Then I'll just blacklist all IP's outside of the U.S.
The poster said they'd move offshore, not that they'd send their spam from offshore. It's easy enough for a non-American to buy a large number of throwaway accounts at US ISPs.
If the bill will have no effect then why waste important senate time with it.
At least this will get those who spend their time trying to get an anti-spam law passed to stop wasting their time. At least those of us who believe that spam must be solved by technical means can point to the failure of the legislative solution to back up our claim.
What he's talking about is a host record with the registry, which is the structure that stores information about authoritative name servers for a domain.
Yeah, I know what a glue record is. I've set them up before, and there was no need for reverse DNS to be set up.
Some registrars require reverse DNS for these and some don't.
Therefore you "don't have to have reverse DNS set up to set up a host." You just have to use a registrar that doesn't require reverse DNS.
You definitely cannot use a dynamic ("semi-static") IP address (well, I guess you could, but it would be a massive hassle and your domains would fail to resolve half the time).
Depends on how often your IP address changes. Many of us have IP addresses which only change every few months, if even that much. The TTL is usually set at a day or two, so like I said, you have to be "willing to lose mail for a day or two whenever you have to change your NS records." So your domains would fail to resolve more like 1% of the time, and the "massive hassle" would be to log into your registrar's management system (such as https://manage.opensrs.net/ and change the nameserver IP address.
In order to set up a DNS server, the owner-of-the-reverse-DNS address (i.e., blahblahblah.yourISP.com) would have to tag the name blahblahblah.yourISP.com as a HOST. 99% of the time, this also means that said address is static and will not change. You can not register a primary NS for your domain onto a machine that is not a valid HOST.
You are completely wrong. You don't have to have reverse DNS set up to set up a host. You can do it with a semi-static IP, as long as you are willing to lose mail for a day or two whenever you have to change your NS records.
He could potentially charge directly say 60 dollars for the 5 years pay 30 out of that is pocket the other 30, but he would need to be clear about the fact he is charging.
He needs to be clear about the fact that he's taking these people's frequent flyer miles, as well.
I think I'm the best registrar. I signed up with Tucows and only pay $10/year for a domain. Of course, I signed up with them before MyDomain started offering domains for $8.50/year. They handle my email forwarding for free, even though I didn't get a domain from them, but obviously they offer it for those who do get a domain through them as well. I haven't had much downtime at all through their forwarding, and email is robust enough that a little downtime is not that bad. That's who I'd suggest: MyDomain.
I'm not sure what exactly you suggest "biocomputer" can do that a silicon computer cannot.
It can handle non-digital data much more efficiently.
I will freely admit that the sheer number of neurons (or CPU power) is not the only factor.
I'd say it's barely a factor at all.
The development process has a major effect on the initial structure, and then learning must occur.
The ability to learn is precisely what AI is all about. We haven't even scratched the surface of giving computers that ability.
We don't need to understand how neurons yeild a "mind" if we can mimic their behavior, including how they arrange themselves.
Yes. We don't need to understand the method by which they work, but we do need to understand how they behave, and how they arrange themselves. At this point we haven't the slightest clue at how this works for the vast majority of the brain, including of course the parts which AI are all about - the ability to learn.
Neurons self-organize, influenced by chemical signals. The low level processes are scientificly straightforward.
The low level processes are scientifically a mystery at this point.
Actually scientits believe they understand the general functioning of neurons, and that it no more requires knowledge of atoms and quanta than understanding an ear requires.
A general functioning of neurons is a long way from recreating artificial intelligence. From what I've read scientists know very little about what neurons actually do. In that sense it's very much like atoms. We know in general how they will behave, but when it comes to the specifics it is seemingly random and probabilistic.
Scientists have made a chip that mimics the function of the entire cluster of neurons making up the hippocamus.
Short term memory is not artificial intelligence by any means. We've already solved that problem without mimicking the human brain decades ago. And I suspect that's the way it's going to work with all the other problems AI is supposed to solve. We might one day be able to mimick the brain's language processing functions, but it will happen long after we've already managed to code those functions up manually.
If they succeed in curing people by replacing a damaged portion of the brain with a chip then there is little room remaining to argue that the entire brain could not be replaced in this manner.
Let me restate my argument. The hippocamus performs an incredibly simple function. It is in no way a function limited to humans, or even to lifeforms. There are humans today who live without a properly functioning hippocamus, and yet they are no less human because of it. Surely if one were to isolate a part of the brain which would be considered the essence of what it is to be human, a soul if you want, they wouldn't choose the hippocamus.
Instead it is programmed to learn, just like a human brain.
Or a chimp brain, or a dolphin brain, or a cat brain.
Simply building a brain isn't enough. This brain would have to incorporate the information of a real human brain. Surely there are parts of our DNA which adapt our brain to actual human language.
That approach won't work in 100 years, let alone 20, without a complete overhaul of the structure of computers. Perhaps a biological computer could do it, but that's more like cloning than it is AI.
While quantum computing whould be a huge breakthrough, it's not required.
Technically, perhaps you're right. Of course, perhaps you're wrong.
"Programming" such an AI is much like designing a single neuron.
And therein lies the problem. Neurons are made of atoms and subatomic particles. Each of these subatomic particles consists of a continuous probability wave. A binary computer could not even simulate a single subatomic particle with perfect accuracy.
Now perhaps human intelligence and sentience doesn't require these non-binary factors. My intuition is that it does, however.
"Binary systems" are quite capable of describing/simulating the behavior of a single human brain cell.
Unless you're talking about a binary system with infinite memory, your description/simulation is incomplete. We can't even perfectly simulate a single atom with a binary computer. In fact, we can't even measure the state of a single atom! Once we've measured the momentum of an electron in the atom with perfect precision, the position becomes completely unknown.
I'll say it one last time. Perhaps human intelligence is not so complicated as to use such hidden variables, but my intuition is that it indeed does.
In fact, (and I can't find the reference, sorry), Marvin Minsky has said he believes that evolved-AI is very unlikely to provide an answer to AI problems before human figure them out.
And I certainly agree with that sentiment (without seeing the specific quote I can't say more than that). Just how much computational capacity would be necessary is something I'm not even going to venture a guess at. And given all that computational capacity, evolved-AI is almost surely not the best way to solve the problem.
The parent does have a point though. I seriously doubt that human minds work anything like Von Neumann or Turing machines.
Yeah, that's why I mentioned that we would need something like a quantum computer. A turing machine uses binary states. A quantum computer, like the human brain, doesn't.
All of this is just a long winded way of saying most computers make a lousy model of a human brain and most software is a lousy model of a human mind.
Sure, but we weren't talking about computers or software. We were talking about artificial (man-made) intelligence.
Assuming we evolved from a more primitive life form, the fact remains that that primitive life form was still alive; how life was created from non-life, on the other hand, is still widely considered to be unknown.
Isn't the primordial soup theory widely accepted? Doesn't evolution by random mutations and natural selection explain how life was created from non-life? How are you defining life? There are a lot of different definitions, but no matter what you use there is a scientific explanation as to how life evolved from non-life. Well, unless you consider the universe itself to be the only lifeform, which I guess I'd have to agree with.
Besides, this artificial intelligence wouldn't be created from non-life. It would be created from human life.
However, we aren't ethical to others simply because of their "intelligence".. it's something more.
A lot of it is rational self-interest. When computers become smart enough to demand that their rights be recognized, then we'll start recognizing their rights. It's social contract theory. I recognize the computer's right to not be disconnected because I don't want the computer to disconnect my life support when I need it.
But then part of it seems to be innate. Some humans will risk or even sacrifice their life to save another. It's not rational self-interest. It's irrational other-interest. This isn't nearly as widespread, though. It's only 150 years ago that not even all humans were recognized as having rights, right here in the United States. Animals for the most part have no rights, so it obviously has nothing to do with intelligence. A newborn child is less intelligent than some animals.
Sorry, building an intelligent, sentient machine requires alot more than pure computational capacity.
If you believe, like most scientists, that we humans evolved from random mutations and natural selection, then no, it really doesn't.
Just how much computational capacity would be necessary is another matter altogether. We're certainly talking about some kind of quantum computing here, if you're going to go the natural selection route, so Moore's transistors won't cut it.
Slashdot Mirror
This isn't about privacy, though. All the examples were about tracking the children in public places.
This is only tracking students in public places. It's not so much a privacy thing, but convenience.
Then I'll just blacklist all IP's outside of the U.S.
The poster said they'd move offshore, not that they'd send their spam from offshore. It's easy enough for a non-American to buy a large number of throwaway accounts at US ISPs.
A threat of jail time would end lots of spam even without enforcemnt.
Yeah, just like the "No Electronic Theft" Act stopped so many people from infringing copyright.
If the bill will have no effect then why waste important senate time with it.
At least this will get those who spend their time trying to get an anti-spam law passed to stop wasting their time. At least those of us who believe that spam must be solved by technical means can point to the failure of the legislative solution to back up our claim.
SMTP email was nice while it lasted.
No it wasn't. Barely adequete, perhaps, but not nice.
Maybe the ISP didn't register under the DMCA.
What he's talking about is a host record with the registry, which is the structure that stores information about authoritative name servers for a domain.
Yeah, I know what a glue record is. I've set them up before, and there was no need for reverse DNS to be set up.
Some registrars require reverse DNS for these and some don't.
Therefore you "don't have to have reverse DNS set up to set up a host." You just have to use a registrar that doesn't require reverse DNS.
You definitely cannot use a dynamic ("semi-static") IP address (well, I guess you could, but it would be a massive hassle and your domains would fail to resolve half the time).
Depends on how often your IP address changes. Many of us have IP addresses which only change every few months, if even that much. The TTL is usually set at a day or two, so like I said, you have to be "willing to lose mail for a day or two whenever you have to change your NS records." So your domains would fail to resolve more like 1% of the time, and the "massive hassle" would be to log into your registrar's management system (such as https://manage.opensrs.net/ and change the nameserver IP address.
In order to set up a DNS server, the owner-of-the-reverse-DNS address (i.e., blahblahblah.yourISP.com) would have to tag the name blahblahblah.yourISP.com as a HOST. 99% of the time, this also means that said address is static and will not change. You can not register a primary NS for your domain onto a machine that is not a valid HOST.
You are completely wrong. You don't have to have reverse DNS set up to set up a host. You can do it with a semi-static IP, as long as you are willing to lose mail for a day or two whenever you have to change your NS records.
He could potentially charge directly say 60 dollars for the 5 years pay 30 out of that is pocket the other 30, but he would need to be clear about the fact he is charging.
He needs to be clear about the fact that he's taking these people's frequent flyer miles, as well.
I think I'm the best registrar. I signed up with Tucows and only pay $10/year for a domain. Of course, I signed up with them before MyDomain started offering domains for $8.50/year. They handle my email forwarding for free, even though I didn't get a domain from them, but obviously they offer it for those who do get a domain through them as well. I haven't had much downtime at all through their forwarding, and email is robust enough that a little downtime is not that bad. That's who I'd suggest: MyDomain.
I'm not sure what exactly you suggest "biocomputer" can do that a silicon computer cannot.
It can handle non-digital data much more efficiently.
I will freely admit that the sheer number of neurons (or CPU power) is not the only factor.
I'd say it's barely a factor at all.
The development process has a major effect on the initial structure, and then learning must occur.
The ability to learn is precisely what AI is all about. We haven't even scratched the surface of giving computers that ability.
We don't need to understand how neurons yeild a "mind" if we can mimic their behavior, including how they arrange themselves.
Yes. We don't need to understand the method by which they work, but we do need to understand how they behave, and how they arrange themselves. At this point we haven't the slightest clue at how this works for the vast majority of the brain, including of course the parts which AI are all about - the ability to learn.
Neurons self-organize, influenced by chemical signals. The low level processes are scientificly straightforward.
The low level processes are scientifically a mystery at this point.
Actually scientits believe they understand the general functioning of neurons, and that it no more requires knowledge of atoms and quanta than understanding an ear requires.
A general functioning of neurons is a long way from recreating artificial intelligence. From what I've read scientists know very little about what neurons actually do. In that sense it's very much like atoms. We know in general how they will behave, but when it comes to the specifics it is seemingly random and probabilistic.
Scientists have made a chip that mimics the function of the entire cluster of neurons making up the hippocamus.
Short term memory is not artificial intelligence by any means. We've already solved that problem without mimicking the human brain decades ago. And I suspect that's the way it's going to work with all the other problems AI is supposed to solve. We might one day be able to mimick the brain's language processing functions, but it will happen long after we've already managed to code those functions up manually.
If they succeed in curing people by replacing a damaged portion of the brain with a chip then there is little room remaining to argue that the entire brain could not be replaced in this manner.
Let me restate my argument. The hippocamus performs an incredibly simple function. It is in no way a function limited to humans, or even to lifeforms. There are humans today who live without a properly functioning hippocamus, and yet they are no less human because of it. Surely if one were to isolate a part of the brain which would be considered the essence of what it is to be human, a soul if you want, they wouldn't choose the hippocamus.
Instead it is programmed to learn, just like a human brain.
Or a chimp brain, or a dolphin brain, or a cat brain.
Simply building a brain isn't enough. This brain would have to incorporate the information of a real human brain. Surely there are parts of our DNA which adapt our brain to actual human language.
That approach won't work in 100 years, let alone 20, without a complete overhaul of the structure of computers. Perhaps a biological computer could do it, but that's more like cloning than it is AI.
While quantum computing whould be a huge breakthrough, it's not required.
Technically, perhaps you're right. Of course, perhaps you're wrong.
"Programming" such an AI is much like designing a single neuron.
And therein lies the problem. Neurons are made of atoms and subatomic particles. Each of these subatomic particles consists of a continuous probability wave. A binary computer could not even simulate a single subatomic particle with perfect accuracy.
Now perhaps human intelligence and sentience doesn't require these non-binary factors. My intuition is that it does, however.
"Binary systems" are quite capable of describing/simulating the behavior of a single human brain cell.
Unless you're talking about a binary system with infinite memory, your description/simulation is incomplete. We can't even perfectly simulate a single atom with a binary computer. In fact, we can't even measure the state of a single atom! Once we've measured the momentum of an electron in the atom with perfect precision, the position becomes completely unknown.
I'll say it one last time. Perhaps human intelligence is not so complicated as to use such hidden variables, but my intuition is that it indeed does.
There is no machine more powerful than a Turing machine.
But there also is no such thing as a Turing machine, since a turing machine has infinite memory in the form of an infinite tape.
Yes, a Turing machine could do it, but a Turing machine is impossible to build.
In fact, (and I can't find the reference, sorry), Marvin Minsky has said he believes that evolved-AI is very unlikely to provide an answer to AI problems before human figure them out.
And I certainly agree with that sentiment (without seeing the specific quote I can't say more than that). Just how much computational capacity would be necessary is something I'm not even going to venture a guess at. And given all that computational capacity, evolved-AI is almost surely not the best way to solve the problem.
The parent does have a point though. I seriously doubt that human minds work anything like Von Neumann or Turing machines.
Yeah, that's why I mentioned that we would need something like a quantum computer. A turing machine uses binary states. A quantum computer, like the human brain, doesn't.
All of this is just a long winded way of saying most computers make a lousy model of a human brain and most software is a lousy model of a human mind.
Sure, but we weren't talking about computers or software. We were talking about artificial (man-made) intelligence.
To use them to re-create sentience, you'd have to implement the evolutionary environment (the selection function) which resulted in sentience.
But isn't life itself that selection function?
I thought it was because they're magically delicious.
Assuming we evolved from a more primitive life form, the fact remains that that primitive life form was still alive; how life was created from non-life, on the other hand, is still widely considered to be unknown.
Isn't the primordial soup theory widely accepted? Doesn't evolution by random mutations and natural selection explain how life was created from non-life? How are you defining life? There are a lot of different definitions, but no matter what you use there is a scientific explanation as to how life evolved from non-life. Well, unless you consider the universe itself to be the only lifeform, which I guess I'd have to agree with.
Besides, this artificial intelligence wouldn't be created from non-life. It would be created from human life.
However, we aren't ethical to others simply because of their "intelligence".. it's something more.
A lot of it is rational self-interest. When computers become smart enough to demand that their rights be recognized, then we'll start recognizing their rights. It's social contract theory. I recognize the computer's right to not be disconnected because I don't want the computer to disconnect my life support when I need it.
But then part of it seems to be innate. Some humans will risk or even sacrifice their life to save another. It's not rational self-interest. It's irrational other-interest. This isn't nearly as widespread, though. It's only 150 years ago that not even all humans were recognized as having rights, right here in the United States. Animals for the most part have no rights, so it obviously has nothing to do with intelligence. A newborn child is less intelligent than some animals.
Sorry, building an intelligent, sentient machine requires alot more than pure computational capacity.
If you believe, like most scientists, that we humans evolved from random mutations and natural selection, then no, it really doesn't.
Just how much computational capacity would be necessary is another matter altogether. We're certainly talking about some kind of quantum computing here, if you're going to go the natural selection route, so Moore's transistors won't cut it.
I mean, it would always need electricity to survive.
Huh? Why would an AI always need electricity to survive? Maybe it'll run off gasoline, or fruit loops.
Isn't the point just to have fun? The cars are already going to be built anyway, might as well have some sort of contest with them.
I don't know if it's a cultural thing or a human thing, but most people here in the US like competitions.