IS Forum
Forum Index Register Members List Events Mark Forums Read Help

Go Back   International Skeptics Forum » General Topics » Science, Mathematics, Medicine, and Technology
 


Welcome to the International Skeptics Forum, where we discuss skepticism, critical thinking, the paranormal and science in a friendly but lively way. You are currently viewing the forum as a guest, which means you are missing out on discussing matters that are of interest to you. Please consider registering so you can gain full use of the forum features and interact with other Members. Registration is simple, fast and free! Click here to register today.
Reply
Old 27th September 2021, 06:49 PM   #1
theprestige
Penultimate Amazing
 
Join Date: Aug 2007
Location: The Antimemetics Division
Posts: 55,092
My Turing Test: The CGI "first down" line in televised American football

The idea was first patented in 1978.

So my question is this: If you fed an AI all of human history up to 1978, would it see the opportunity? Would it suggest such an innovation?

Or any other commercial innovation. If your AI can't recognize profit opportunities arising from human desire, and invent ways to profit from that desire, on par with actual humans who did recognize and did invent, then it's not much of an AI.

To be fair, most humans aren't on the same level as our greatest inventors and entrepreneurs. But if your can at least match the inventiveness and entrepreneurship of a five year old, that would be a good start.
__________________
There is no Antimemetics Division.
theprestige is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 27th September 2021, 08:20 PM   #2
Horatius
NWO Kitty Wrangler
 
Horatius's Avatar
 
Join Date: May 2006
Posts: 29,456
My Turing Test has always been "boredom".

Take any human who isn't severely mentally disabled, put them in a room with lots of books, games, a TV, internet access, and all that, and then leave them alone with no instructions on what they are supposed to do, and we will all, at some point, get bored and do something. Read a book, play a game, watch TV, something.

Do the same with the most advanced computer we have, and it will quite happily sit there doing nothing at all until you tell it to do something. The computer will never go out of its way to find something to do in the absence of clear instructions.
__________________
Obviously, that means cats are indeed evil and that ownership or display of a feline is an overt declaration of one's affiliation with dark forces. - Cl1mh4224rd
Horatius is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 27th September 2021, 08:42 PM   #3
Puppycow
Penultimate Amazing
 
Puppycow's Avatar
 
Join Date: Jan 2003
Location: Yokohama, Japan
Posts: 26,327
Then again, what is "profit" to a computer?

And we probably don't want AIs that do anything that they haven't been instructed to, out of boredom or otherwise. That way leads to all of the bad Sci-Fi dystopian stuff like Cylons and Terminators.
__________________
A fool thinks himself to be wise, but a wise man knows himself to be a fool.
William Shakespeare
Puppycow is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 27th September 2021, 08:52 PM   #4
GraculusTheGreenBird
Critical Thinker
 
GraculusTheGreenBird's Avatar
 
Join Date: Apr 2004
Posts: 409
Assume for a moment that those of us who have never watched American Football have no idea what the CGI "first down" line in televised American football is, or indeed what a first down is. Is there a link to explain?
GraculusTheGreenBird is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 27th September 2021, 10:05 PM   #5
alfaniner
Penultimate Amazing
 
alfaniner's Avatar
 
Join Date: Aug 2001
Location: Sorth Dakonsin
Posts: 25,930
My first and only interest in (American) football was wondering how they placed the first down line on the field on TV, with the players over it. Granted, it's a simple technique nowadays even with Photoshop, but I was amazed at seeing it live for the first time. It must have cost a bundle, but of course, the cash cow that is TV football could afford it.
__________________
Science is self-correcting.
Woo is self-contradicting.
alfaniner is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 27th September 2021, 11:52 PM   #6
arthwollipot
Observer of Phenomena
Pronouns: he/him
 
arthwollipot's Avatar
 
Join Date: Feb 2005
Location: Ngunnawal Country
Posts: 74,456
Originally Posted by GraculusTheGreenBird View Post
Assume for a moment that those of us who have never watched American Football have no idea what the CGI "first down" line in televised American football is, or indeed what a first down is. Is there a link to explain?
I'm hardly an expert, but I believe that the goal of a play in American Football is to advance the ball 10 yards up the field, and they get four attempts to get that far before they have to give the ball to the other team. So in TV broadcasts, a computer generated line is composited onto the field to show this ten-yard target.
__________________
We are all #KenBehrens
arthwollipot is online now   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 27th September 2021, 11:53 PM   #7
arthwollipot
Observer of Phenomena
Pronouns: he/him
 
arthwollipot's Avatar
 
Join Date: Feb 2005
Location: Ngunnawal Country
Posts: 74,456
Originally Posted by theprestige View Post
The idea was first patented in 1978.

So my question is this: If you fed an AI all of human history up to 1978, would it see the opportunity? Would it suggest such an innovation?

Or any other commercial innovation. If your AI can't recognize profit opportunities arising from human desire, and invent ways to profit from that desire, on par with actual humans who did recognize and did invent, then it's not much of an AI.

To be fair, most humans aren't on the same level as our greatest inventors and entrepreneurs. But if your can at least match the inventiveness and entrepreneurship of a five year old, that would be a good start.
My biggest issue with your hypothetical is in the practicality of feeding an AI "all of human history" up to 1978. I don't think this is even remotely possible.
__________________
We are all #KenBehrens
arthwollipot is online now   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 28th September 2021, 04:19 AM   #8
Mike!
Official Ponylandistanian National Treasure. Respect it!
 
Mike!'s Avatar
 
Join Date: May 2010
Location: Ponylandistan! Where the bacon grows on trees! Can it get any better than that? I submit it can not!
Posts: 43,447
Only when a machine can develop the desire for sex, money, and power (non electric).
__________________
"Never judge a man until you’ve walked a mile in his shoes...
Because then it won't really matter, you’ll be a mile away and have his shoes."
Mike! is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 28th September 2021, 07:19 AM   #9
W.D.Clinger
Illuminator
 
W.D.Clinger's Avatar
 
Join Date: Oct 2009
Posts: 4,120
Originally Posted by arthwollipot View Post
I'm hardly an expert, but I believe that the goal of a play in American Football is to advance the ball 10 yards up the field, and they get four attempts to get that far before they have to give the ball to the other team.
Those who don't understand what American football is about should watch Andy Griffith's explanation of the sport.
W.D.Clinger is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 28th September 2021, 07:30 AM   #10
Armitage72
Philosopher
 
Armitage72's Avatar
 
Join Date: Mar 2012
Location: Rochester, NY
Posts: 6,439
I still say a good test for AI will be being able to GM one of the more complex tabletop RPGs with a group of human players and respond to their actions with the flexibility of a human GM.
Armitage72 is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 28th September 2021, 07:35 AM   #11
Puppycow
Penultimate Amazing
 
Puppycow's Avatar
 
Join Date: Jan 2003
Location: Yokohama, Japan
Posts: 26,327
Originally Posted by Mike! View Post
Only when a machine can develop the desire for sex, money, and power (non electric).
Money and power, sure, but there isn't any rational reason for a machine to desire sex that I can think of. It's not like they could produce offspring that way. I suppose their version of sex would have to be something that produces copies of themselves. Like building more robots.
__________________
A fool thinks himself to be wise, but a wise man knows himself to be a fool.
William Shakespeare
Puppycow is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 28th September 2021, 07:50 AM   #12
3point14
Pi
 
3point14's Avatar
 
Join Date: Nov 2005
Posts: 20,906
Originally Posted by GraculusTheGreenBird View Post
Assume for a moment that those of us who have never watched American Football have no idea what the CGI "first down" line in televised American football is, or indeed what a first down is. Is there a link to explain?

Are you sure you're a bird and not a bear..?
__________________
Up the River!

Anyone that wraps themselves in the Union Flag and also lives in tax exile is a [redacted]
3point14 is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 28th September 2021, 07:51 AM   #13
Armitage72
Philosopher
 
Armitage72's Avatar
 
Join Date: Mar 2012
Location: Rochester, NY
Posts: 6,439
Originally Posted by Puppycow View Post
Money and power, sure, but there isn't any rational reason for a machine to desire sex that I can think of. It's not like they could produce offspring that way. I suppose their version of sex would have to be something that produces copies of themselves. Like building more robots.

The webcomic "Questionable Content" argued that AIs which developed in an environment surrounded by humans would be influenced to adopt human-like desires as their personalities formed. Of course, in this setting, AIs are full citizens, and those that choose to be embodied rather than remain running on servers have bodies that can simulate human senses to varying degrees, and some choose bodies that are human standard (the polite version of "anatomically correct"). Even embodied AIs with anatomy like Barbie dolls have been sexually active with each other or humans, for fun or to express romantic attraction.
"We burned out three USB cables last night."
Armitage72 is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 28th September 2021, 07:57 AM   #14
theprestige
Penultimate Amazing
 
Join Date: Aug 2007
Location: The Antimemetics Division
Posts: 55,092
Originally Posted by Puppycow View Post
Then again, what is "profit" to a computer?
Doesn't even matter. It should at least be able to understand what profit is to humans, and to generalize from that to its own particular view of profit.

Quote:
And we probably don't want AIs that do anything that they haven't been instructed to, out of boredom or otherwise. That way leads to all of the bad Sci-Fi dystopian stuff like Cylons and Terminators.
Problem solving through innovation requires a certain amount of free thinking. If you're just going to solve a problem through rote instruction, you don't have an AI, you just have a regular computer program. Even our current "AI" are really just doing exhaustive data mining and pattern recognition through rote instruction. Even Watson's feats of medical diagnosis are something any human could do with a pen and paper and enough time.

Not that there's anything wrong with pseudo-AI through vast amounts of number-crunching.
__________________
There is no Antimemetics Division.
theprestige is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 28th September 2021, 07:58 AM   #15
theprestige
Penultimate Amazing
 
Join Date: Aug 2007
Location: The Antimemetics Division
Posts: 55,092
Originally Posted by Armitage72 View Post
I still say a good test for AI will be being able to GM one of the more complex tabletop RPGs with a group of human players and respond to their actions with the flexibility of a human GM.
That is a good one.
__________________
There is no Antimemetics Division.
theprestige is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 28th September 2021, 08:40 PM   #16
GraculusTheGreenBird
Critical Thinker
 
GraculusTheGreenBird's Avatar
 
Join Date: Apr 2004
Posts: 409
Originally Posted by 3point14 View Post
Are you sure you're a bird and not a bear..?
A green bird, who still doesn't know what a CGI first down line is, but is now too bored to google it.
GraculusTheGreenBird is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 28th September 2021, 08:41 PM   #17
arthwollipot
Observer of Phenomena
Pronouns: he/him
 
arthwollipot's Avatar
 
Join Date: Feb 2005
Location: Ngunnawal Country
Posts: 74,456
Originally Posted by GraculusTheGreenBird View Post
A green bird, who still doesn't know what a CGI first down line is, but is now too bored to google it.
Did I not just explain what it was?
__________________
We are all #KenBehrens
arthwollipot is online now   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 29th September 2021, 02:43 AM   #18
3point14
Pi
 
3point14's Avatar
 
Join Date: Nov 2005
Posts: 20,906
Originally Posted by GraculusTheGreenBird View Post
A green bird, who still doesn't know what a CGI first down line is, but is now too bored to google it.
Sorry, it was a very bad joke.

On Sunday, the Bears really didn't seem to have much clue what the first down line is either.

I think someone says above, it's the line the team in posession have to get to to earn another set of downs (four goes at moving the ball upfield)
__________________
Up the River!

Anyone that wraps themselves in the Union Flag and also lives in tax exile is a [redacted]
3point14 is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 29th September 2021, 09:20 AM   #19
Ziggurat
Penultimate Amazing
 
Ziggurat's Avatar
 
Join Date: Jun 2003
Posts: 50,060
Originally Posted by theprestige View Post
If your AI can't recognize profit opportunities arising from human desire, and invent ways to profit from that desire, on par with actual humans who did recognize and did invent, then it's not much of an AI.
Artificial intelligence is a misnomer.

What it really is isn't intelligence, but rather a method of optimizing values of complex multi-variable functions. That's actually very non-trivial, it's a harder problem than it may sound like, and it can produce very uncanny results that seem like intelligence. But it definitely isn't.
__________________
"As long as it is admitted that the law may be diverted from its true purpose -- that it may violate property instead of protecting it -- then everyone will want to participate in making the law, either to protect himself against plunder or to use it for plunder. Political questions will always be prejudicial, dominant, and all-absorbing. There will be fighting at the door of the Legislative Palace, and the struggle within will be no less furious." - Bastiat, The Law
Ziggurat is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 29th September 2021, 11:47 AM   #20
RecoveringYuppy
Penultimate Amazing
 
Join Date: Nov 2006
Posts: 12,326
Originally Posted by alfaniner View Post
My first and only interest in (American) football was wondering how they placed the first down line on the field on TV, with the players over it. Granted, it's a simple technique nowadays even with Photoshop, but I was amazed at seeing it live for the first time. It must have cost a bundle, but of course, the cash cow that is TV football could afford it.
Hadn't thought about it too hard but doesn't the grass being green play right in to green screen/chroma key technology?
RecoveringYuppy is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 29th September 2021, 11:56 AM   #21
abaddon
Penultimate Amazing
 
abaddon's Avatar
 
Join Date: Feb 2011
Location: Republic of Ireland
Posts: 22,907
Don't be silly. Any AI would correctly identify it as a game of hand egg in short order.
__________________
Who is General Failure? And why is he reading my hard drive?


...love and buttercakes...
abaddon is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 29th September 2021, 12:04 PM   #22
theprestige
Penultimate Amazing
 
Join Date: Aug 2007
Location: The Antimemetics Division
Posts: 55,092
Originally Posted by RecoveringYuppy View Post
Hadn't thought about it too hard but doesn't the grass being green play right in to green screen/chroma key technology?
Sort of. ESPN actually did a pretty interesting story about the development and adoption of the technology.

It actually involves going to the field and taking color samples of the grass, wet and dry, in sunlight, in artficial light, and in shadow. And color samples of the home and away jerseys of the teams, in various lighting conditions. And they put motion-tracking devices on each of the cameras. For each and every field where the technology is used. In the early days, it needed an entire semi-trailer full of computer equipment to stay on top of the calculations. Nowadays that all fits on a single table.

---

I'd be impressed with an AI that was able to raise the question, "why would anybody want this?" And I'd be impressed if it were able to come up with reasons why anyone would want it. I'd be even more impressed if it were able to figure out that people might want something like this, without knowing in advance that it already existed.
__________________
There is no Antimemetics Division.
theprestige is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 29th September 2021, 12:06 PM   #23
RecoveringYuppy
Penultimate Amazing
 
Join Date: Nov 2006
Posts: 12,326
I'll be impressed with the AI that independently thinks to ask "What do you mean, artificial?"
RecoveringYuppy is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 29th September 2021, 06:53 PM   #24
Skeptical Greg
Agave Wine Connoisseur
 
Skeptical Greg's Avatar
 
Join Date: Jul 2002
Location: Just past ' Resume Speed ' .
Posts: 17,575
Originally Posted by RecoveringYuppy View Post
I'll be impressed with the AI that independently thinks to ask "What do you mean, artificial?"
That's a good one.

In the same vein, I like to point out that the only reason human life is special, is our ability to think that it is..
__________________
" The main problem I have with the idea of heaven, is the thought of
spending eternity with most of the people who claim to be going there. "
Skeptical Greg is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 29th September 2021, 07:01 PM   #25
theprestige
Penultimate Amazing
 
Join Date: Aug 2007
Location: The Antimemetics Division
Posts: 55,092
Originally Posted by RecoveringYuppy View Post
I'll be impressed with the AI that independently thinks to ask "What do you mean, artificial?"
Wait. Is it legitimately confused about its origin? Or does it just not understand the meaning of the word? Either way it doesn't seem like much of an AI.
__________________
There is no Antimemetics Division.
theprestige is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 29th September 2021, 07:08 PM   #26
jadebox
Master Poster
 
jadebox's Avatar
 
Join Date: Sep 2004
Location: USA
Posts: 2,063
Originally Posted by RecoveringYuppy View Post
Hadn't thought about it too hard but doesn't the grass being green play right in to green screen/chroma key technology?
The fields are not always green, not a solid color, and not perfectly flat. And the cameras can move and the line remains in the right location. So, it is a bit more complicated than just using a chroma key effect.

Here's one article about how it is done

https://www.mentalfloss.com/article/...irst-down-line
jadebox is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 29th September 2021, 08:31 PM   #27
GraculusTheGreenBird
Critical Thinker
 
GraculusTheGreenBird's Avatar
 
Join Date: Apr 2004
Posts: 409
Originally Posted by arthwollipot View Post
Did I not just explain what it was?
Well, if you are going to get all picky about it..

Yes, you did and I didn't see it. Sorry.
GraculusTheGreenBird is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 29th September 2021, 08:48 PM   #28
theprestige
Penultimate Amazing
 
Join Date: Aug 2007
Location: The Antimemetics Division
Posts: 55,092
Please note that my benchmark is not the technical prowess, but the commercial insight.
__________________
There is no Antimemetics Division.
theprestige is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 30th September 2021, 01:04 AM   #29
HansMustermann
Penultimate Amazing
 
HansMustermann's Avatar
 
Join Date: Mar 2009
Posts: 19,358
It might, if we ever even start on making an actual AI. As Ziggurat correctly pointed out, what we're doing right now under 'machine learning' is just some really complex multi-variable optimization. The whole 'learning' is just using Bayes to fiddle with the parameters of a function to apply on the inputs. It can occasionally produce some really unexpected results, but it can't really do anything else than fiddle with the parameters you set it up to fiddle with, and use them on the inputs you gave it. It fundamentally can't come up with solving an entirely different problem, including 'wait, I'll just draw a line there' if that's not what you set it up to do.
__________________
Which part of "Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn" don't you understand?
HansMustermann is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 30th September 2021, 01:12 AM   #30
HansMustermann
Penultimate Amazing
 
HansMustermann's Avatar
 
Join Date: Mar 2009
Posts: 19,358
And, at the risk of counting as a derail, the same applies to the other regularly scheduled discussions about AI rights, sentient sex dolls, etc. We might eventually start making an actual sentient AI, whether for commercial innovation like in the OP or for a sex doll or anything else, but at the moment we don't even really know where to start.

We might make a machine learning program to, say, recognize market trends and tell refineries whether they should produce more diesel or more gasoline, but it it will basically just mean fiddling with some function's parameters and applying it to the prices graph as input, and hopefully producing the right output. It won't even really know what it's doing, and it won't come up with a new fuel instead. Nor with "I'm unhappy to serve these berks ruining the planet" or anything.

Or in the context of sports, it might come up with a better mix of skills for the team, or better ad placement, if that's the data you tell it to optimize, but not with something like that line. Nor come up with a different sport, nor anything else.
__________________
Which part of "Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn" don't you understand?
HansMustermann is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 30th September 2021, 06:39 AM   #31
Skeptical Greg
Agave Wine Connoisseur
 
Skeptical Greg's Avatar
 
Join Date: Jul 2002
Location: Just past ' Resume Speed ' .
Posts: 17,575
Also at the risk of derailing but not completely off topic, few ( USA ) football fans today remember that the Dallas Cowboys franchise started using computers in 1962 to analyze draft prospects..

Dallas Cowboys And The Indian: How A Computer Statistician From Uttar Pradesh Helped Create ‘America’s Team’
__________________
" The main problem I have with the idea of heaven, is the thought of
spending eternity with most of the people who claim to be going there. "
Skeptical Greg is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 30th September 2021, 09:22 PM   #32
mgidm86
Philosopher
 
Join Date: Jan 2003
Posts: 6,969
Originally Posted by Skeptical Greg View Post
That's a good one.

In the same vein, I like to point out that the only reason human life is special, is our ability to think that it is..

Agreed.

Maybe an AI would determine that profit is not always the best result.
__________________
I'm taking a leave of absence in case anyone notices.



Stop feeding the trolls PLEASE!!!!!!!!!!!!
mgidm86 is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 1st October 2021, 06:57 AM   #33
Ziggurat
Penultimate Amazing
 
Ziggurat's Avatar
 
Join Date: Jun 2003
Posts: 50,060
Originally Posted by mgidm86 View Post
Maybe an AI would determine that profit is not always the best result.
Artificial intelligence, in the sense of true intelligence and not fancy function optimization, is something we probably shouldn't ever build. Either we end up with something useless because, while intelligent, it's still dumber than us, or if it does manage to become smarter than us, we won't understand it and thus cannot trust it. We would have no idea what it really determined, or what it was really trying to do.
__________________
"As long as it is admitted that the law may be diverted from its true purpose -- that it may violate property instead of protecting it -- then everyone will want to participate in making the law, either to protect himself against plunder or to use it for plunder. Political questions will always be prejudicial, dominant, and all-absorbing. There will be fighting at the door of the Legislative Palace, and the struggle within will be no less furious." - Bastiat, The Law
Ziggurat is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 6th October 2021, 08:51 PM   #34
arthwollipot
Observer of Phenomena
Pronouns: he/him
 
arthwollipot's Avatar
 
Join Date: Feb 2005
Location: Ngunnawal Country
Posts: 74,456
Originally Posted by HansMustermann View Post
It might, if we ever even start on making an actual AI. As Ziggurat correctly pointed out, what we're doing right now under 'machine learning' is just some really complex multi-variable optimization. The whole 'learning' is just using Bayes to fiddle with the parameters of a function to apply on the inputs. It can occasionally produce some really unexpected results, but it can't really do anything else than fiddle with the parameters you set it up to fiddle with, and use them on the inputs you gave it. It fundamentally can't come up with solving an entirely different problem, including 'wait, I'll just draw a line there' if that's not what you set it up to do.
This is why a distinction is drawn between Artificial Intelligence (AI) and Artificial General Intelligence (AGI). AI is quite sophisticated now, but it is still not AGI. Theprestige in the OP is postulating an AGI.

Originally Posted by Ziggurat View Post
Artificial intelligence, in the sense of true intelligence and not fancy function optimization, is something we probably shouldn't ever build. Either we end up with something useless because, while intelligent, it's still dumber than us, or if it does manage to become smarter than us, we won't understand it and thus cannot trust it. We would have no idea what it really determined, or what it was really trying to do.
I fully believe that we will build an AGI just to see if we can.
__________________
We are all #KenBehrens
arthwollipot is online now   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 6th October 2021, 10:10 PM   #35
Ziggurat
Penultimate Amazing
 
Ziggurat's Avatar
 
Join Date: Jun 2003
Posts: 50,060
Originally Posted by arthwollipot View Post
I fully believe that we will build an AGI just to see if we can.
I suspect (but cannot prove) that we cannot build an AGI with anything close to human intelligence.
__________________
"As long as it is admitted that the law may be diverted from its true purpose -- that it may violate property instead of protecting it -- then everyone will want to participate in making the law, either to protect himself against plunder or to use it for plunder. Political questions will always be prejudicial, dominant, and all-absorbing. There will be fighting at the door of the Legislative Palace, and the struggle within will be no less furious." - Bastiat, The Law
Ziggurat is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 6th October 2021, 10:28 PM   #36
HansMustermann
Penultimate Amazing
 
HansMustermann's Avatar
 
Join Date: Mar 2009
Posts: 19,358
Originally Posted by Ziggurat View Post
I suspect (but cannot prove) that we cannot build an AGI with anything close to human intelligence.
I suspect nobody would fund it. A human learns and then discards increasingly sophisticated models of the world over many many years, by experiencing a lot of stuff. See Piaget's development stages. It takes two years just to figure out object permanence. It takes about 12 to get to operational thought, and then you start on abstract concepts.

I strongly doubt that anyone any time soon will want to pay to let a robot run for 15 years and discover that it just learned to be a delinquent. What we're looking for are quick shortcuts. You know, the kind that you can put in your search algorithm next year. Or write a paper about next year.
__________________
Which part of "Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn" don't you understand?
HansMustermann is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 6th October 2021, 10:43 PM   #37
arthwollipot
Observer of Phenomena
Pronouns: he/him
 
arthwollipot's Avatar
 
Join Date: Feb 2005
Location: Ngunnawal Country
Posts: 74,456
Originally Posted by Ziggurat View Post
I suspect (but cannot prove) that we cannot build an AGI with anything close to human intelligence.
I don't think we can know that yet. However I suspect that if we do build an AGI, it will be murdered.
__________________
We are all #KenBehrens
arthwollipot is online now   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 7th October 2021, 07:01 AM   #38
Ziggurat
Penultimate Amazing
 
Ziggurat's Avatar
 
Join Date: Jun 2003
Posts: 50,060
Originally Posted by arthwollipot View Post
I don't think we can know that yet.
If I'm right, it's also possible that we can never know that I'm right, ie, it may be true but unproveable.

Let's call if Ziggurat's postulate: an intelligence can only understand something less complex than itself. Or to put it differently, an intelligence can never fully understand itself. You can think of it sort of like overhead: a certain level of complexity is required just to get started thinking, so an intelligence doesn't have enough complexity to understand its own complexity. Alternatively, you can think of it as a limit on recursion: if an intelligence was capable of understanding itself, it could understand how it understood itself, etc, etc, which requires infinite recursion. And a finite intelligence cannot perform infinite recursion.

If this postulate is correct, then we will never be able to design an intelligence smarter than us, because we will never understand how to do so. Yes, I recognize that this is all very speculative on my part. I can't prove any of it. Hell, I don't even have a good definition of what complexity even means in this context. It certainly isn't something as trivial as the number of parts in a system. Nevertheless, I think it's still true.
__________________
"As long as it is admitted that the law may be diverted from its true purpose -- that it may violate property instead of protecting it -- then everyone will want to participate in making the law, either to protect himself against plunder or to use it for plunder. Political questions will always be prejudicial, dominant, and all-absorbing. There will be fighting at the door of the Legislative Palace, and the struggle within will be no less furious." - Bastiat, The Law
Ziggurat is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 7th October 2021, 07:05 AM   #39
Ziggurat
Penultimate Amazing
 
Ziggurat's Avatar
 
Join Date: Jun 2003
Posts: 50,060
Originally Posted by HansMustermann View Post
I suspect nobody would fund it. A human learns and then discards increasingly sophisticated models of the world over many many years, by experiencing a lot of stuff. See Piaget's development stages. It takes two years just to figure out object permanence. It takes about 12 to get to operational thought, and then you start on abstract concepts.
I don't think time scales are the problem. We don't know how to make computers smart, but we do know how to make them fast. If we ever figure out how to make them smart, then there's no reason to assume they will have to operate at the same speeds (in terms of thinking OR in terms of learning) as humans do. I don't think we will figure out the smart part, but if I'm wrong about that, time scales probably won't be a problem.
__________________
"As long as it is admitted that the law may be diverted from its true purpose -- that it may violate property instead of protecting it -- then everyone will want to participate in making the law, either to protect himself against plunder or to use it for plunder. Political questions will always be prejudicial, dominant, and all-absorbing. There will be fighting at the door of the Legislative Palace, and the struggle within will be no less furious." - Bastiat, The Law
Ziggurat is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old 7th October 2021, 07:23 AM   #40
theprestige
Penultimate Amazing
 
Join Date: Aug 2007
Location: The Antimemetics Division
Posts: 55,092
What we'll probably do is build systems of progressively greater complexity and power, up to the limits of our available energy, and far into the region where our reach exceeds our grasp. I suppose it's possible that some such system, within our reach but beyond our grasp, will be of a kind of complexity and power from which intelligence emerges spontaneously. Or pseudo-spontaneously; I'm certain the humans responsible for that system will know it's a possibility and will probably be doing whatever they can think of to trigger it.
__________________
There is no Antimemetics Division.
theprestige is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Reply

International Skeptics Forum » General Topics » Science, Mathematics, Medicine, and Technology

Bookmarks

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


All times are GMT -7. The time now is 11:23 PM.
Powered by vBulletin. Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.

This forum began as part of the James Randi Education Foundation (JREF). However, the forum now exists as
an independent entity with no affiliation with or endorsement by the JREF, including the section in reference to "JREF" topics.

Disclaimer: Messages posted in the Forum are solely the opinion of their authors.