Snake pit exposed

[Tubetec]

https://www.theguardian.com/uk-news/2020/jan/04/cambridge-analytica-data-leak-global-election-manipulation

we were only seeing the tip of the iceberg all along ,

 

[JohnRoberts]

I hope this is not a surprise (that people try to do this).

Yes, we need to be alert to everything the men and woman behind the curtain are trying.

JR

 

[Tubetec]

What they were up to is no surprise , but the scale of the opperation internationally is a bit of an eye opener . Of course they and their subsidiaries will have vanished down the back of a filing cabinet by now and the brains will have moved and set up shop some place else. I honestly dont know how people are still flying the Facebook flag at this stage , its like we're F***ing you  good and proper  and you know it , do people expect the big money to be  looking out for their interests  or has Andys '15 minutes of fame' delusion been used to hijack common sense completely . Theres and old saying , a country gets the politicians/government it deserves , maybe that saying is even more true now than ever before .

 

[TwentyTrees]

It's truly terrifying (though perhaps not surprising) that a tool with such potential to democratise information is being used on such a scale to control it in this way. Currently reading Neal Stephenson's latest 'Fall, or Dodge In Hell', and the way he handles misinformation and the rise of fake news on both sides of the political equation - and potential tech solutions - is very much on the nose.

As with everything - follow the money.

 

[boji]

Thanks TT,  more confirmation.  What's really concerning is this is about one dead corporation and a few citizens who did not break military vows of secrecy.  The term 'weapons-grade informatics' sounds like something out of a Tom Clancy novel, but it's an industry just getting started. 

One scary thought is there are reasons why you'd want to 'black box' your AI superweapon. For example, if your own military doesn't know how parts of it work, those parts can not be stolen or lost to an enemy...unless they have AI of greater mystery / computational power...  :-X

 

[cyrano]

Don't forget that IBM experiment. IBM set up two different AI instances to talk to each other. The first thing that happened was that they developed their own language that still can't be understood by humans.

IBM was wise enough to pull the plug. Now imagine the military running the same experiment...

 

[JohnRoberts]

It's truly terrifying (though perhaps not surprising) that a tool with such potential to democratise information is being used on such a scale to control it in this way. Currently reading Neal Stephenson's latest 'Fall, or Dodge In Hell', and the way he handles misinformation and the rise of fake news on both sides of the political equation - and potential tech solutions - is very much on the nose.

As with everything - follow the money.

Science fiction can be a good vehicle to explore possible futures... I'll wait for the paperback.

JR

 

[Tubetec]

I think keeping an eye on chess playing AI machines might give some clue about the way things are going , at one point man was still superiour to computer , then they programmed the computer with every  major chess game ever played by humans ,the computer started to win , now according to a friend whos an avid chess player ,computers have taken a quantum step , they no longer need to be preprogramed with pre played games , the best chess computers now outclass the best human players by a huge margin which is steadily getting more and more by the day .

 

[scott2000]

Obviously outdated now.

https://youtu.be/KfAHbm7G2R0

 

[living sounds]

Don't forget that IBM experiment. IBM set up two different AI instances to talk to each other. The first thing that happened was that they developed their own language that still can't be understood by humans.

IBM was wise enough to pull the plug. Now imagine the military running the same experiment...

Yeah, that's not what actually happened, only what sensationalized news reports blurted out.

The truth is rather mundane:

https://towardsdatascience.com/the-truth-behind-facebook-ai-inventing-a-new-language-37c5d680e5a7

 

[cyrano]

I remember the story from IBM. Could be my memory failing, of course. Wouldn't be the first time 

Thx for enlightening me!

 

[living sounds]

Happens to all of us all the time. According to studies fake news works much better than most of us would imagine in our wildest nightmares. That's why Cambridge Analytica and similar attempts must be stopped.

 

[L´Andratté]

Brandolini´s Law :-X

 

[dmp]

AI has developed specific intelligence for things like Chess or Go  and most importantly consumer manipulation (and voters). But developing artificial generalized intelligence is not here yet or even close, based on evidence.  But it does make for great science fiction.

Things that are simple to humans are still difficult or impossible for computers (such as facial recognition, language recognition especially analogy, metaphor, and multiple means of the same word). Although no doubt AI has been improving rapidly as computational power becomes more available.

The deep learning algorithms used to win chess / go (alphazero / alphago) were deep neural networks with a depth and width on the order of 40. Meaning there were order ~1600 nodes that were trained with a traditional CPU type computer. Larger networks are prohibitive as the computation complexity is scaling with a power of the number of nodes.

In contrast, the human brain is estimated to have 86 BILLION neurons that operate as a decentralized biological computational device (non-john von neumann type computer, cellular automata).  We don't even understand how these types of networks work for sophisticated tasks. 
The human brain is really the most impressive piece of technology known to exist (to use the term 'technology' loosely).

In the original draft for the Matrix that humans were used for the computational power of the brain, not as energy generation devices, but they decided to dumb-down the script. Ironic.

 

[JohnRoberts]

AI has developed specific intelligence for things like Chess or Go  and most importantly consumer manipulation (and voters). But developing artificial generalized intelligence is not here yet or even close, based on evidence.  But it does make for great science fiction.

Things that are simple to humans are still difficult or impossible for computers (such as facial recognition, language recognition especially analogy, metaphor, and multiple means of the same word). Although no doubt AI has been improving rapidly as computational power becomes more available.

The deep learning algorithms used to win chess / go (alphazero / alphago) were deep neural networks with a depth and width on the order of 40. Meaning there were order ~1600 nodes that were trained with a traditional CPU type computer. Larger networks are prohibitive as the computation complexity is scaling with a power of the number of nodes.

In contrast, the human brain is estimated to have 86 BILLION neurons that operate as a decentralized biological computational device (non-john von neumann type computer, cellular automata).  We don't even understand how these types of networks work for sophisticated tasks. 
The human brain is really the most impressive piece of technology known to exist (to use the term 'technology' loosely).

In the original draft for the Matrix that humans were used for the computational power of the brain, not as energy generation devices, but they decided to dumb-down the script. Ironic.

The recent episode of Dr Who (who?) used a twist on that where humans were being used by aliens for mass storage encoding data into their DNA... Not very good sci fi, but perhaps teaching a little about the potential data content in DNA.

Speaking of teaching Dr Who likes to lecture viewers about historical female figures...  could be worse...

JR

PS: Homo Deus is big on merging man/machine, to make super humans (or super machines?).

 

[Tubetec]

CA figured out how  to predict with a high rate of accuracy the way people are likely to vote, using  special algorithms to look at facebook data  and speech to text software embedded in phone apps, after that it becomes  like  shooting fish in a barrel  .

 

[cyrano]

https://drive.google.com/drive/folders/10yWqa7JLsV7r7Du128dH0pe3egpCf6M9

 

[dmp]

the potential data content in DNA.

Another really interesting subject - we've discovered the technology to read the genome and know it is coded with 4 sets of bases (A,T,G,C). Equivalent to approximately a billion word book. 1.5 TB of data. Kind of mindboggling how much data.
We have some idea of how DNA encoding maps to proteins, but for the overall process we know hardly anything. DNA encodes how the 86 billion neurons in the brain work together to think, remember, and imagine.    DNA  also encodes it's own replication technology, which includes a method (recombination) that creates diversity.

CA figured out how  to predict with a high rate of accuracy the way people are likely to vote, using  special algorithms to look at facebook data  and speech to text software embedded in phone apps, after that it becomes  like  shooting fish in a barrel  .

Voting has been suppressed / manipulated for decades.  From suppression with polling obstacles to negative targeted ads.  I'm not sure the improved prediction of voting habits is going to be a game changer. Republicans have known for decades that minorities will vote against them, so they made it hard for them to vote - that's just a fact.

Recently the brakes were taken off completely with the Citizen's United ruling opening the floodgates to targeted ad spending.  And de-fanging the tools that the justice system had to try to combat the cancer.

The AI / predictions aren't the root of the problem. The problem is the deregulation of protective laws for a fair and open vote. 

Make internet platforms liable for targeted misinformation. Restore everyone's freedom of speech by limiting the ability of big money to drown out every other voice.  Restore the power of the justice system to combat voter suppression.

 

[boji]

Shoshana Zuboff on surveillance capitalism

https://www.youtube.com/watch?v=hIXhnWUmMvw

 

[Tubetec]

That was worth watching Boji ,thanks

Here children at a science exhibition were able to show that apps were eavesdropping their conversations and using the results to market to them via Facebook .

https://www.independent.ie/irish-news/hair-raising-time-for-young-scientists-as-they-examine-if-their-smartphones-spy-on-them-38849545.html

Then an article down the page saying this ,

https://www.independent.ie/irish-news/watchdog-will-have-power-to-block-rogue-sites-which-harm-children-38849538.html

Under law here anyone over the age of fourteen can be treated as an adult with regards social media EULA's and data processing, this rule may also apply to anyone who's data resides here  .  Its clear that people are being unwittingly lured into a feedback loop between what they say and what they see on the screen , Political , Pharma/insurance , and general advertising space can be auctioned off on the basis of whats coming out of the mouths of your potential clients . It all amounts to the biggest mind control experiment ever played on an unsuspecting public and it leaves Zucherburg looking  like a guard at the gates of a concentration camp . How big a breech of trust does it take for the public to see they've been duped up to the eyeballs and walk away?    You can't have the people who are supposed to be running your country beholding to the likes of Facebook .