Anyone who has not heard of "models of social influence" (also known as "social dynamics") should educate themselves

Anyone who has not heard of "models of social influence" (also known as "social dynamics") should educate themselves.

Computer scientists since the 90s have been creating simulations in an attempt to model public opinion. For the most it's been a meme field with little results other than a few interesting ideas and some novel math.

For an overview on the field, check out this 2017 paper. jasss.soc.surrey.ac.uk/20/4/2.html

During the 2016 election google and friends were pouring massive amounts of resources into group opinion modelling in an attempt to predict and manipulate the election. At the time they weren't able to come up with anything substantial (other than some glorified google trends statistics, which also ended up being wrong).

Today however the story is very different.

Google built a computer capable of designing a computer that was capable of running their models. The meta-computer they called AutoML. The computer it designed has already been assembled in has been running for several months.

fastcodesign.com/90152427/googles-new-ai-designs-ai-better-than-humans

The story they give for their new supercomputer is that it's for image processing, which is bullshit. They say this for publicity purposes only since they are afraid if they couldn't explain its purpose autists would start to get nosey.

The truth is that their new supercomputer was created to run massive social dynamic models. And it has blown away all expectations. No it's not sentient. No it's not an AGI. But what it can do is simulate public opinion with a remarkable accuracy.

Other urls found in this thread:

youtu.be/0l20Pac9YWM
twitter.com/NSFWRedditVideo

Personality weightings are generated using a separate neural network. This NN uses all known data related to the agent as input. This includes: income, bank accounts, credit card purchases, youtube/google searches, education, criminal record, GPS location history, all sent/received emails, hundreds of photographs and videos, along DNA samples if available (including data gathered from airport body scanners).

They've been testing out their new tool in many ways. One that I know of is Bitcoin, which ever since it started to explode has been a massive source of concern among the (((the elite))).

They game is this: they pump out news articles for a short time (either for or against Bitcoin), predict what will happen with their model, then watch see how close the result matches.

What was found was that news articles are too small a lever to manipulate public opinion in a meaningful way. Either a catalyst is needed, aka a false flag, or a more subtle process they call blending, which I will get into in a later post.

I also know they've been testing the impact of UFOs and UFO sightings. I'm not sure if this is just to test the capabilities of the system and how conspiracies spread, or it's some kind of trial run to actually go public with the ayy lmaos.

Apparently access is severely limited. They person I spoke to claims to have worked on several related projects but said they were only given access in the context of specific problems.

First for the singularity

We need a civil war

>computers can predict the future
this isn't news

Is this stuff good for the Jews?

its only natural given they are a gigantic company that specialises in AI. this type of research is not surprising at all.

what I will be interested in learning is how they alter their actions based on the predictions. the most optimal course of action would be to go with the flow whenever it doesnt affect their bottom line, but what if the overton window overcompensates (as it usually does) by going too far right?

no

interesting theory, but I think you're jumping to conclusions

youtu.be/0l20Pac9YWM

I program these systems. They don't work as well as you think, because they don't have access to enough data. Say, for example, that AutoML has access to 95% of the world's sentiment. (which is very generous) even then, that unpredictable 5% is enough to undo just about every major policy decision made based on AutoML's determinations.

This is nothing new...Social dynamics, systems dynamics, all around since the 50's. Every idiot generation swears that computers are powerful enough to reach singularity in their generation. There's a trick.

Humans calculate in 5 states, minimum, even the dummies. Computers have 2 states. It's not even a fucking comparison. You might wanna curb your enthusiasm with this AI garbage.

Humans with enough information could predict the future. Perhaps it's too complex now - perhaps it isn't.

It's not. I can predict the future to about a week.

same
Soros will die within the next week

>I program these systems.
then you wouldnt say:
>They don't work as well as you think, because they don't have access to enough data.
they dont work because they struggle with local optima

GOP should get good with google then

Rothschild dies tomorrow

If kek is with us. Yes.

checked

...

>then you wouldnt say
Snownigger what in the actual fuck would you know about this garbage? Local optima is the prime problem. It's the only open problem. The math and computation power have been trivial since the 70s.

>I program these things

No you don't. By your post you prove you don't even understand the fundementals of a neural network.

I notice you didn't elaborate in any way.
>muh fundamentals
Just admit AI is delusional.

>built a computer capable of designing a computer

You sound fucking retarded. I'm pretty sure all they did was build a genetic algorithm that reconfigures layers in a tensorflow model. It's conceptually super simple, but difficult to execute without a program that burns compute time with no results. It's definitely less impressive than you think.

(this sounds cool though)

>Humans calculate in 5 states, minimum, even the dummies. Computers have 2 states. It's not even a fucking comparison.

Is this post copypasta or decent trolling?

Please, poland, elaborate on how you deduced this. Was it with your fancy neural network, or are you speaking far above your station?
You ever play with wetware, leaf?

No why

computation in >3 states is why. Magical stuff. God's creations are truly great.