Good comparison between GUI and command line! Selecting from predefined, context-dependent choices is far, far simpler than unconstrained command entry for the user. It is also about the limit of what voice-recognition can do these days.
Incidentally, calling the outer interface layer the "OS" already shows that the writers here are utterly clueless.
There are some rather strong indications it will not go very far at all. They are not reliable proof, sure, but proving a negative is notoriously hard. One is that at this time, after half a century of research into it, there still is no credible theory how intelligence could be generated artificially. The only thing we have that can mimic some aspects of intelligence is automated theorem proving, and that cannot scale up to what a smart human can do in this universe, not enough matter and energy available. There is nothing else. The other indicator is that actual intelligence is only observable together with consciousness, and nobody has the least clue what that is. In fact, the current state-of-the art in physics would indicate that consciousness is impossible, which is clearly wrong. (There are some morons that pretend to be scientists in neuro-"science" that claim it is an "emergent property", which is just a scientifically sounding and dishonest way of saying "we do not know".) And then there is the little fact that AI has made exactly no progress at all in the direction of anything that can reasonably be called intelligence. All they have (which is still useful, no argument about that) is clever ways to fake intelligence and doing things without intelligence where we once thought that intelligence was required. And that is really all they have.
Bullshit. The observed thrust is rather close to the margin or error. Nobody halfway competent plans an experiment that way. Incidentally, NASA did not do any test on this at all. NASA Eagleworks, which is not NASA, did the tests.
Well, to be fair, if a scaling QC ever materializes, AES-128 may be just barely vulnerable (2^64 effort). But AES-256 will still have a very comfortable security margin.
The crypto-morons that think everything is easy and they of course understand the questions that takes an actual expert a decade or so to really get will never die out. They are a close cousin to the morons that think coding is easy.
Indeed. On the plus side, we already have QC-proof symmetric encryption today. It just gives you a square-root improvement, so AES-256 is proof against a QC. The moron above probably does not know that, as a one-time pad is symmetrical encryption and hence does not improve against AES-256 in the presence og working, scaling QC in actual reality.
Of QC ever gets real. Strikes me a lot like "AI", which looks these days as it may actually be impossible in this universe if you want something at least as smart as a human moron. Quantum factoring has gone from 4 bits to 16 bits in 25 years or so. Even if it continues to scale like that (which it will not, there is indication it scales inverse-exponentially, so 30-100 bits or so may be the absolute upper limit), it will not be a threat to modern encryption for 50-100 years, and that is only if we continue to use 1024 bit primes.
There can be no test in orbit until you get significant thrust for a reasonable amount of energy. This thing cannot do that. It gives you a tiny trust for a lot of energy, that is if it works at all. Because the thrust is so tiny, smart people actually expect this will turn out to be an equally tiny measurement error.
No, not at all. The thrust (if it is there) if far, far too low for that. Do you people even read the description of what this thing maybe does before you start to fantasize?
Ah, so there is a second group of fraudsters trying to capitalize on the success of the first group to con the public into believing something with grossly insufficient evidence?
Not at all. At this time there is a tiny chance this thing will produce a minuscule amount of thrust if fed a lot of energy. The far larger chance is that there is a measurement error that us tricky to find.
That entirely depends on the gif and who it is sent to and with what intent. Understanding this does require some actual understanding of the facts involved though, and you obviously have none.
You have no clue how a test of actual scientific value works. You are a complete moron. Now, if you repeat that test with a few 1000 people with different types and intensity of epilepsy and nobody has any ill effects, then there may be some validity to the hypothesis that he is lying.
I would also be fine with whenever something like this is known to the person posting it online, that they will have to add a warning. Like a food-manufacturer is required to give allergy-warnings for the more common allergy-triggers, like the epilepsy-warnings that come with a lot of games.
Slashdot Mirror
Good comparison between GUI and command line! Selecting from predefined, context-dependent choices is far, far simpler than unconstrained command entry for the user. It is also about the limit of what voice-recognition can do these days.
Incidentally, calling the outer interface layer the "OS" already shows that the writers here are utterly clueless.
There are some rather strong indications it will not go very far at all. They are not reliable proof, sure, but proving a negative is notoriously hard. One is that at this time, after half a century of research into it, there still is no credible theory how intelligence could be generated artificially. The only thing we have that can mimic some aspects of intelligence is automated theorem proving, and that cannot scale up to what a smart human can do in this universe, not enough matter and energy available. There is nothing else. The other indicator is that actual intelligence is only observable together with consciousness, and nobody has the least clue what that is. In fact, the current state-of-the art in physics would indicate that consciousness is impossible, which is clearly wrong. (There are some morons that pretend to be scientists in neuro-"science" that claim it is an "emergent property", which is just a scientifically sounding and dishonest way of saying "we do not know".) And then there is the little fact that AI has made exactly no progress at all in the direction of anything that can reasonably be called intelligence. All they have (which is still useful, no argument about that) is clever ways to fake intelligence and doing things without intelligence where we once thought that intelligence was required. And that is really all they have.
Bullshit. The observed thrust is rather close to the margin or error. Nobody halfway competent plans an experiment that way. Incidentally, NASA did not do any test on this at all. NASA Eagleworks, which is not NASA, did the tests.
Well, to be fair, if a scaling QC ever materializes, AES-128 may be just barely vulnerable (2^64 effort). But AES-256 will still have a very comfortable security margin.
The crypto-morons that think everything is easy and they of course understand the questions that takes an actual expert a decade or so to really get will never die out. They are a close cousin to the morons that think coding is easy.
Indeed. On the plus side, we already have QC-proof symmetric encryption today. It just gives you a square-root improvement, so AES-256 is proof against a QC. The moron above probably does not know that, as a one-time pad is symmetrical encryption and hence does not improve against AES-256 in the presence og working, scaling QC in actual reality.
Protip: If anybody in an encryption-debate brings up the one-time pad, then they have just outed themselves as clueless amateurs.
Of QC ever gets real. Strikes me a lot like "AI", which looks these days as it may actually be impossible in this universe if you want something at least as smart as a human moron. Quantum factoring has gone from 4 bits to 16 bits in 25 years or so. Even if it continues to scale like that (which it will not, there is indication it scales inverse-exponentially, so 30-100 bits or so may be the absolute upper limit), it will not be a threat to modern encryption for 50-100 years, and that is only if we continue to use 1024 bit primes.
Most insight-less comment of the day. No wonder you post as AC.
There can be no test in orbit until you get significant thrust for a reasonable amount of energy. This thing cannot do that. It gives you a tiny trust for a lot of energy, that is if it works at all. Because the thrust is so tiny, smart people actually expect this will turn out to be an equally tiny measurement error.
No, not at all. The thrust (if it is there) if far, far too low for that. Do you people even read the description of what this thing maybe does before you start to fantasize?
Ah, so there is a second group of fraudsters trying to capitalize on the success of the first group to con the public into believing something with grossly insufficient evidence?
Not at all. At this time there is a tiny chance this thing will produce a minuscule amount of thrust if fed a lot of energy. The far larger chance is that there is a measurement error that us tricky to find.
What tests by credible (!) scientific organizations are you talking about? AFAIK there were none.
Maybe that will send a message...
I agree. Manual "fixes" will make the issue worse.
If they are identified, I think they should be fined $1 and then be given a medal.
It will get interesting when/if that happens. A possible approach would be to teach the AI to lie. That may not turn out well...
That entirely depends on the gif and who it is sent to and with what intent. Understanding this does require some actual understanding of the facts involved though, and you obviously have none.
The display does not flicker, but what is displayed on it sure can.
You have no clue how a test of actual scientific value works. You are a complete moron. Now, if you repeat that test with a few 1000 people with different types and intensity of epilepsy and nobody has any ill effects, then there may be some validity to the hypothesis that he is lying.
Speech used to drive somebody to suicide intentionally and hate-speech are prosecutable. This is similar, but worse.
I would also be fine with whenever something like this is known to the person posting it online, that they will have to add a warning. Like a food-manufacturer is required to give allergy-warnings for the more common allergy-triggers, like the epilepsy-warnings that come with a lot of games.
"prosecution for criminal acts" - is sending an animation a criminal act?
If it is done with the intent and a reasonable expectation to cause serious harm, it is.
If somebody makes a credible attempt to cause serious harm to somebody else, limitations to any expectation of privacy apply.