"First, for it to be secure, a theory which we know is flawed (Quantum Theory, does not account for Gravity) needs to hold up to an extremely precise level"
Quantum Field Theory makes extremely precise mathematical predictions, which have been shown correct in many experiments. Measuring the gravitational effect upon a particles momentum is nigh impossible due to how incredibly weak gravity is compared to the other forces (notice this is different from measuring the time dilation effects of different gravitational field strengths).
Precision:
These happen to be too low for crypto. For crypto we would need around 256 bits, i.e. around 1 in 10^76. Even only 128 bit crypto would be around 1 in 10^38. I do agree that the level of precision is _very_ impressive for Physics, it is just not enough by a very long shot for secure crypto. At the precision level needed for crypto, gravity matters very much.
I'd also like to suggest that experts in AI are, by definition, embracing AI. Since, you know, they have devoted significant time in their lives to becoming experts in the subject. Can you name a single "expert" in AI that doesn't "embrace AI"? What would that even look like?
Hahahaha, you are soooo badly off about this one. This is Science, not Religion.
Indeed. One problem is that you are typically not aware of all paths of attack or that the level you can secure different paths is different. Hence it boils down to risk management. For small aspects, you can get 100% security though. For example a 100 bit entropy password with Argon2 and reasonable parameters is unlikely to be breakable, ever. But it does nothing about the alternate attack path a keylogger offers.
Password protection (e.g. by Argon2 or to a lesser degree PBKDF2) brings this down quite a bit. Sure, if you want to memorize an encryption key directly, you would need to go to 256 bit or so. But for a password, "absolute security" against brute-forcing starts somewhere around 100 bits. Your 91 bit entropy password will not be broken. It may get observed or circumvented though.
No, they cannot. Chess games use "quality" metrics to decide which move to make. These are not compatible with "brute force" approaches. A "brute force" algorithm just tries everything and can only recognize success or failure, nothing in between. The problem here may be that in CS, the term "brute force" has a well-defined meaning, while in general usage it does not.
(Oh, but the way: You, too, are just an automation; at some point, automation becomes so complex that it is indistinguishable from "sentience".)
You can spread your religion-surrogate crap somewhere else please. There is no scientific basis to your claim.
You appear to have it backwards. Nearly all if not all of our scientific knowledge points to humans being an automation just like any other robot, although much more complex and with fundamentally different hardware. Religion or religion-surrogates are the only source of belief to the contrary.
No, I do not. You are misinterpreting the Science. It actually says no such thing and the question is wide open. Now, I do not propose to put a "god" or any such nonsense in there (I am an atheist), but Physicalism is usually practiced and defended much the same way religion is. It is "obvious truth", and it assumes known Physics covers everything (when not even Physics makes such a claim and rather points out that it does not). People that do not buy it are accused of being religious (in the sense of "wrong god"). Rather strong evidence to the contrary is completely ignored (the nature of consciousness is completely unknown, how intelligence works in a smart person is completely unknown and consciousness and real intelligence are _only_ observable together) and available facts are misinterpreted to support the belief ("nearly if not all of our scientific knowledge...").
Except for the absence of a "god", Physicalism is religion. Similar variations of the religious meme have existed before, see, e.g. some variants of communism or national socialism. These are religions in all but name and about as well-funded on actual facts.
There is pretty strong evidence that we are hybrids. Quite a few things clearly run on automation, but some things do not. It takes an act of will to decide to "think about something" though and many people manage that only rarely. Of course, many people falsely assume that they are using free will all the time and even for trivial decisions, but that is just not true.
The problem here is that there is no mechanism for consciousness on physics and there is no known mechanism that can produce the intelligence of a smart human being physically. In fact, after many decades of AI research, it looks like "strong AI" may actually not be possible physically in this universe. You are falling for a "black box" argument, that says that anything observable at the interface of a black box must be crated in there. The problem here is that you may not be able to observe everything. In the age of the cell-hone, that idea should not be so hard to understand. The other problem is that the physicalist argument is circular: Assume everything is physical and you can proof everything is physical. That is not science, that is religion. At this time, the question is wide open.
And that is just my point. While my own research area is not really related, I have following AI research for now around 30 years and that is what I see. Of course, people mistaking some one-shot implementation optimization for an "algorithmic advance" are just as common today as they were back then.
Well, maybe. As to an UBI, it is certainly feasible financially in some countries and maybe in a lot more, as it basically is a simplification of the welfare system. The most serious problem with it is that many people will not be able to live well with not having a job. It gives meaning, substance and structure. I think we will see riots and massively increased mental health issues when UBIs are implemented carelessly. Ultimately, there really is no other choice I can see, as automation (not AI) will take a lot of jobs in the next decade or two, possibly more than half of them. But people (except for a very small minority) will need to find something worthwhile they can do with their time and doing that by themselves is difficult for many.
Well, probably the thing is that murders are just humans having made a mistake, while sex offenders clearly are monsters that will rape, pillage and murder...oh, wait.
First, for it to be secure, a theory which we know is flawed (Quantum Theory, does not account for Gravity) needs to hold up to an extremely precise level. Second, the engineering needs to be secure as well, and most instances of this have been broken in the past. And third, it is nonsense anyways, since after the key exchange, you have to revert to conventional encryption for the actual data transmission anyways.
Why this BS still gets any attention is really beyond me. People that want to believe in magic?
Actually, it is a _lot_ less of an improvement than it sounds and it has very limited scalability. All these algorithms have sub-linear performance and a factor of 10 or even 100 may not give you anything substantially better.
I fully agree to that. Well, mostly, as really understanding large training data sets is difficult. But using a tool without understanding its limits and where it can have surprising behavior is dangerous and extremely unprofessional.
Useless metric. The actual scientific advancement from these papers is a tiny faction of what was known before. Science is unfortunately not immune to hypes, but they do not produce much advancement.
Oh, it is nice than I can buy a reasonable (hobbyist use only) 3D printer kit for $300 or so. Gives me something to tinker with. The whole bubble was just exceptionally stupid though.
Oh yes. That would help so incredible much. And then make sure all developers, designers and architects either get some real basic understanding on security and have somebody competent they can ask. In most cases, that will be a consultant, as even large enterprises cannot keep in-house experts current. There is just not enough variance in one application landscape. Consultants, on the other hand, see a lot of different situations in a lot of different places.
Of course, the question of getting competent IT security consultants is a pretty hard one as well. There are too many that just work on the surface, with buzzwords and no clear understanding. My minimal list for the technical skills of an IT security consultant is at least intermediary skills in: Networking, system administration (Unix/Linux a must), software development in several languages on several platforms, solid algorithm and data-structures background, secure software development, use of crypto, attack techniques, economics of attacks, distributed systems. Clear writing, ability to explain things, high analytical skills, people skills, etc. are a must as well.
Of course, this is a list where you need real-life "Chief Engineer" with a special interest in security. But nothing much less will do the job and even that Chief Engineer will regularly need to discuss things with some equally qualified colleagues. My advice: Ignore all the large consultancies (in particular also IBM), because while they may have a few such people, they will not give you access. Then look for non-flashy, matter of fact small ones. A significant number of PhDs in IT Security related fields from a good university and is good indicator of skill, but only when coupled with hands-on skills from the above list. Of course, these people will be expensive and will not have a lot of time. But hiring them will be very much worth it.
Indeed. As this is published now, it is in the "likely, but not certain" class. This is a call for assistance, may it be with possible flaws in the measurement or mathematics, may it be with more experiments. Also keep in mind that "experiment" can mean some targeted searching though the absolute huge amounts of data a collider like the LHC produces. This is how particle physicists firm up things, this is a global cooperation and that is the only way it works.
Slashdot Mirror
"First, for it to be secure, a theory which we know is flawed (Quantum Theory, does not account for Gravity) needs to hold up to an extremely precise level"
Quantum Field Theory makes extremely precise mathematical predictions, which have been shown correct in many experiments. Measuring the gravitational effect upon a particles momentum is nigh impossible due to how incredibly weak gravity is compared to the other forces (notice this is different from measuring the time dilation effects of different gravitational field strengths).
Precision:
These happen to be too low for crypto. For crypto we would need around 256 bits, i.e. around 1 in 10^76. Even only 128 bit crypto would be around 1 in 10^38.
I do agree that the level of precision is _very_ impressive for Physics, it is just not enough by a very long shot for secure crypto. At the precision level needed for crypto, gravity matters very much.
The other ones only implicitly....
There are tons of books around that are filled with complete nonsense.
I'd also like to suggest that experts in AI are, by definition, embracing AI. Since, you know, they have devoted significant time in their lives to becoming experts in the subject. Can you name a single "expert" in AI that doesn't "embrace AI"? What would that even look like?
Hahahaha, you are soooo badly off about this one. This is Science, not Religion.
Indeed. One problem is that you are typically not aware of all paths of attack or that the level you can secure different paths is different. Hence it boils down to risk management. For small aspects, you can get 100% security though. For example a 100 bit entropy password with Argon2 and reasonable parameters is unlikely to be breakable, ever. But it does nothing about the alternate attack path a keylogger offers.
Password protection (e.g. by Argon2 or to a lesser degree PBKDF2) brings this down quite a bit. Sure, if you want to memorize an encryption key directly, you would need to go to 256 bit or so. But for a password, "absolute security" against brute-forcing starts somewhere around 100 bits. Your 91 bit entropy password will not be broken. It may get observed or circumvented though.
No, they cannot. Chess games use "quality" metrics to decide which move to make. These are not compatible with "brute force" approaches. A "brute force" algorithm just tries everything and can only recognize success or failure, nothing in between. The problem here may be that in CS, the term "brute force" has a well-defined meaning, while in general usage it does not.
(Oh, but the way: You, too, are just an automation; at some point, automation becomes so complex that it is indistinguishable from "sentience".)
You can spread your religion-surrogate crap somewhere else please. There is no scientific basis to your claim.
You appear to have it backwards. Nearly all if not all of our scientific knowledge points to humans being an automation just like any other robot, although much more complex and with fundamentally different hardware. Religion or religion-surrogates are the only source of belief to the contrary.
No, I do not. You are misinterpreting the Science. It actually says no such thing and the question is wide open. Now, I do not propose to put a "god" or any such nonsense in there (I am an atheist), but Physicalism is usually practiced and defended much the same way religion is. It is "obvious truth", and it assumes known Physics covers everything (when not even Physics makes such a claim and rather points out that it does not). People that do not buy it are accused of being religious (in the sense of "wrong god"). Rather strong evidence to the contrary is completely ignored (the nature of consciousness is completely unknown, how intelligence works in a smart person is completely unknown and consciousness and real intelligence are _only_ observable together) and available facts are misinterpreted to support the belief ("nearly if not all of our scientific knowledge...").
Except for the absence of a "god", Physicalism is religion. Similar variations of the religious meme have existed before, see, e.g. some variants of communism or national socialism. These are religions in all but name and about as well-funded on actual facts.
There is pretty strong evidence that we are hybrids. Quite a few things clearly run on automation, but some things do not. It takes an act of will to decide to "think about something" though and many people manage that only rarely. Of course, many people falsely assume that they are using free will all the time and even for trivial decisions, but that is just not true.
The problem here is that there is no mechanism for consciousness on physics and there is no known mechanism that can produce the intelligence of a smart human being physically. In fact, after many decades of AI research, it looks like "strong AI" may actually not be possible physically in this universe. You are falling for a "black box" argument, that says that anything observable at the interface of a black box must be crated in there. The problem here is that you may not be able to observe everything. In the age of the cell-hone, that idea should not be so hard to understand. The other problem is that the physicalist argument is circular: Assume everything is physical and you can proof everything is physical. That is not science, that is religion. At this time, the question is wide open.
And that is just my point. While my own research area is not really related, I have following AI research for now around 30 years and that is what I see.
Of course, people mistaking some one-shot implementation optimization for an "algorithmic advance" are just as common today as they were back then.
I don't. I am criticizing the reporting here.
Well, maybe. As to an UBI, it is certainly feasible financially in some countries and maybe in a lot more, as it basically is a simplification of the welfare system. The most serious problem with it is that many people will not be able to live well with not having a job. It gives meaning, substance and structure. I think we will see riots and massively increased mental health issues when UBIs are implemented carelessly. Ultimately, there really is no other choice I can see, as automation (not AI) will take a lot of jobs in the next decade or two, possibly more than half of them. But people (except for a very small minority) will need to find something worthwhile they can do with their time and doing that by themselves is difficult for many.
Well, probably the thing is that murders are just humans having made a mistake, while sex offenders clearly are monsters that will rape, pillage and murder...oh, wait.
First, for it to be secure, a theory which we know is flawed (Quantum Theory, does not account for Gravity) needs to hold up to an extremely precise level. Second, the engineering needs to be secure as well, and most instances of this have been broken in the past. And third, it is nonsense anyways, since after the key exchange, you have to revert to conventional encryption for the actual data transmission anyways.
Why this BS still gets any attention is really beyond me. People that want to believe in magic?
Well, MS seems to have learned some honesty finally.
That very much is it. Not the only hype going on, but one of the worst ones at the moment for sheer non-understanding.
That is why some will get rich (off other investors), even if there will not be any significant advances in products.
Actually, it is a _lot_ less of an improvement than it sounds and it has very limited scalability. All these algorithms have sub-linear performance and a factor of 10 or even 100 may not give you anything substantially better.
I fully agree to that. Well, mostly, as really understanding large training data sets is difficult. But using a tool without understanding its limits and where it can have surprising behavior is dangerous and extremely unprofessional.
Useless metric. The actual scientific advancement from these papers is a tiny faction of what was known before. Science is unfortunately not immune to hypes, but they do not produce much advancement.
Oh, it is nice than I can buy a reasonable (hobbyist use only) 3D printer kit for $300 or so. Gives me something to tinker with. The whole bubble was just exceptionally stupid though.
(Oh, but the way: You, too, are just an automation; at some point, automation becomes so complex that it is indistinguishable from "sentience".)
You can spread your religion-surrogate crap somewhere else please. There is no scientific basis to your claim.
That thing does not exist. Algorithms are improving glacially slow or not at all at this time.
Oh yes. That would help so incredible much. And then make sure all developers, designers and architects either get some real basic understanding on security and have somebody competent they can ask. In most cases, that will be a consultant, as even large enterprises cannot keep in-house experts current. There is just not enough variance in one application landscape. Consultants, on the other hand, see a lot of different situations in a lot of different places.
Of course, the question of getting competent IT security consultants is a pretty hard one as well. There are too many that just work on the surface, with buzzwords and no clear understanding. My minimal list for the technical skills of an IT security consultant is at least intermediary skills in: Networking, system administration (Unix/Linux a must), software development in several languages on several platforms, solid algorithm and data-structures background, secure software development, use of crypto, attack techniques, economics of attacks, distributed systems. Clear writing, ability to explain things, high analytical skills, people skills, etc. are a must as well.
Of course, this is a list where you need real-life "Chief Engineer" with a special interest in security. But nothing much less will do the job and even that Chief Engineer will regularly need to discuss things with some equally qualified colleagues. My advice: Ignore all the large consultancies (in particular also IBM), because while they may have a few such people, they will not give you access. Then look for non-flashy, matter of fact small ones. A significant number of PhDs in IT Security related fields from a good university and is good indicator of skill, but only when coupled with hands-on skills from the above list. Of course, these people will be expensive and will not have a lot of time. But hiring them will be very much worth it.
Indeed. As this is published now, it is in the "likely, but not certain" class. This is a call for assistance, may it be with possible flaws in the measurement or mathematics, may it be with more experiments. Also keep in mind that "experiment" can mean some targeted searching though the absolute huge amounts of data a collider like the LHC produces. This is how particle physicists firm up things, this is a global cooperation and that is the only way it works.