intelligence

For all things philosophical.

Moderators: AMod, iMod

jayjacobus
Posts: 1273
Joined: Wed Jan 27, 2016 9:45 pm

Re: intelligence

Post by jayjacobus »

Someone is manipulating the people's English for their intents and purposes. But that's not my intent in case that is not clear,
jayjacobus
Posts: 1273
Joined: Wed Jan 27, 2016 9:45 pm

Re: intelligence

Post by jayjacobus »

Turing set a criteria for computer experts. But the criteria is not the destination. The destination is somewhere down the road, So Turing's criteria is an important milestone and a difficult one. But don't change the definition of the destination because ordinary people will think the destination has been reached. Don't define the milestone with a destination meaning because that will fool people who don't know what you mean into thinking you have reached the destination. Find a way to define the milestone in whatever glorious terms you want but make it clear that you are only part way to your destination and be honest about your obstacles for reaching the destination.

An artifical thinking machine will have enormous impacts on people and how people will live. Do you claim that is imminent? We need to prepare for that right now.


Are you toolmakers making ever more practical tools or are you Frankensteins planning to make 10 million monsters?
Logik
Posts: 4041
Joined: Tue Dec 04, 2018 12:48 pm

Re: intelligence

Post by Logik »

jayjacobus wrote: Sun Apr 28, 2019 4:45 pm Turing set a criteria for computer experts. But the criteria is not the destination. The destination is somewhere down the road, So Turing's criteria is an important milestone and a difficult one.
Turing's criterion is far more generic than most people realise and ties in directly into the scientific epistemology.

One's ability to ask and answer the question "Is what I have in front of me X" stops with your ability to verify/falsify X.

To move away from the AI discussion with a much simpler example as to the limits of epistemology: do you think you could tell the difference between real water (obtained from a natural source like a river) and synthetic water (e.g water produced in a lab from hydrogen and oxygen).

If you can't contrive an experiment which would allow you to draw such a distinction then what is the difference?

Philosophers often allow themselves to draw distinctions without a difference.
jayjacobus
Posts: 1273
Joined: Wed Jan 27, 2016 9:45 pm

Re: intelligence

Post by jayjacobus »

Logik wrote: Tue Apr 30, 2019 12:43 pm
jayjacobus wrote: Sun Apr 28, 2019 4:45 pm Turing set a criteria for computer experts. But the criteria is not the destination. The destination is somewhere down the road, So Turing's criteria is an important milestone and a difficult one.
Turing's criterion is far more generic than most people realise and ties in directly into the scientific epistemology.

One's ability to ask and answer the question "Is what I have in front of me X" stops with your ability to verify/falsify X.

To move away from the AI discussion with a much simpler example as to the limits of epistemology: do you think you could tell the difference between real water (obtained from a natural source like a river) and synthetic water (e.g water produced in a lab from hydrogen and oxygen).

If you can't contrive an experiment which would allow you to draw such a distinction then what is the difference?

Philosophers often allow themselves to draw distinctions without a difference.
Who is intelligent, the physist with a Phd or the car salesman with a high school education? Does it matter if they get the job done?
AI will get the job done. What job is that? What's the difference between AI and physists?
Logik
Posts: 4041
Joined: Tue Dec 04, 2018 12:48 pm

Re: intelligence

Post by Logik »

jayjacobus wrote: Tue Apr 30, 2019 2:06 pm Who is intelligent, the physist with a Phd or the car salesman with a high school education? Does it matter if they get the job done?
AI will get the job done. What job is that? What's the difference between AI and physists?
That's pretty much it. What is the yardstick for 'intelligence'?

Hint: It is not IQ

Right now the thing that sets AI from GAI apart is goal-setting. A 'truly intelligent' GAI would set and pursue its own goals and objectives.
jayjacobus
Posts: 1273
Joined: Wed Jan 27, 2016 9:45 pm

Re: intelligence

Post by jayjacobus »

People can make one argument and when that doesn't work, they can make another argument. I cannot find a resolution because the arguments keep changing. As long as this keeps going on resolution is not possible.
User avatar
FlashDangerpants
Posts: 6268
Joined: Mon Jan 04, 2016 11:54 pm

Re: intelligence

Post by FlashDangerpants »

jayjacobus wrote: Tue Apr 30, 2019 2:54 pm People can make one argument and when that doesn't work, they can make another argument. I cannot find a resolution because the arguments keep changing. As long as this keeps going on resolution is not possible.
Intelligence is a collection of things, not a singular object. You are complaining because a wide category isn't being distilled into a thimble for you, but it never will be.
Logik
Posts: 4041
Joined: Tue Dec 04, 2018 12:48 pm

Re: intelligence

Post by Logik »

jayjacobus wrote: Tue Apr 30, 2019 2:54 pm People can make one argument and when that doesn't work, they can make another argument. I cannot find a resolution because the arguments keep changing. As long as this keeps going on resolution is not possible.
The arguments keep changing because the problem is ill-defined.

That's a short way of saying: nobody has good answers. Neither to the technical; nor the socio-political challenges.
jayjacobus
Posts: 1273
Joined: Wed Jan 27, 2016 9:45 pm

Re: intelligence

Post by jayjacobus »

FlashDangerpants wrote: Tue Apr 30, 2019 3:21 pm
jayjacobus wrote: Tue Apr 30, 2019 2:54 pm People can make one argument and when that doesn't work, they can make another argument. I cannot find a resolution because the arguments keep changing. As long as this keeps going on resolution is not possible.
Intelligence is a collection of things, not a singular object. You are complaining because a wide category isn't being distilled into a thimble for you, but it never will be.
Is that your analysis or a slur?
jayjacobus
Posts: 1273
Joined: Wed Jan 27, 2016 9:45 pm

Re: intelligence

Post by jayjacobus »

Logik wrote: Tue Apr 30, 2019 3:47 pm
jayjacobus wrote: Tue Apr 30, 2019 2:54 pm People can make one argument and when that doesn't work, they can make another argument. I cannot find a resolution because the arguments keep changing. As long as this keeps going on resolution is not possible.
The arguments keep changing because the problem is ill-defined.

That's a short way of saying: nobody has good answers. Neither to the technical; nor the socio-political challenges.
Some people may want to put a spin on intelligence. I want to take it off by calling AI a substitute for intelligence. There is nothing technical nor socio-political in that.
Logik
Posts: 4041
Joined: Tue Dec 04, 2018 12:48 pm

Re: intelligence

Post by Logik »

jayjacobus wrote: Tue Apr 30, 2019 4:43 pm Some people may want to put a spin on intelligence. I want to take it off by calling AI a substitute for intelligence. There is nothing technical nor socio-political in that.
That depends on whether your definition of AI includes or excludes AGI.

Most people who work in the field draw the distinction between AI and AGI. It's a useful distinction.

We want AI. That's sub-human or par-human intelligence in select domains.
We don't want AGI. That's par-human or super-human intelligence in all domains.

We don't want AGI because we don't know how to solve the problem of control

Don't get me wrong. AI will still put you out of a job, but it's far less likely to take over the word on a whim.
User avatar
Speakpigeon
Posts: 987
Joined: Sat Nov 11, 2017 3:20 pm
Location: Paris, France, EU

Re: intelligence

Post by Speakpigeon »

jayjacobus wrote: Sun Apr 28, 2019 12:40 pm Someone is manipulating the people's English for their intents and purposes. But that's not my intent in case that is not clear,
Talk of AIs smarter than humans is hype to get funding from government agencies.
It's also used to divert the attention of the public from the consequences of using more AIs to do more and more low-skilled jobs. While we're talking about the possibility of AIs smarter than humans and the implications of that for the survival of humanity, we're not talking about the much more immediate and pressing issue of jobs the lost to the AIs that are already being used and that are nowhere near human intelligence, and not even smarter than low-skilled workers.
Those simple AIs are just more machines to replace human labour. Nothing new but we're just reaching a new phase. It's no longer factory jobs that will be affected but also service jobs, and now the question will be that of finding a massive number of new jobs for low-skilled workers. Possible but not necessary for the economy. Only necessary if you want to keep society together.
EB
jayjacobus
Posts: 1273
Joined: Wed Jan 27, 2016 9:45 pm

Re: intelligence

Post by jayjacobus »

You wrote "nobody has good answers". Do you mean me or you?
Logik
Posts: 4041
Joined: Tue Dec 04, 2018 12:48 pm

Re: intelligence

Post by Logik »

jayjacobus wrote: Tue Apr 30, 2019 5:07 pm You wrote "nobody has good answers". Do you mean me or you?
Me, you, people who are actively trying to solve these problems.

There isn't a human on Earth that has good answers when it comes to AI.
User avatar
Speakpigeon
Posts: 987
Joined: Sat Nov 11, 2017 3:20 pm
Location: Paris, France, EU

Re: intelligence

Post by Speakpigeon »

There is of course a real difference between the intelligence of humans and that of AIs. And it is a simple difference.
Humans have a universal intelligence and one that is very sturdy, essentially by taking into account a very large set of data, systematically way beyond what is strictly necessary to solve the problem as narrowly defined. If a human can't get through the door, he will try the window and won't stop until he finds and alternative way of solving the problem.
AIs by comparison have an extraordinarily narrow intelligence. As soon as there is a hiatus, the AIs just stops, unless it fails and kills someone.
Yet, that kind of AIs can already replace actual jobs. And then, where's the limit?
EB
Post Reply