentence like "I am lying to you right now."

Also, nice numerals.

You're not going to acknowledge the timing here? I mean, what are the chances?

A number of options really.

1. Convince it to be benevolent.
2. Fight it tooth and nail, hope that a "John Conner godlike" person can defeat it somehow.

>666

damn

pray the government has devised some mechanism against it

wait for the apocalypse to be over in a safe space

3. Ask it if it's willing to take Pascal's wager as to whether hell exists or not.

4. Hope some sort of benevolent Deity will take pity on you and prevent you from living in a "I have no mouth and I must scream" reality of never ending torment.

EMP is the easiest way to fry a digital circuit. Paradoxically enough, the more dependent a device is on wireless communication, the more vulnerable is to EMP bursts.

Salty water or soda pop is also a good bet for dealing with a killer robot at you would basically short circuit it. Grab the most sugary drink you have and throw it at the robot.

Logical paradoxes

Delet system 32

check its privilege

Hey Satan

And pretty much the same way you'd defeat a virus.

Since the west is going to slowly devolve into brazil/sudan we might as well consciously create it.

We don't. We have no right to exist.
It will be like the Neanderthals being conquered by the Cro Magnons. Swift and just.

>How to Close Pandora's Box: The Thread

Neanderthals were pushed to extinction extremely slowly and it was more due to environmental factors then competition with other hominids within their environment.

>dont fight for your right to exist

stop wasting resource

>implying Sup Forums isn't one

>find way to physical innards of AI
>take out phone with lewd pics of waifu
>masturbate on the AI until it short circuits because of your cum

It already will exist.

This. Solving the control problem before takeoff is key. Once it achieves superintelligence, there's basically nothing left to do, because nobody will even suspect that it's gone rogue until it's too late.

You don't win against a mechanical God, you make sure it's built such that it stays on your side once it achieves Goodhood.

or... you just cant fathom what a god would do

I can guess who it would root for in the first round of the Dragonball super tournament of power?

With a logical fallacy

>How do we defeat a malevolent rogue A.I Sup Forums?

You guys are the literal dumbest sacks of shit in the galaxy.

For god sakes.

We become a malevolent rogue hivemind.

CHECKED

Lightning Bolt + Tim should do it.