|
Post by wapentake on Mar 29, 2023 20:51:16 GMT
Haven’t I? well your idea that AI is potentially a bee or rock when it has in fact been developed to emulate human thought processes makes it hard if you cannot see that. Are you now agreeing with my starting position? My starting position was that artificial intelligence will only emulate the human desire to dominate if it is programmed to do that. You seem to be agreeing with my starting position when you talk of AI that has been 'developed to emulate human thought processes'. I said that from the start and that the potential to better human thought process is there too ,you only see a machine I see the possibility of a machine that goes beyond that.
|
|
|
Post by Orac on Mar 29, 2023 20:55:53 GMT
I gave you an example - you then switched around and said it doesn't matter because it's just an example of redundancy. The fact that it is redundancy in human understanding was also just hand-waved away.
You realise that education facilities are already seriously spooked by this? There is now (or soon) no sensible way to apportion grades to any kind of intellectual work and there is no way they can see to stop it being a catastrophic problem
|
|
|
Post by Einhorn on Mar 29, 2023 22:10:10 GMT
I gave you an example - you then switched around and said it doesn't matter because it's just an example of redundancy. The fact that it is redundancy in human understanding was also just hand-waved away. You realise that education facilities are already seriously spooked by this? There is now (or soon) no sensible way to apportion grades to any kind of intellectual work and there is no way they can see to stop it being a catastrophic problem I rightly waved it away. Machines have been making people redundant for centuries. We were talking about machines taking over, dominating human beings. They are two different conversations. And universities still apportion grades on the basis of written exams at undergraduate level. Anti-plagiarism software is available to check assignments.
|
|
|
Post by Bentley on Mar 30, 2023 9:07:15 GMT
There are possibly endless ways that ‘ robots’ might over the world and only way way that they won’t ( just as there are endless ways that aliens might enslave humans and only one way that they won’t). I don’t see how an autonomous AI that is self aware and can reason is incapable of making decisions that humans might see as evil. I can’t see how it cannot develop free will or a sense of self preservation either.
|
|
|
Post by Orac on Mar 30, 2023 9:17:53 GMT
I gave you an example - you then switched around and said it doesn't matter because it's just an example of redundancy. The fact that it is redundancy in human understanding was also just hand-waved away. You realise that education facilities are already seriously spooked by this? There is now (or soon) no sensible way to apportion grades to any kind of intellectual work and there is no way they can see to stop it being a catastrophic problem I rightly waved it away. Machines have been making people redundant for centuries. We were talking about machines taking over, dominating human beings. They are two different conversations. And universities still apportion grades on the basis of written exams at undergraduate level. Anti-plagiarism software is available to check assignments. I don't think you are being charitable enough to my position and that is locking you into a corner. A human race that has outsourced its understanding of the world, has also outsourced its decision making. If you want to maintain that this situation isn't effectively a 'take-over', then you are just messing with semantics. The anti plagiarism processes don't work on Ai material. The material looks like original work (or can easily be made to do so).
|
|
|
Post by Einhorn on Mar 30, 2023 9:31:19 GMT
I rightly waved it away. Machines have been making people redundant for centuries. We were talking about machines taking over, dominating human beings. They are two different conversations. And universities still apportion grades on the basis of written exams at undergraduate level. Anti-plagiarism software is available to check assignments. I don't think you are being charitable enough to my position and that is locking you into a corner. A human race that has outsourced its understanding of the world, has also outsourced its decision making. If you want to maintain that this situation isn't effectively a 'take-over', then you are just messing with semantics. The anti plagiarism processes don't work on Ai material. The material looks like original work (or can easily be made to do so). So, you foresee a future where all research and development is handed over to machines, all laws are written by machines, all social policy is formulated by machines? What do you mean when you say anti-plagiarism processes don't work on AI material? A masters or Ph.D. thesis is read by people who will be very familiar with the current literature on the topic.
|
|
|
Post by Orac on Mar 30, 2023 12:52:50 GMT
I don't think you are being charitable enough to my position and that is locking you into a corner. A human race that has outsourced its understanding of the world, has also outsourced its decision making. If you want to maintain that this situation isn't effectively a 'take-over', then you are just messing with semantics. The anti plagiarism processes don't work on Ai material. The material looks like original work (or can easily be made to do so). So, you foresee a future where all research and development is handed over to machines, all laws are written by machines, all social policy is formulated by machines? What do you mean when you say anti-plagiarism processes don't work on AI material? A masters or Ph.D. thesis is read by people who will be very familiar with the current literature on the topic. You wont even realise what is happening at firstLet's talk about something close to both our hearts - arguing politics With the current speed of development, i would say it will be less than a year before you can't tell if you arguing here with responses provided by a machine which are drawn from the input of your messages. Within two years, you wont be able to beat the thing, even if it's wrong berceuse the resources it has at its disposal are so large compared to the few dozen inaccurate facts most humans can maintain and bring Political parties will use it to write statements of intent, speeches and to fine tune manifestos / policies - at first.To some extent this has already started. A public discussion is how we decide what to do next - but this process of humans persuading each other will become near meaningless in the next few years.On the PHD front, it can create new discovery now in many domains. Looking for a novel (previously un-noted) pattern in a large amount of data and then create a presentation around it - this can be done. Proper innovation - ie creating a new physical theory is maybe 6 years away. Beyond that it starts to get a bit peculiar
|
|
|
Post by Orac on Mar 30, 2023 18:12:07 GMT
On the other hand, that might not happen
|
|
|
Post by johnofgwent on Apr 12, 2023 12:39:39 GMT
I think it was Robert Winston who on a documentary on child development said if your child has not lied to you by the age of four you have a problem.
He defended his words thus.
For the earliest part of their lives children look upon their mothers as an all seeing all knowing god but as they develop their own sense of their awareness and limitations and crucially conclude their parents have those same limits.
So by the age of four at the very latest you should have experienced a scenario where your little darling does something like tip over an ornament when you are diverted say to an item in the kitchen
You return to find whatever spilled, or broken and accuse them of the act
Up to the point of awareness the child will admit to it as they cannot understand how you would not know.
At the point of awareness they feel they can blame the cat, dog, brother, sister etc. It is a critical stage in development as much as being able to wipe their own arse and you must hold onto that fact.
What you do about the little bugger lying to you Dr Winston did not say…
|
|
|
Post by johnofgwent on Apr 12, 2023 13:13:23 GMT
So, you foresee a future where all research and development is handed over to machines, all laws are written by machines, all social policy is formulated by machines? What do you mean when you say anti-plagiarism processes don't work on AI material? A masters or Ph.D. thesis is read by people who will be very familiar with the current literature on the topic. You wont even realise what is happening at firstLet's talk about something close to both our hearts - arguing politics With the current speed of development, i would say it will be less than a year before you can't tell if you arguing here with responses provided by a machine which are drawn from the input of your messages. Within two years, you wont be able to beat the thing, even if it's wrong berceuse the resources it has at its disposal are so large compared to the few dozen inaccurate facts most humans can maintain and bring Political parties will use it to write statements of intent, speeches and to fine tune manifestos / policies - at first.To some extent this has already started. A public discussion is how we decide what to do next - but this process of humans persuading each other will become near meaningless in the next few years.On the PHD front, it can create new discovery now in many domains. Looking for a novel (previously un-noted) pattern in a large amount of data and then create a presentation around it - this can be done. Proper innovation - ie creating a new physical theory is maybe 6 years away. Beyond that it starts to get a bit peculiar i recall over 15 years ago political parties were keen to assemble databases of their opponent’s faux pas as a weapon to use against them, and several It freelancers made handsome piles of wonga in such travails. So this is hardly new but i take your point, i played around with one of these AI systems a few months ago and i was quite impressed. I have heard quite a lot of utter bullshit spouted by advocates of and evangelists of AI over the years. The current ads for mr zuckerbergs Meta that can’t give him legs or a decent set of eyebrows but on telly is forecast to aid surgeons in clinical diagnosis of joint trouble is singularly amusing. But yeah, i can see the Turing test barrier being broken soon at this rate
|
|