Nothing more can be done. Stupid people will do stupid things, even when warned repeatedly. Excessive warnings may even be counter-productive, because all those Dunning-Kruger sufferers out there will assume that this large and well-visible warning most certainly is a warning only for stupid people and hence cannot apply to _them_.
One compile-time warning (for interpreted languages: a developer flag or the like), one warning in the documentation and that is enough. More warnings will not get more people to understand what is going on.
These bugs are mostly irrelevant. The ones that matter is where the programmer did not understand the code he/she wrote. And that is a problem before the keyboard.
Most people are stupid and do not recognize the importance and validity of negative results. Unfortunately, scientific publishing has adopted that unscientific attitude and today things get tried over and over again because nobody managed to publish that they do not work.
Simplicity is not definable. It is a judgment call, i.e. it requires an intelligent and wise person to make the determination. Most people do not qualify as either. (Yes, that is a judgment call as well...)
Not only according to Quantum Mechanics. There are millions of people that play the lottery. And you know what, some even win! It is still a gross waste of money to do so, but countless people that do not understand probabilities do not realize that.
Very, very wrong. While Occam's Razor can be abused to justify Reductionism, it is not Reductionism in itself. It merely is the fundamental scientific principle that one should look at (and prove or disprove) simple explanations first in order to be efficient. It leads to other derived sound scientific principles, for example "extraordinary claims require extraordinary proof", which serves to reduce scientific fraud and errors.
Science is not just hard facts. Science is a mind-set, and a set of approaches and principles on how to arrive at hard facts. And sometimes it is soft facts, because nothing better can be obtained with the resources available.
Deal with it and stop misusing "Science!" as a surrogate religion. That violates the idea of science and leads to bad scientific practices.
Example soft fact: "God does almost certainly not exist. [Dawnkins]" That is science. The hard statement "God does not exist" is not scientifically viable and qualifies only as religion as there is now way to prove that. It is still a very reasonable assumption (but only that), and the main way to arrive at it is by a variation of Occam's Razor: "If all simple ways to prove a claim up to a certain, reasonably high complexity level have been exhausted without result, it is reasonable to assume the claim is almost certainly untrue." As Quantum mechanics show, "reasonable high complexity" depends on the circumstances and can be very high in practice.
While Occam's Razor is decidedly a good scientific principle, it cannot be used to prove anything. It is most useful for formulating hypotheses and prioritizing competing research directions.
However, a language that makes it too easy to write code will always have this problem. I tend to not blame the language for that, but the causality is there.
You can write secure code in any almost any language (unless the run-time system is insecure, see for example the history of Java), and you can write insecure code in any language (yes, even in Rust, Swift and Go and other newfangled but not really better hype-languages). The difference is not the language. The difference are the people doing architecture, design and implementation. If some languages have more security problems, that is primarily because these languages attract less competent coders.
Incidentally, absolute numbers are irrelevant. What we need is issues per application.
No, we do not know how to build them. We know how to push a pre-made button and then watch in awe. Building something is a bit more than to tell some mechanism to do it for you. Or do you think telling a contractor to build a house for you means that you built it? Anyways, you completely missed the point.
No so. I did identify you as not intelligent. I win either way. Again, some actual intelligence required to see that, so I guess you will prattle on meaninglessly.
a) But can you, really? From what I've seen, nobody has truly defined intelligence yet.
There is no hard definition, but there are pretty good descriptions by way of what it can do. Lets call it a "working definition" subject to improvement once we know more.
b) We don't "know" that, but because of a) the discussion becomes meaningless. I'll grant you, for creating a "true AI", creating an "AI consciousness" will probably help alot.
If that is possible. Somehow I doubt it, because while intelligence can at least be described by its effects, consciousness is even more difficult as it seems it requires experiencing it in order to understand what it is about.
Kudos for your last point. Reductionism has huge limitations and worshipping it has only rarely provided breakthroughs, inventions and discoveries, if ever. It seems those who glorify it are clinging to what is known, thus unable to make up and validate new knowledge. Equating reductionism with science is ignorance.
Thanks. And yes, I agree that fear of the unknown seems to be a major cause of this mind-set.
Cloning a thing by triggering the built-in replication mechanisms and building a thing from scratch is quite different. It does take some minimal actual intelligence to see that though. I guess you are lacking that.
And do you have any basis for these grand claims other than your own delusions and wishes? You know, maybe something that qualifies as "fact" in the scientific sense?
Funny study done a while ago (lost the reference, sorry): These "electro-sensitive" people cannot distinguish wireless routers going full power from empty cases with LEDs set to simulate it being an active wireless router.
The condition is real, but it is a psychological one, no connection to actual EM activity.
Indeed. Faking it does work for low-quality interaction with human beings (gaming, advertising, etc.) as a) humans are willing to add their own imagination to the mix when wanting to be entertained and b) most humans do not have that much natural intelligence anyways.
However, faking it does not allow AI to ever exceed anything mediocre-skilled human beings already can do. Sure, for a lot of menial tasks that is nice. but the proper term for these machines is "automation", not AI.
Hence AI is not moving "in the wrong direction", it is moving in the only direction known. The two things we know for intelligence are
a) we can describe what it can do and
b) we know that it only manifests itself together with consciousness.
Especially the second thing gets routinely ignored. But everything we know indicates that it is either a critical element, or that, in fact, intelligence and consciousness may be aspects of the same thing. Given this, it is not really a surprise that we do not even have a mathematical model how intelligence could be created artificially. It is quite possible that true/strong AI is quite infeasible in this universe.
And to all physicalists: Your believe that everything observable this universe is physical (and hence it must be possible to create artificial intelligence) is religion, not science. Science is far more open-minded than you are.
Slashdot Mirror
Nothing more can be done. Stupid people will do stupid things, even when warned repeatedly. Excessive warnings may even be counter-productive, because all those Dunning-Kruger sufferers out there will assume that this large and well-visible warning most certainly is a warning only for stupid people and hence cannot apply to _them_.
One compile-time warning (for interpreted languages: a developer flag or the like), one warning in the documentation and that is enough. More warnings will not get more people to understand what is going on.
These bugs are mostly irrelevant. The ones that matter is where the programmer did not understand the code he/she wrote. And that is a problem before the keyboard.
Most people are stupid and do not recognize the importance and validity of negative results. Unfortunately, scientific publishing has adopted that unscientific attitude and today things get tried over and over again because nobody managed to publish that they do not work.
Simplicity is not definable. It is a judgment call, i.e. it requires an intelligent and wise person to make the determination. Most people do not qualify as either. (Yes, that is a judgment call as well...)
Not only according to Quantum Mechanics. There are millions of people that play the lottery. And you know what, some even win! It is still a gross waste of money to do so, but countless people that do not understand probabilities do not realize that.
Very, very wrong. While Occam's Razor can be abused to justify Reductionism, it is not Reductionism in itself. It merely is the fundamental scientific principle that one should look at (and prove or disprove) simple explanations first in order to be efficient. It leads to other derived sound scientific principles, for example "extraordinary claims require extraordinary proof", which serves to reduce scientific fraud and errors.
Science is not just hard facts. Science is a mind-set, and a set of approaches and principles on how to arrive at hard facts. And sometimes it is soft facts, because nothing better can be obtained with the resources available.
Deal with it and stop misusing "Science!" as a surrogate religion. That violates the idea of science and leads to bad scientific practices.
Example soft fact: "God does almost certainly not exist. [Dawnkins]" That is science. The hard statement "God does not exist" is not scientifically viable and qualifies only as religion as there is now way to prove that. It is still a very reasonable assumption (but only that), and the main way to arrive at it is by a variation of Occam's Razor: "If all simple ways to prove a claim up to a certain, reasonably high complexity level have been exhausted without result, it is reasonable to assume the claim is almost certainly untrue." As Quantum mechanics show, "reasonable high complexity" depends on the circumstances and can be very high in practice.
While Occam's Razor is decidedly a good scientific principle, it cannot be used to prove anything. It is most useful for formulating hypotheses and prioritizing competing research directions.
If fully agree. The problem is incompetent and semi-competent people "quickly kludging new functionality into fairly simple web pages".
Indeed. PHP coders suck (on average).
However, a language that makes it too easy to write code will always have this problem. I tend to not blame the language for that, but the causality is there.
You can write secure code in any almost any language (unless the run-time system is insecure, see for example the history of Java), and you can write insecure code in any language (yes, even in Rust, Swift and Go and other newfangled but not really better hype-languages). The difference is not the language. The difference are the people doing architecture, design and implementation. If some languages have more security problems, that is primarily because these languages attract less competent coders.
Incidentally, absolute numbers are irrelevant. What we need is issues per application.
No, we do not know how to build them. We know how to push a pre-made button and then watch in awe. Building something is a bit more than to tell some mechanism to do it for you. Or do you think telling a contractor to build a house for you means that you built it? Anyways, you completely missed the point.
No so. I did identify you as not intelligent. I win either way. Again, some actual intelligence required to see that, so I guess you will prattle on meaninglessly.
And do learn C for when you actually need performance and/or a small memory footprint. Embeds well into Python, BTW.
a) But can you, really? From what I've seen, nobody has truly defined intelligence yet.
There is no hard definition, but there are pretty good descriptions by way of what it can do. Lets call it a "working definition" subject to improvement once we know more.
b) We don't "know" that, but because of a) the discussion becomes meaningless. I'll grant you, for creating a "true AI", creating an "AI consciousness" will probably help alot.
If that is possible. Somehow I doubt it, because while intelligence can at least be described by its effects, consciousness is even more difficult as it seems it requires experiencing it in order to understand what it is about.
Kudos for your last point. Reductionism has huge limitations and worshipping it has only rarely provided breakthroughs, inventions and discoveries, if ever. It seems those who glorify it are clinging to what is known, thus unable to make up and validate new knowledge. Equating reductionism with science is ignorance.
Thanks. And yes, I agree that fear of the unknown seems to be a major cause of this mind-set.
Cloning a thing by triggering the built-in replication mechanisms and building a thing from scratch is quite different. It does take some minimal actual intelligence to see that though. I guess you are lacking that.
I think this is indeed all about grant money.
What we do know to do with gene-therapy is not even remotely on the level of the sorcerer's apprentice.
Give this question another 50 years, and maybe we will know a bit more and maybe even be able to make predictions that have some merit.
And do you have any basis for these grand claims other than your own delusions and wishes? You know, maybe something that qualifies as "fact" in the scientific sense?
We know no such thing. And no real scientists claims we do. Even live cannot be created artificially at this time.
Funny study done a while ago (lost the reference, sorry): These "electro-sensitive" people cannot distinguish wireless routers going full power from empty cases with LEDs set to simulate it being an active wireless router.
The condition is real, but it is a psychological one, no connection to actual EM activity.
Indeed. And basically almost none of these ideas will ever pan out and the few that do will not do so anytime soon. Swarm-stupidity at work.
You just failed Physics 101. Speed has no relation to g-forces, unless you are going in a curve.
Sounds like nonsense to me as well. Like the whole project.
You just have fallen for this fallacy:
https://www.washingtonpost.com...
Indeed. Faking it does work for low-quality interaction with human beings (gaming, advertising, etc.) as a) humans are willing to add their own imagination to the mix when wanting to be entertained and b) most humans do not have that much natural intelligence anyways.
However, faking it does not allow AI to ever exceed anything mediocre-skilled human beings already can do. Sure, for a lot of menial tasks that is nice. but the proper term for these machines is "automation", not AI.
Hence AI is not moving "in the wrong direction", it is moving in the only direction known. The two things we know for intelligence are
a) we can describe what it can do and
b) we know that it only manifests itself together with consciousness.
Especially the second thing gets routinely ignored. But everything we know indicates that it is either a critical element, or that, in fact, intelligence and consciousness may be aspects of the same thing. Given this, it is not really a surprise that we do not even have a mathematical model how intelligence could be created artificially. It is quite possible that true/strong AI is quite infeasible in this universe.
And to all physicalists: Your believe that everything observable this universe is physical (and hence it must be possible to create artificial intelligence) is religion, not science. Science is far more open-minded than you are.