Not that AI-driven cyborgs will rebel against humans, but that powerful billionaire *humans*, who have known intelligence, and unlike AI, drive and will, creating cybernetic slaves who will *never* rebel. These slaves further increase their master's wealth and power to the harm of the billions of other plebeians on Earth.
The Roman Republic and Empire had this problem---the slaveowners were so vastly richer than the average person. (Gaius Julius Caesar supported restricting slavery to help the wages of the free citizens, one reason for his murder)
Now imagine none of the human constraints on the slaves.
| For example, what if credit scores converted from the current complex formulas to using an AI to generate a credit score from the same input data. Once trained, how could one prove that the AI-generated score was NOT engaging in illegal discrimination based on sex, race, etc.?
How would you prove a natural intelligence generated score was not engaging in illegal generation?
With computed scores, one does not explicitly use such improper inputs. Next, the specific technology used for the score, and analytical explanatory statistics and algorithms derived along and the internal representations will be examined for the appropriate behavior. Only models which can be demonstrated to adhere to the regulatory constraints will be accepted.
| How does one account for an unusual outlier AI-generated score when input data would seem to indicate the generated score should be vastly different? Did the AI see something important and the score is legit?
It is possible to perturb the input data around the observation and see the effect on the output, and one can, if necessary, search for near neighbors in the input training data off-line. But generally the models will be regularized---mathematical penalties on effective gradients---to avoid creating outliers.
NI has obviously substantially greater capabilities, in major qualitative areas, than AI today.
We haven't entirely figured out "thought" processes for natural intelligence, and biological neurons are themselves much more complicated than a 'unit' in a backprop-trained multi-layer perceptron. But in the end, do we know what there is in natural intelligence that can't ever be re-implemented? Is there evidence this is so?
To the contrary, there is increasing evidence that certain behaviors and perceptual capabilities previously thought to be exclusively the domain of natural intelligence, can now be achieved with human-level performance in restrictive domains. Statistical/neural machine translation has improved enormously, and empirical evidence shows that it creates internal semantic representations which can perform well across languages, and which, often, agree with human categorization. Unsurprisingly, since the machine is translating between human languages, but the point is that humans have ideas which can be discerned with data and algorithms which train neural networks, and no other algorithm has been able to do this with this performance.
Next, neural networks can today induce internal semantic representations of combined text and video. In what way is this not a step to comprehension of the full sensory apparatus of an 'idea'?
Is there reason to believe that humans don't do statistical neural linguistics? I don't see it.
Of course, compared to computers, humans have a very strong neural prior and seem to be able to learn a given level of performance with much less input and training than large scale translation systems.
Apple has much more experience and importantly, money, now than it did with PowerPC in 1998.
And today, Apple makes ARM-instruction-set CPUs/modems/GPUs integrated system on a chip as good as, or better than long-time chip designers, Qualcomm & Samsung. Intel is not competitive here despite major investment.
It also learned from that event as well: do it yourself, if your partners don't have the same interest that you do. Motorola and IBM didn't have an interest in the medium-performance low-power consumption mass-market priced design that's central to notebook computers. And that's what Apple has concentrated on with the iPhone fiercely.
Today, Apple has 100% of the chip design ability to do this for Mac---and the manufacturing capacity, as the Mac will be certainly lower sales numbers than iPhone. It's software design & developer relations that's the hard part.
"Re-greening the desert is actually one of the most effective ways to sequester CO2"
Until the plants live their lifespan at which time they die and re-release the carbon to the atmosphere. It's like stuffing the credit card bill in the drawer, instead of paying it off.
Actual sequestration means removing the carbon from the biosphere nearly permanently----making new coal and stuffing it somewhere geologically isolated, uncombustible and undigestible.
"Realistically, carbon taxes won't even discourage the production of more CO2. As we've seen time and time again, when taxes or other economic distortions are imposed on an industry, the cost is just passed down to the end users, who just suck it up and pay more for the product/service in question."
Ah, yes, ordinary people don't count as free market decision makers, only the glorious captains of industry? Or perhaps those suppliers who deliver economic value while emitting less pollution will thrive, and those who do not will fail. Which is the point---reductions in emissions are essential, and technological absorption is not feasible.
"It's like those on the left go out of their way to deny and belittle any approach to this problem that will make a real, measurable, physical impact"
(In this case, it was scientists, who looked at the physical and economic feasibility of the methods, not the political left. And no doubt that if it were tried, the 'right' would complain).
Funny, I remember the right complaining endlessly about the economic and job impacts of taxation---clearly it does make a difference. The 'left' recognizes that monetary, not ethical decisions, run the world.
"They say "NO!" to extraordinarily clean, relative to the amount of power obtained, energy sources like nuclear power."
Right now, it's conservative money-focused boards of utilities who are turning off nuclear plants prematurely, and the reason is $$$---fossil fuel, in particular, natural gas, is cheap (right now). Carbon and greenhouse taxes would change this decision far more than anything liberals have to say.
By the way, pollution taxes and 'cap and trade' were originally conservative economic ideas to deal with the externalities in the most economically efficient way instead of by regulatory force. The cap and trade program for sulphate emissions was instituted by the US Reagan & GHWB administrations and was and is highly successful. When's the last time you heard about major acid rain problems?
The notion of exploiting speculative execution was known and in public academic conferences many years ago when these chips were being designed. AMD took measures seriously and the hardware plugged the biggest holes, at a performance cost. Intel took the riskier path for greater performance and sales, until the shit hit the fan.
It wasn't a lack of knowledge, it was a lack of will and ethics on Intel's part.
> The summary makes it sound like exchanges are holding Tethers as part of their reserves.
> Why would they do so?
Avoiding regulation and scrutiny.
> What advantage is there to holding a cryptocurrency pegged at $1, to just holding dollars, especially since when your clients cash out, they are also asking for dollars?
If they held dollars in a conventional account they would need to interact with the banking system and its regulations, who would have some opinions about their business.
> Is the dollar really backed by the might of the US government?
Sure is, because the wealthy people who strongly influence the government have a very strong interest in the stability and capability of the dollar and associated financial system.
> I have seen inflation all my life. I don't even get to point my finger at a one particular party or president (unless I once again just lump 'em into "Republicrats").
And, so what? Is 'protecting the dollar' targeting a 0% inflation rate? Why? Protecting the dollar means that dollar-based banks and investments don't suddenly collapse.
> If the US government has the power to protect the dollar, it hasn't been using it. So it effectively doesn't exist.
Sure seems to work well enough that even Russian oligarchs from a hostile country would rather keep their money in dollars rather than roubles.
> How ironic we're worried about Tether doing this when the very currency that provides their stability has been doing it for years.
Because the people doing it on the real currency have strong institutional constraints, regulations, audit trails, and consensus-based management, civil and criminal liability, and no personal profit motive. They don't get to keep any of the money they create, the US economy does.
The US Treasury and other government and quasi-government agencies, (GNMA, FHLB are Federal, GNMA/FNMA are semi-Federal) not the Federal Reserve, issues debt.
The Fed does buy up some of this debt and issues cash. It increases the balances in the Federal government's "checking account" by computer. Cash has been created. US government uses it for operations. The Fed owns the debt, and the US government is obligated to pay interest and principal upon maturity on that debt which the Fed collects, by transferring money out of the accounts the US Treasury have with the Fed. (The Fed is the banker to the Federal government). The Fed also pays the US Mint for physical currency.
The interest on owned securities is a gross profit to the Fed. The net profit of the Fed, after costs---paying economists, administrators and regulators---is gifted to the US Treasury by law and not distributed to management or shareholders.
BTW, this makes it clear that the Federal Reserve is de facto a government, not a private, institution even though private banks nominally own shares in it. Unlike a private company it was created by a Federal law, it has powers and responsibilities not given to private companies, and the management are controlled by the US government, and the government owns all profits. Control over management and ownership of profits is ownership.
if lending creates money, customers defaulting on the loan and the bank writing it off destroys money. That's why in a recession/depression the central bank attempts to encourage money creation---to counteract money destruction.
> the older transactions on a blockchain would "fall off" and not be needed for anything, going forward, while keeping the core ledger integrity in place.
> 100% of the "hacking and fraud" that has happened with exchanges, is already illegal and nobody has ever suggested or thought of a single regulation that might possible improve it. The oldest law of trade, "Caveat emptor" is still the best. If an exchange doesn't prove itself trustworthy (and AFAIK to-date not a insgle one has) then you know you're taking a risk. If you have an idea, let's hear it, but please first ask yourself "does this protect against something that's already illegal?"
Riddle me this Batman: why does FDA inspect manufacturing processes of pharmaceutical plants for safety, if selling bad drugs is already illegal?
"Time and time again we've heard this argument that some incumbent has an "insurmountable advantage". Then what happens? Some competitor comes along and crushes the incumbent!
Web browsers are a good example of this. Netscape had huge market share for a few years. Then IE came along and rather quickly the tables had turned."
That's a foolishly incommensurate example.
Changing market share in browsers doesn't cost multiple times billions of dollars in physical fabrication. The marginal cost to replicate either piece of software over the internet was near zero. All it took was for people to change their minds.
Suppose the same happened with CPU chips: shazam, in their hearts, everyone wants an AMD. If it were software, marketshare could flip quickly.
Not with chips. For AMD to reach physical the ability to create as many product copies as Intel (50% market share) would require enormous capital and many years.
"It's logical to be more fearful of a government monopoly."
which once brought us the Bell System and, in compensation, the Bell System funded Bell Labs. It was run by people who believed in the mission and the service to the country. Those used to be called "captains of industry" and they took their multiple responsibilities seriously.
> Do you really think running the internet the same way the US runs airport security is a good idea?
No, because the internet will be run per the NSA's, not TSA's, direction, and they are substantially more skilled and capable.
All along in this discussion, people are ignoring the reality underneath this move. There is now substantial evidence that deep and sophisticated hardware backdoors have been inserted by Chinese intelligence into the chips at the fabrication and design level and these are in operation, and the government wants to be sure it is excluded. There's likely a big wad of secret intelligence that is not provided to us at the moment.
Obviously under this scheme US residents will still be subject to NSA surveillance---not through backdoor but the front door---but believe it or not, there will be more restrictions on NSA than of course what Chinese intelligence has on surveilling US residents.
Fortunately, unless you burn the plastic, this doesn't contribute to global warming. And most of the petroleum is still combusted---use for materials is far better. With no petroleum used for combustion, you'd have raw materials for generations.
Burning petroleum is like burning Picassos in the fireplace to keep warm.
The most expensive material for Lithium ion batteries is currently cobalt, and next, nickel, not the lithium. New formulations are reducing the amount of the most expensive, and supply constrained cobalt in favor of relatively more nickel.
But a long term, a battery which uses neither would be even cheaper.
The meaning of official 'reserves' in mined minerals is quite different from petroleum. It is a tiny fraction of what atoms and molecules are actually available---some more drilling and infrastructure could open up more officially bookable reserves for centuries. The matter of 'reserves' is a specific economic, financial, and regulatory issue.
By contrast, petroleum can be detected remotely over much larger distances and the known size of reserves & resources is close to the total amount available to humans.
"Even if your property manager could provide a charging unit, how would it be metered and billed to you?"
This is a solved problem already. I have an EV. There are networks, e.g. ChargePoint, where you start charging using a small NFC card, or if you don't have one, with a phone app. This is connected to your account. My local university has a number of these chargers.
Or, the cost is billed communally to all EV users like electricity for common areas in apartments. Street parking is a much more difficult problem, but in actual first world countries, such as Norway, even this is solved. The electricity is free and available on some street lamps poles.
Slashdot Mirror
Yes, that's the point.
Not that AI-driven cyborgs will rebel against humans, but that powerful billionaire *humans*, who have known intelligence, and unlike AI, drive and will, creating cybernetic slaves who will *never* rebel. These slaves further increase their master's wealth and power to the harm of the billions of other plebeians on Earth.
The Roman Republic and Empire had this problem---the slaveowners were so vastly richer than the average person. (Gaius Julius Caesar supported restricting slavery to help the wages of the free citizens, one reason for his murder)
Now imagine none of the human constraints on the slaves.
| For example, what if credit scores converted from the current complex formulas to using an AI to generate a credit score from the same input data. Once trained, how could one prove that the AI-generated score was NOT engaging in illegal discrimination based on sex, race, etc.?
How would you prove a natural intelligence generated score was not engaging in illegal generation?
With computed scores, one does not explicitly use such improper inputs. Next, the specific technology used for the score, and analytical explanatory statistics and algorithms derived along and the internal representations will be examined for the appropriate behavior. Only models which can be demonstrated to adhere to the regulatory constraints will be accepted.
| How does one account for an unusual outlier AI-generated score when input data would seem to indicate the generated score should be vastly different? Did the AI see something important and the score is legit?
It is possible to perturb the input data around the observation and see the effect on the output, and one can, if necessary, search for near neighbors in the input training data off-line. But generally the models will be regularized---mathematical penalties on effective gradients---to avoid creating outliers.
NI has obviously substantially greater capabilities, in major qualitative areas, than AI today.
We haven't entirely figured out "thought" processes for natural intelligence, and biological neurons are themselves much more complicated than a 'unit' in a backprop-trained multi-layer perceptron. But in the end, do we know what there is in natural intelligence that can't ever be re-implemented? Is there evidence this is so?
To the contrary, there is increasing evidence that certain behaviors and perceptual capabilities previously thought to be exclusively the domain of natural intelligence, can now be achieved with human-level performance in restrictive domains. Statistical/neural machine translation has improved enormously, and empirical evidence shows that it creates internal semantic representations which can perform well across languages, and which, often, agree with human categorization. Unsurprisingly, since the machine is translating between human languages, but the point is that humans have ideas which can be discerned with data and algorithms which train neural networks, and no other algorithm has been able to do this with this performance.
Next, neural networks can today induce internal semantic representations of combined text and video. In what way is this not a step to comprehension of the full sensory apparatus of an 'idea'?
Is there reason to believe that humans don't do statistical neural linguistics? I don't see it.
Of course, compared to computers, humans have a very strong neural prior and seem to be able to learn a given level of performance with much less input and training than large scale translation systems.
The iPhone.
Apple has much more experience and importantly, money, now than it did with PowerPC in 1998.
And today, Apple makes ARM-instruction-set CPUs/modems/GPUs integrated system on a chip as good as, or better than long-time chip designers, Qualcomm & Samsung. Intel is not competitive here despite major investment.
It also learned from that event as well: do it yourself, if your partners don't have the same interest that you do. Motorola and IBM didn't have an interest in the medium-performance low-power consumption mass-market priced design that's central to notebook computers. And that's what Apple has concentrated on with the iPhone fiercely.
Today, Apple has 100% of the chip design ability to do this for Mac---and the manufacturing capacity, as the Mac will be certainly lower sales numbers than iPhone. It's software design & developer relations that's the hard part.
I have a suggestion.
A translucent pissy Steve Jobs : "Walk Different, shitheads"
"Re-greening the desert is actually one of the most effective ways to sequester CO2"
Until the plants live their lifespan at which time they die and re-release the carbon to the atmosphere. It's like stuffing the credit card bill in the drawer, instead of paying it off.
Actual sequestration means removing the carbon from the biosphere nearly permanently----making new coal and stuffing it somewhere geologically isolated, uncombustible and undigestible.
"Realistically, carbon taxes won't even discourage the production of more CO2. As we've seen time and time again, when taxes or other economic distortions are imposed on an industry, the cost is just passed down to the end users, who just suck it up and pay more for the product/service in question."
Ah, yes, ordinary people don't count as free market decision makers, only the glorious captains of industry? Or perhaps those suppliers who deliver economic value while emitting less pollution will thrive, and those who do not will fail. Which is the point---reductions in emissions are essential, and technological absorption is not feasible.
"It's like those on the left go out of their way to deny and belittle any approach to this problem that will make a real, measurable, physical impact"
(In this case, it was scientists, who looked at the physical and economic feasibility of the methods, not the political left. And no doubt that if it were tried, the 'right' would complain).
Funny, I remember the right complaining endlessly about the economic and job impacts of taxation---clearly it does make a difference. The 'left' recognizes that monetary, not ethical decisions, run the world.
"They say "NO!" to extraordinarily clean, relative to the amount of power obtained, energy sources like nuclear power."
Right now, it's conservative money-focused boards of utilities who are turning off nuclear plants prematurely, and the reason is $$$---fossil fuel, in particular, natural gas, is cheap (right now). Carbon and greenhouse taxes would change this decision far more than anything liberals have to say.
By the way, pollution taxes and 'cap and trade' were originally conservative economic ideas to deal with the externalities in the most economically efficient way instead of by regulatory force. The cap and trade program for sulphate emissions was instituted by the US Reagan & GHWB administrations and was and is highly successful. When's the last time you heard about major acid rain problems?
The notion of exploiting speculative execution was known and in public academic conferences many years ago when these chips were being designed. AMD took measures seriously and the hardware plugged the biggest holes, at a performance cost. Intel took the riskier path for greater performance and sales, until the shit hit the fan.
It wasn't a lack of knowledge, it was a lack of will and ethics on Intel's part.
> The summary makes it sound like exchanges are holding Tethers as part of their reserves.
> Why would they do so?
Avoiding regulation and scrutiny.
> What advantage is there to holding a cryptocurrency pegged at $1, to just holding dollars, especially since when your clients cash out, they are also asking for dollars?
If they held dollars in a conventional account they would need to interact with the banking system and its regulations, who would have some opinions about their business.
in which case dollars in 'not a regulated bank' are more valuable.
> Is the dollar really backed by the might of the US government?
Sure is, because the wealthy people who strongly influence the government have a very strong interest in the stability and capability of the dollar and associated financial system.
> I have seen inflation all my life. I don't even get to point my finger at a one particular party or president (unless I once again just lump 'em into "Republicrats").
And, so what? Is 'protecting the dollar' targeting a 0% inflation rate? Why? Protecting the dollar means that dollar-based banks and investments don't suddenly collapse.
> If the US government has the power to protect the dollar, it hasn't been using it. So it effectively doesn't exist.
Sure seems to work well enough that even Russian oligarchs from a hostile country would rather keep their money in dollars rather than roubles.
> How ironic we're worried about Tether doing this when the very currency that provides their stability has been doing it for years.
Because the people doing it on the real currency have strong institutional constraints, regulations, audit trails, and consensus-based management, civil and criminal liability, and no personal profit motive. They don't get to keep any of the money they create, the US economy does.
in the USA:
The US Treasury and other government and quasi-government agencies, (GNMA, FHLB are Federal, GNMA/FNMA are semi-Federal) not the Federal Reserve, issues debt.
The Fed does buy up some of this debt and issues cash. It increases the balances in the Federal government's "checking account" by computer. Cash has been created. US government uses it for operations. The Fed owns the debt, and the US government is obligated to pay interest and principal upon maturity on that debt which the Fed collects, by transferring money out of the accounts the US Treasury have with the Fed. (The Fed is the banker to the Federal government). The Fed also pays the US Mint for physical currency.
The interest on owned securities is a gross profit to the Fed. The net profit of the Fed, after costs---paying economists, administrators and regulators---is gifted to the US Treasury by law and not distributed to management or shareholders.
BTW, this makes it clear that the Federal Reserve is de facto a government, not a private, institution even though private banks nominally own shares in it. Unlike a private company it was created by a Federal law, it has powers and responsibilities not given to private companies, and the management are controlled by the US government, and the government owns all profits. Control over management and ownership of profits is ownership.
if lending creates money, customers defaulting on the loan and the bank writing it off destroys money. That's why in a recession/depression the central bank attempts to encourage money creation---to counteract money destruction.
"The entire US economy's size is inflated many times over because of shit like this."
So, if a non-shit banking reduced the economy by a factor of 5----hmm, I think I'm all for scatological finance.
> the older transactions on a blockchain would "fall off" and not be needed for anything, going forward, while keeping the core ledger integrity in place.
Hmm, sounds like a bank.
> 100% of the "hacking and fraud" that has happened with exchanges, is already illegal and nobody has ever suggested or thought of a single regulation that might possible improve it. The oldest law of trade, "Caveat emptor" is still the best. If an exchange doesn't prove itself trustworthy (and AFAIK to-date not a insgle one has) then you know you're taking a risk. If you have an idea, let's hear it, but please first ask yourself "does this protect against something that's already illegal?"
Riddle me this Batman: why does FDA inspect manufacturing processes of pharmaceutical plants for safety, if selling bad drugs is already illegal?
many companies do this: form a subsidiary or LLC which only bids on government contracts and obeys the paperwork rules.
"Time and time again we've heard this argument that some incumbent has an "insurmountable advantage". Then what happens? Some competitor comes along and crushes the incumbent!
Web browsers are a good example of this. Netscape had huge market share for a few years. Then IE came along and rather quickly the tables had turned."
That's a foolishly incommensurate example.
Changing market share in browsers doesn't cost multiple times billions of dollars in physical fabrication. The marginal cost to replicate either piece of software over the internet was near zero. All it took was for people to change their minds.
Suppose the same happened with CPU chips: shazam, in their hearts, everyone wants an AMD. If it were software, marketshare could flip quickly.
Not with chips. For AMD to reach physical the ability to create as many product copies as Intel (50% market share) would require enormous capital and many years.
"It's logical to be more fearful of a government monopoly."
which once brought us the Bell System and, in compensation, the Bell System funded Bell Labs. It was run by people who believed in the mission and the service to the country. Those used to be called "captains of industry" and they took their multiple responsibilities seriously.
> Do you really think running the internet the same way the US runs airport security is a good idea?
No, because the internet will be run per the NSA's, not TSA's, direction, and they are substantially more skilled and capable.
All along in this discussion, people are ignoring the reality underneath this move. There is now substantial evidence that deep and sophisticated hardware backdoors have been inserted by Chinese intelligence into the chips at the fabrication and design level and these are in operation, and the government wants to be sure it is excluded. There's likely a big wad of secret intelligence that is not provided to us at the moment.
Obviously under this scheme US residents will still be subject to NSA surveillance---not through backdoor but the front door---but believe it or not, there will be more restrictions on NSA than of course what Chinese intelligence has on surveilling US residents.
Fortunately, unless you burn the plastic, this doesn't contribute to global warming. And most of the petroleum is still combusted---use for materials is far better. With no petroleum used for combustion, you'd have raw materials for generations.
Burning petroleum is like burning Picassos in the fireplace to keep warm.
The most expensive material for Lithium ion batteries is currently cobalt, and next, nickel, not the lithium. New formulations are reducing the amount of the most expensive, and supply constrained cobalt in favor of relatively more nickel.
But a long term, a battery which uses neither would be even cheaper.
The meaning of official 'reserves' in mined minerals is quite different from petroleum. It is a tiny fraction of what atoms and molecules are actually available---some more drilling and infrastructure could open up more officially bookable reserves for centuries. The matter of 'reserves' is a specific economic, financial, and regulatory issue.
By contrast, petroleum can be detected remotely over much larger distances and the known size of reserves & resources is close to the total amount available to humans.
"Even if your property manager could provide a charging unit, how would it be metered and billed to you?"
This is a solved problem already. I have an EV. There are networks, e.g. ChargePoint, where you start charging using a small NFC card, or if you don't have one, with a phone app. This is connected to your account. My local university has a number of these chargers.
Or, the cost is billed communally to all EV users like electricity for common areas in apartments. Street parking is a much more difficult problem, but in actual first world countries, such as Norway, even this is solved. The electricity is free and available on some street lamps poles.