Sep 7, 2023·edited Sep 7, 2023Liked by Gary Marcus
This is the world we live in, yes. It is good to know all this. The next question is what to do about it.
Absolute privacy does not exist, and likely it is not even desirable. People who will benefit most from it are criminals, child abusers, and so on.
We don't get a cut when our info is sold, but we do benefit from free services, including Twitter and Facebook, which may not be profitable if they can't monetize their knowledge in some way. As a rule, people prefer the loss of some privacy and exposure to ads that know too much about you, rather than be willing to pay Twitter a fee.
So, the sensible approach is to focus on preventing abuse of privacy. Either by the government or corporations.
The fact that they find all this in privacy statements is more or less the lawyers creating legal room, but that in itself doesn’t mean they are doing it. But the fact that the commercial car data hubs are booming (and what these leeches purport to be able to sell) suggests they are. I also wonder how this works in the EU where there are stricter rules. Both here and in GAI we are starting to see that Joseph Stalin was right when he said “quantity has its own quality”. And that is the league these surveillance capitalists are in.
Some more sketchy behavior in A.I. land is in mathematics and automated theorem provers. Read from Peter Woit's blog:
"A company called Prolific is advertising work paying 20-25 pounds/hour doing tasks in Lean 3. This company is in exactly the business described in the NY mag articles, hiring people to do tasks as part of “studies”, which often are generating AI training data.
One unusual thing about this whole industry is that if you sign up for this work you often have no idea who your employer really is, or what your work will be used for, and you sign a non-disclosure agreement to not discuss what you do."
When the car companies were building robots to help build cars, they bragged that it would make employee’s lives easier. They lost their jobs. AI is going to do the same thing, only on a huge scale. It’s even learning how to write poetry!
Due to the nature of the human mind, any human enterprise is typically going to include both the good, the bad, and the ugly. As the scale of power available to us grows, the bad and the ugly increasingly obtain the ability to erase the good.
The good is very real, it does exist. AI is likely to achieve constructive miracles in many arenas. But those constructive miracles depend for their existence upon a stable society, which becomes ever harder to maintain as the scale of power available to the bad increases.
To the degree the above is true, our focus should extend beyond particular technologies to our relationship with knowledge and power as a whole. It doesn't really matter if we make one technology safe if dozens of others are left running wild.
Currently, we seem to be making the mistake of confronting the bad properties of technology on a case by case basis, such as the privacy invasions of the car industry as referenced in this article. This case by case approach seems doomed to failure, because an accelerating knowledge explosion will likely present such challenges faster than we can meet them.
How much power do we intend to give our imperfect selves? Is there any limit to that? Are we assuming that we can safely manage ANY amount of power delivered at ANY rate?
If there is any hope of meeting the challenge presented by advanced technology through the processes of reason, such questions must be addressed.
Yes, surveillance is a different animal than "AI", which was always about mentoring and grooming the children.
Having lived a lot in china the past 40 years, I can say that 40 years ago there were mil/cops every where with guns ( same in south-korea ), and now past +10 years I go to places like Kunming and never see a 'cop' ( its all cameras now ), and there is NO ICE, its all EV motorcycles (scooters), which drive you crazy cuz they drive on sidewalks with lights off ( conserve power )
...
But like I said, replace PIGS on the street with "CAMS" is what it is, but this GREEN/RED kill ID thing that MUSK&Peter-Thiel sell to US-GOV to kill people all over the world called "Clearview" by Palantir ( pre-crime AI too ) is terrifying.
Thanks ... I dont think that the connection between AI and surveillance capitalism gets enough attention.
This is the world we live in, yes. It is good to know all this. The next question is what to do about it.
Absolute privacy does not exist, and likely it is not even desirable. People who will benefit most from it are criminals, child abusers, and so on.
We don't get a cut when our info is sold, but we do benefit from free services, including Twitter and Facebook, which may not be profitable if they can't monetize their knowledge in some way. As a rule, people prefer the loss of some privacy and exposure to ads that know too much about you, rather than be willing to pay Twitter a fee.
So, the sensible approach is to focus on preventing abuse of privacy. Either by the government or corporations.
An AI that feeds from data is the worst paradigm we ever had. And at the same time, this obfuscate the real things to be done for AGI.
The fact that they find all this in privacy statements is more or less the lawyers creating legal room, but that in itself doesn’t mean they are doing it. But the fact that the commercial car data hubs are booming (and what these leeches purport to be able to sell) suggests they are. I also wonder how this works in the EU where there are stricter rules. Both here and in GAI we are starting to see that Joseph Stalin was right when he said “quantity has its own quality”. And that is the league these surveillance capitalists are in.
I drive a 1991 Acura. It ain't collecting a thing :)
Hi Gary.
Some more sketchy behavior in A.I. land is in mathematics and automated theorem provers. Read from Peter Woit's blog:
"A company called Prolific is advertising work paying 20-25 pounds/hour doing tasks in Lean 3. This company is in exactly the business described in the NY mag articles, hiring people to do tasks as part of “studies”, which often are generating AI training data.
One unusual thing about this whole industry is that if you sign up for this work you often have no idea who your employer really is, or what your work will be used for, and you sign a non-disclosure agreement to not discuss what you do."
https://www.math.columbia.edu/~woit/wordpress/?p=13550
When the car companies were building robots to help build cars, they bragged that it would make employee’s lives easier. They lost their jobs. AI is going to do the same thing, only on a huge scale. It’s even learning how to write poetry!
Regulate 3rd party access to data. It covers privacy as well as ML safety (trademarks, accountability, disclosure, etc).
Due to the nature of the human mind, any human enterprise is typically going to include both the good, the bad, and the ugly. As the scale of power available to us grows, the bad and the ugly increasingly obtain the ability to erase the good.
The good is very real, it does exist. AI is likely to achieve constructive miracles in many arenas. But those constructive miracles depend for their existence upon a stable society, which becomes ever harder to maintain as the scale of power available to the bad increases.
To the degree the above is true, our focus should extend beyond particular technologies to our relationship with knowledge and power as a whole. It doesn't really matter if we make one technology safe if dozens of others are left running wild.
Currently, we seem to be making the mistake of confronting the bad properties of technology on a case by case basis, such as the privacy invasions of the car industry as referenced in this article. This case by case approach seems doomed to failure, because an accelerating knowledge explosion will likely present such challenges faster than we can meet them.
How much power do we intend to give our imperfect selves? Is there any limit to that? Are we assuming that we can safely manage ANY amount of power delivered at ANY rate?
If there is any hope of meeting the challenge presented by advanced technology through the processes of reason, such questions must be addressed.
Yes, surveillance is a different animal than "AI", which was always about mentoring and grooming the children.
Having lived a lot in china the past 40 years, I can say that 40 years ago there were mil/cops every where with guns ( same in south-korea ), and now past +10 years I go to places like Kunming and never see a 'cop' ( its all cameras now ), and there is NO ICE, its all EV motorcycles (scooters), which drive you crazy cuz they drive on sidewalks with lights off ( conserve power )
...
But like I said, replace PIGS on the street with "CAMS" is what it is, but this GREEN/RED kill ID thing that MUSK&Peter-Thiel sell to US-GOV to kill people all over the world called "Clearview" by Palantir ( pre-crime AI too ) is terrifying.