This is that state-of-the-art. And there actually is nothing on the horizon that could do any better. Hence the threat of AI is mostly to people doing jobs that are on the same level. Anybody actually reasonably good at problem solving will not be replaced by AI anytime soon and possibly not ever.
Indeed. The only alternative is the collapse of Capitalism (because people will not be able to buy enough things) and we have no replacement for Capitalism at this time. And it needs to be at least a step up from the absolute minimum needed to survive for the same reason. It will be a long and slow process though, because many people just cannot stand that somebody else gets anything for free.
Hahahha, think of what an actually working hypothetical AI would conclude in this situation: The customer/user is a moron and has no clue! That will drive AI adoption like crazy! (If it ever happens at all....)
Indeed. And here we reach the the limits of Capitalism to distribute the wealth to the population. At the same time Capitalism depends critically on people being able to buy things, hence wealth must continue to be distributed to the population or things collapse. It is no accident that an unconditional basic income for everybody is seriously being discussed now in some countries and it is not idealists with their heads in the clouds that drive this discussion. (They are routinely trying to hijack it though, which somewhat obscures that this is about a critically important problem).
The main problem today seems to be one of irrational envy: For example most ( > 80%) people in Switzerland say they would continue working with an unconditional basic income, even if that allows them to live reasonably well already. The problem is that most people think that not so may of their fellow citizens would do so. Still, long-term, there really is no alternative to it if we want to keep civilization going.
Indeed. I made excellent experiences with "open notes" exams as the lecturer of an EE course over several years. (That was the first time I was the primary examiner and could do that.) For one, students take better notes and ask questions during the lecture if things are unclear. And you can ask more difficult things, which makes the exams better overall. I also got very positive feedback from students, saying that while things did not get easier, they understood more and generally felt the course was more worthwhile taking as they could focus on understanding things and not on remembering them. And while you have to ask new questions every time, I did not find that difficult or hard to do.
Personally, I will only do "open notes" in the future whenever the decision is up to me and, if the lecture is based on a book, "open book".
Unless the degree aimed for is worthless in itself (a far to common occurrence these days), the only effect this has is students hurting their own skills. And once you have an aural exam or an advanced exam that requires actual understanding, that will come back to bite the cheaters.
Exactly, thank you. Error correction is not magic. Error correction is what keeps QC research going (very, very slowly) at this time, because without it there would be absolutely no point.
Actually, exponential growth for computer speeds has stopped a while ago and was never as good as advertised before. The thing is that actual experts understand that many important problems cannot be parallelized and hence single-thread performance is what determines speed. That has mostly stalled in the last 10 years or so.
Kurzweil is an incompetent moron with a grand vision he sells well. Kind of a bit like Trump, although I do not think Trump is stupid enough to believe the things he says. With Kurzweil I am not so sure.
You mean the arrogant, incompetent and clueless morons that think they are the kings of the world? No thanks. The Pi community sucks even worse than the Pi hardware.
Indeed. The Pi sucks badly interface-wise. Still not fixed on the Pi3, because they stay with an inferior Broadcom SoC. Get real USB and real networking instead of the unreliable crap the Pi has. And some competitors even have SATA.
The available evidence seems to indicate that strongly. Also take into account how they are failing everywhere else, and they really, really will try very hard to make Win10 a success. If they still fail under those conditions, then they do not have the technological skills. Another indicator is how hard they push Win10 in not merit-based ways, see free upgrades, (almost) forced upgrades, etc. My take is they have lost mastery of their platform and know it, and hence they are running deeply scared now.
That is naive. You assume maintaining entanglement gets less than linearly more difficult and that noise is independent of the number of qbits. Both are not reasonable assumptions.
I agree. At this time, we cannot even know whether the physics itself holds up. Factoring 15 is something that can be done with a conventional analog computer, no actual quantum effects needed. So there are two hard road-blocks to this ever threatening RSA of real sizes: a) it may not actually be possible to use quantum effects for computations and what we currently observe may be something different and b) quantum computers may not scale to the required bit-sizes, ever. We see these hard scalability limits in every other technology, there is no reason to believe quantum computing will magically not be subject to them. And with the lack of progress in scalability in the last 20 years or so, It seems the limits may be very, very low, for example at 100 qbits.
From the lack of scaling in the last 20 years or so of quantum computing research, I would put 50 years for low RSA bit-counts (e.g. 768 bits, requiring > 1000 qbits if you take error correction into account) as lower limit. It may also well be "never".
And scaling this up beyond 100 qubits with full quantum coherence and quantum control of qubit operations (from all reports e.g. the D-Wave machine "only" does quantum annealing with ~500 qubits, and doesn't implement a universal quantum computer) is something that's still quite a bit away. How long? I don't think anybody can really predict. Could be 5 years, could be 10, could be 50.
Could also very well be "never". Just look at the lengths CPU manufacturers have to go to get to 5GHz. A bit more is likely feasible, but, say, 100GHz is likely completely infeasible unless a mythical new technology presents itself. It has not, despite now 50 years of intense research, so what we currently have in CPUs may very well be close to the end of the line in this universe. It is quite likely that quantum computing (if it even works at all, factoring 15 could well be some other effect), runs into pretty hard scalability limits at 100 qbits or so and will never be a threat even to yesterday's RSA key lengths.
That key has eluded researchers for a few decades now. It looks very much like there is an upper limit on the number of qbits that can be entangled in practice if computations are to be performed and as if that upper limit is somewhere around 100. With that, not even very old and outdated RSA-768 is threatened.
That is why these stories are so utterly demented. They are akin to claiming the invention of the logic gate will make 2048-bit computers possible that run at 1000GHz. As we now see in practice, 64 bit at 5GHz is pretty much the viable limit for low-cost and it does not go much further with extreme hardware. In reality, things do not scale after a certain limit and for quantum computing, that limit will be very low.
First, most encryption is not even really affected. For block-ciphers a working and large enough QC halves the key-length. AES-256 would still be perfectly secure and AES-128 would still be hard (but maybe possible) to break. And second, factoring RSA-2048 (which is regarded as too short today) would need around 2200 qbits to factor with this "breakthrough". They are at 5 qbits now. Where where they 10 years ago? Oh, right, at the same low number. If progress is made at this rate, they will be able to break RAS-2048 in x years, where x goes towards infinity, i.e. _never_.
This is about as valid as claiming the invention of paper threatens RSA, after all you can do attacks far faster with paper than with stone tablets.
Can we please stop the moronic and false "success" stories about quantum computing?
And that is one core reason I will not "upgrade" anytime soon. I would have absolutely no problem paying more for a "professional" edition, if that then actually is professional and gives me the control I can expect as a professional. But they do not offer that anymore. So I will stay on Win7 (pro) as my game-launcher and do all other work (except some MS office) on Linux as I have done for 20 years now.
Slashdot Mirror
This is that state-of-the-art. And there actually is nothing on the horizon that could do any better. Hence the threat of AI is mostly to people doing jobs that are on the same level. Anybody actually reasonably good at problem solving will not be replaced by AI anytime soon and possibly not ever.
Indeed. The only alternative is the collapse of Capitalism (because people will not be able to buy enough things) and we have no replacement for Capitalism at this time. And it needs to be at least a step up from the absolute minimum needed to survive for the same reason. It will be a long and slow process though, because many people just cannot stand that somebody else gets anything for free.
Hahahha, think of what an actually working hypothetical AI would conclude in this situation: The customer/user is a moron and has no clue! That will drive AI adoption like crazy! (If it ever happens at all....)
While I agree that most developers _are_ incompetent, what we have in AI these days is a lot worse.
Indeed. And here we reach the the limits of Capitalism to distribute the wealth to the population. At the same time Capitalism depends critically on people being able to buy things, hence wealth must continue to be distributed to the population or things collapse. It is no accident that an unconditional basic income for everybody is seriously being discussed now in some countries and it is not idealists with their heads in the clouds that drive this discussion. (They are routinely trying to hijack it though, which somewhat obscures that this is about a critically important problem).
The main problem today seems to be one of irrational envy: For example most ( > 80%) people in Switzerland say they would continue working with an unconditional basic income, even if that allows them to live reasonably well already. The problem is that most people think that not so may of their fellow citizens would do so. Still, long-term, there really is no alternative to it if we want to keep civilization going.
That is the only sane way to test actual skills.
Indeed. I made excellent experiences with "open notes" exams as the lecturer of an EE course over several years. (That was the first time I was the primary examiner and could do that.) For one, students take better notes and ask questions during the lecture if things are unclear. And you can ask more difficult things, which makes the exams better overall. I also got very positive feedback from students, saying that while things did not get easier, they understood more and generally felt the course was more worthwhile taking as they could focus on understanding things and not on remembering them. And while you have to ask new questions every time, I did not find that difficult or hard to do.
Personally, I will only do "open notes" in the future whenever the decision is up to me and, if the lecture is based on a book, "open book".
Unless the degree aimed for is worthless in itself (a far to common occurrence these days), the only effect this has is students hurting their own skills. And once you have an aural exam or an advanced exam that requires actual understanding, that will come back to bite the cheaters.
There clearly is no new MS, just an old MS that is getting a bit desperate and finds what some of its customers want increasingly hard to ignore.
... a little bit. I guess MS has reversed its stance that Linux is "a cancer" and generally bad.
Exactly, thank you. Error correction is not magic. Error correction is what keeps QC research going (very, very slowly) at this time, because without it there would be absolutely no point.
Actually, exponential growth for computer speeds has stopped a while ago and was never as good as advertised before. The thing is that actual experts understand that many important problems cannot be parallelized and hence single-thread performance is what determines speed. That has mostly stalled in the last 10 years or so.
Kurzweil is an incompetent moron with a grand vision he sells well. Kind of a bit like Trump, although I do not think Trump is stupid enough to believe the things he says. With Kurzweil I am not so sure.
Indeed. And both Ethernet and USB that randomly stops working has been my experience with the RPi, I am _not_ complaining about speed.
You mean the arrogant, incompetent and clueless morons that think they are the kings of the world? No thanks. The Pi community sucks even worse than the Pi hardware.
Indeed. The Pi sucks badly interface-wise. Still not fixed on the Pi3, because they stay with an inferior Broadcom SoC. Get real USB and real networking instead of the unreliable crap the Pi has. And some competitors even have SATA.
The available evidence seems to indicate that strongly. Also take into account how they are failing everywhere else, and they really, really will try very hard to make Win10 a success. If they still fail under those conditions, then they do not have the technological skills. Another indicator is how hard they push Win10 in not merit-based ways, see free upgrades, (almost) forced upgrades, etc. My take is they have lost mastery of their platform and know it, and hence they are running deeply scared now.
That is why you need good lawyers ;-)
That is naive. You assume maintaining entanglement gets less than linearly more difficult and that noise is independent of the number of qbits. Both are not reasonable assumptions.
I agree. At this time, we cannot even know whether the physics itself holds up. Factoring 15 is something that can be done with a conventional analog computer, no actual quantum effects needed. So there are two hard road-blocks to this ever threatening RSA of real sizes: a) it may not actually be possible to use quantum effects for computations and what we currently observe may be something different and b) quantum computers may not scale to the required bit-sizes, ever. We see these hard scalability limits in every other technology, there is no reason to believe quantum computing will magically not be subject to them. And with the lack of progress in scalability in the last 20 years or so, It seems the limits may be very, very low, for example at 100 qbits.
From the lack of scaling in the last 20 years or so of quantum computing research, I would put 50 years for low RSA bit-counts (e.g. 768 bits, requiring > 1000 qbits if you take error correction into account) as lower limit. It may also well be "never".
And scaling this up beyond 100 qubits with full quantum coherence and quantum control of qubit operations (from all reports e.g. the D-Wave machine "only" does quantum annealing with ~500 qubits, and doesn't implement a universal quantum computer) is something that's still quite a bit away. How long? I don't think anybody can really predict. Could be 5 years, could be 10, could be 50.
Could also very well be "never". Just look at the lengths CPU manufacturers have to go to get to 5GHz. A bit more is likely feasible, but, say, 100GHz is likely completely infeasible unless a mythical new technology presents itself. It has not, despite now 50 years of intense research, so what we currently have in CPUs may very well be close to the end of the line in this universe. It is quite likely that quantum computing (if it even works at all, factoring 15 could well be some other effect), runs into pretty hard scalability limits at 100 qbits or so and will never be a threat even to yesterday's RSA key lengths.
That key has eluded researchers for a few decades now. It looks very much like there is an upper limit on the number of qbits that can be entangled in practice if computations are to be performed and as if that upper limit is somewhere around 100. With that, not even very old and outdated RSA-768 is threatened.
That is why these stories are so utterly demented. They are akin to claiming the invention of the logic gate will make 2048-bit computers possible that run at 1000GHz. As we now see in practice, 64 bit at 5GHz is pretty much the viable limit for low-cost and it does not go much further with extreme hardware. In reality, things do not scale after a certain limit and for quantum computing, that limit will be very low.
First, most encryption is not even really affected. For block-ciphers a working and large enough QC halves the key-length. AES-256 would still be perfectly secure and AES-128 would still be hard (but maybe possible) to break. And second, factoring RSA-2048 (which is regarded as too short today) would need around 2200 qbits to factor with this "breakthrough". They are at 5 qbits now. Where where they 10 years ago? Oh, right, at the same low number. If progress is made at this rate, they will be able to break RAS-2048 in x years, where x goes towards infinity, i.e. _never_.
This is about as valid as claiming the invention of paper threatens RSA, after all you can do attacks far faster with paper than with stone tablets.
Can we please stop the moronic and false "success" stories about quantum computing?
Apparently MS does not even have good lawyers anymore or they would not have messed up the NDA thing!
And that is one core reason I will not "upgrade" anytime soon. I would have absolutely no problem paying more for a "professional" edition, if that then actually is professional and gives me the control I can expect as a professional. But they do not offer that anymore. So I will stay on Win7 (pro) as my game-launcher and do all other work (except some MS office) on Linux as I have done for 20 years now.