Are AI abilities limitless?

[swansont]

11 minutes ago, internetcynic said:

 Now lets make an observation which, when you look closely is startlingly true: Science fiction has a habit of coming true. Look at, for example Star Trek, from the 1960's. Almost every bit of 'space' technology they had, is now in use, or under development, - communicators, tricorders, matter transportation, sub light drives, cloaking, even warp engines (the Alcubierre drive).

Same name, perhaps, but not the same actual application, and not the same mechanism.  Quantum teleportation, for instance, is nothing like the teleporters on Star Trek.

At the very basic level, all it requires is that technology become more advanced, and some of what seemed to be science fiction will become technology. 

[internetcynic]

13 minutes ago, swansont said:

Same name, perhaps, but not the same actual application, and not the same mechanism.  Quantum teleportation, for instance, is nothing like the teleporters on Star Trek.

At the very basic level, all it requires is that technology become more advanced, and some of what seemed to be science fiction will become technology. 

well, semantics. How do you know Quantum teleportation wont be scaled up at some stage, with the arrival of current unknown new knowledge? Fact is, you dont.  The past evidence suggests it may well be. thats preciesly whats being worked on in some Labs

 

Im open to all ideas, but the one statement  i will mock, set fire to an wee  on is a claim somthing is 'impossible' or will never happen. You have no knowledge of future knowledge that may change the ball game.  Even FLT, i believe wil have a workround. Espcially if as  i espouce elsewhere on here, time doesnt exist, its merely an illusion.

Edited April 9, 2019 by internetcynic

[Strange]

10 minutes ago, internetcynic said:

well, semantics. How do you know Quantum teleportation wont be scaled up at some stage, with the arrival of current unknown new knowledge?

Semantics, yes: the meaning of terms is important. Quantum teleportation is not the same thing as Star-Trek style teleportation. Scaling it up would not change that.

12 minutes ago, internetcynic said:

Im open to all ideas, but the one statement  i will mock, set fire to an wee  on is a claim somthing is 'impossible' or will never happen.

No one will ever square a circle or trisect an angle (using only a finite number of steps with compass and straightedge). It is impossible.

No one will ever find the final digit of pi, because it is impossible.

(Just trying to drag this vaguely back on topic)

Also, just because some things that were thought to be impossible were later achieved doesn't mean that everything is possible. Some things may always be impossible.

[swansont]

18 minutes ago, internetcynic said:

well, semantics. How do you know Quantum teleportation wont be scaled up at some stage, with the arrival of current unknown new knowledge? Fact is, you dont.  The past evidence suggests it may well be. thats preciesly whats being worked on in some Labs

Fact is, I do. Quantum teleportation does not involve moving matter from one point to another. A different beast than what happens on Star Trek.

18 minutes ago, internetcynic said:

Im open to all ideas, but the one statement  i will mock, set fire to an wee  on is a claim somthing is 'impossible' or will never happen.

Conservation of energy is a pretty solid barrier to things being possible or impossible. It literally requires that the laws of physics change over time.

18 minutes ago, internetcynic said:

You have no knowledge of future knowledge that may change the ball game.  Even FLT, i believe wil have a workround. Espcially if as  i espouce elsewhere on here, time doesnt exist, its merely an illusion.

Opinions on time are a dime a dozen, and, in my experience, end up being philosophy and having little to do with physics. At which point I don't care very much.

[DrP]

One of the hardest Star Trek inventions to replicate would be the Universal Translator.  It sounds easy...   it IS easy if you are using it for know languages. But to have it interpret a totally alien language never heard before that could have a language built on totally different rules and make ups to what we know as language and turn it instantly into English is one of the more far fetched inventions on the program imo... I'd go as far as to think it totally impossible. Even after warp and transportation devices the universal translator is pure fiction I would think (for a first contact anyway).

 

 

Edited April 9, 2019 by DrP

[Prometheus]

1 minute ago, DrP said:

One of the hardest Star Trek inventions to replicate would be the Universal Translator.  It sounds easy...   it IS easy if you are using it for know languages. But to have it interpret a totally alien language never heard before that could have a language built on totally different rules and make ups to what we know as language and turn it instantly into English is one of the more far fetched inventions on the program imo... I'd go as far as to think it totally impossible. Even after warp and transportation devices the universal translator is pure fiction I would think (for a first contact anyway).

And the Star Trek writers weren't really trying to predict the future with universal translators - it would just be laborious to have every episode include that initial confusion. There are a few episodes in which translators don't work when they do want to explore the language barrier in cross cultural communication.

Also given the breadth of Sci fi it's reasonable to say a great many, and conflicting, outcomes have been 'predicted'. For every sci-fi that 'predicts' faster than light travel, another 'predicts' sub-light speed travel or worm holes or... 

[CharonY]

!

Moderator Note

Considering that OP is about AI, can we please go back to that topic rather than other Sci-Fi inventions? Open a separate thread if there is a desire to explore this area further. Thank you very much.

 

[swansont]

1 hour ago, internetcynic said:

 In  1890 you could not have predicted cars, airplanes, space rockets,  world war 1 and 2, electronics, sattelites,  mobile phones, television the internet , computers, heart surgery , genetic engineering, nanotechnology, lasers, missiles,  LED's, transistors, ......and on and on.

Benz invented the first auto in ~1885, so that first one is obviously false. People were thinking about flight (they must have, to work on it) before being successful, so it seems dubious that someone could not have predicted them in 1890.

Seeing as you have counted science fiction as a predictor of future tech, and not limiting yourself to exact matches, you have to acknowledge that space rockets existed in the genre before 1890. That should include satellites, too, as being "predicted" 

War was nothing new, so the prediction of WWI and WWII are about when and who, and that's not technology. Certainly predicting that war would come again is a no-brainer.

Missiles existed hundreds of years before 1890, in the form of fireworks. (In the US, we know of the "rockets' red glare" from the war of 1812, mentioned in the Star-Spangled Banner)

I will cede things depending on QM, since that's weird and could not be predicted. Lasers, LEDs and nanotech. DNA hadn't been discovered, so certain genetic engineering is out, but OTOH you have Frankenstein. Which might cover heart surgery as well.  

So I'd say you are about half right. 

[wtf]

12 hours ago, internetcynic said:

I  love statements that assert something is 'impossible'. Such statement are utter bollox.

The unsolvability of the Halting problem is a mathematical theorem. It couldn't be falsified in the future anymore than the Pythagorean theorem could. It's not an empirical result. 

https://en.wikipedia.org/wiki/Halting_problem

[Sensei]

17 hours ago, DrP said:

One of the hardest Star Trek inventions to replicate would be the Universal Translator.  It sounds easy...   it IS easy if you are using it for know languages. But to have it interpret a totally alien language never heard before that could have a language built on totally different rules and make ups to what we know as language and turn it instantly into English is one of the more far fetched inventions on the program imo... I'd go as far as to think it totally impossible. Even after warp and transportation devices the universal translator is pure fiction I would think (for a first contact anyway). 

Google Translator A.I. is fed by original sentence, and translated sentence by humans. From these two versions of the same message but in two or more languages, algorithm is building database. Later client of Google Translator enters sentence, which could be nothing like during learning phase (which is still ongoing). and Google Translator is trying to match corresponding translated equivalent. Problematic are words which means something else depending on context in which they are used though. Or idioms.

How Egyptians hieroglyphics have been decoded? Because there was found stone with the same message written in couple languages which were known.

How two people not knowing their languages are learning and communicating? One is pointing on a thing and says and writes name of a thing, second one is repeating it but in his/her own language. They gain knowledge how different things are called in different languages.

[Strange]

7 minutes ago, Sensei said:

Google Translator A.I. is fed by original sentence, and translated sentence by humans. From these two versions of the same message but in two or more languages, algorithm is building database. Later client of Google Translator enters sentence, which could be nothing like during learning phase (which is still ongoing). and Google Translator is trying to match corresponding translated equivalent. Problematic are words which means something else depending on context in which they are used though. Or idioms.

How Egyptians hieroglyphics have been decoded? Because there was found stone with the same message written in couple languages which were known.

The problem is, in both of these examples, you already have the unknown language and a translation into a known language. 

But humans all learn a language from scratch, knowing nothing about it. The way babies learn is by doing a statistical analysis of sounds to identify morphemes, and then a statistical analysis of how and when those morphemes are used to associate meaning with them.

Which implies that, given time, an AI system could learn a completely new language from scratch and then act as a translator of that language. But it might take two years of intense exposure to the new environment.

14 minutes ago, Strange said:

But humans all learn a language from scratch, knowing nothing about it.

Thinking about it, that may not be completely true. There may be built in assumptions about what it is possible for a language to be like, which reduces the "search space". It isn't clear that a human baby would pick up an alien language based on completely different concepts for encoding and transmitting information.

I suspect linguists would be among the scientists who are most excited if we ever meet an alien race!

[Sensei]

8 minutes ago, Strange said:

The problem is, in both of these examples, you already have the unknown language and a translation into a known language. 

That's because Google Translator A.I. has no cameras nor microphones. Operator is limited by interfaces present in a device. If it would have camera, operator teaching A.I. would show e.g. apple, and say word apple, A.I. would make connection of that image from camera and sound from microphones with object.

Considering that A.I. (with cameras and microphones), could watch simple video prepared for AIs by humans, with thing picture, word name and sound, it could learn ~86400 words per day (limited by duration of sound). In my language there is 3 millions of word (I know because I made dictionary, really troublesome project, exceeding memory on smartphones even such as Google Pixel 2 ?! Heap Size is 512 MB). That's ~35 days long video.

[dimreepr]

36 minutes ago, Sensei said:

That's because Google Translator A.I. has no cameras nor microphones. Operator is limited by interfaces present in a device. If it would have camera, operator teaching A.I. would show e.g. apple, and say word apple, A.I. would make connection of that image from camera and sound from microphones with object.

Considering that A.I. (with cameras and microphones), could watch simple video prepared for AIs by humans, with thing picture, word name and sound, it could learn ~86400 words per day (limited by duration of sound). In my language there is 3 millions of word (I know because I made dictionary, really troublesome project, exceeding memory on smartphones even such as Google Pixel 2 ?! Heap Size is 512 MB). That's ~35 days long video.

9

can it learn dolphin?

[DrP]

39 minutes ago, Sensei said:

If it would have camera, operator teaching A.I. would show e.g. apple, and say word apple, A.I. would make connection of that image from camera and sound from microphones with object.

But an alien 'language' might not have a word for 'Apple'. It might have apples in it's knowledge but the alien word for it could be something like "earth alien reproductive plant growth destroyed and assimilated for energy by alien ingestion' or 'biological nurture pod for land growth organism reproductive mechanism'  -   or it might emit a smell that represents 'an apple' or a 'basic food source'  or 'reproductive offspring nurture pod for land based biological growths'.     

I am sure that AI could eventually figure out any language if there was a 2 way attempt at communication... especially if it was an earth based language (but not in the 2 seconds the universal translator has to work it all out in, for an alien language at that).  

1 hour ago, Strange said:

I suspect linguists would be among the scientists who are most excited if we ever meet an alien race!

The film 'Contact' springs to mind.  :-)  

[Sensei]

1 hour ago, Strange said:

Thinking about it, that may not be completely true. There may be built in assumptions about what it is possible for a language to be like, which reduces the "search space". It isn't clear that a human baby would pick up an alien language based on completely different concepts for encoding and transmitting information. 

If sound emitted by alien would exceed 16-20 kHz human baby would be unable to hear alien..

It's similar issue like with A.I. without proper interfaces ("Operator is limited by interfaces present in a device.")

[Prometheus]

32 minutes ago, DrP said:

The film 'Contact' springs to mind. 

Check out Arrival too - focuses on a linguist trying to decipher a non-lingual alien language.

[Strange]

2 minutes ago, Prometheus said:

Check out Arrival too - focuses on a linguist trying to decipher a non-lingual alien language.

Both great films!

[Sensei]

36 minutes ago, DrP said:

But an alien 'language' might not have a word for 'Apple'.

..you don't have to search for extraterrestrial life forms, to find out that there are many places on the Earth where there is no apple tree, so natives never developed word for "apple" (or whatever else), not present in their region of the world.. instead they adopted word from foreign language..

Edited April 10, 2019 by Sensei

[dimreepr]

3 minutes ago, Sensei said:

..you don't have to search for extraterrestrial life forms, to find out that there are many places on the Earth where there is no apple tree, so natives never developed word for "apple" (or whatever else), not present in their region of the world.. instead they adopted word from foreign language..

5

miss the point, much...

they may have apples but ''it might emit a smell that represents 'an apple''

[Strange]

There was at least one Star Trek episode (Darmok) where the universal translator failed because, even hough it could translate the words that were spoken, no one could understand the translation.

The Tamarians communicate entirely in quotations and references to mythology, so when one of them says something that is translated as "Darmok and Jalad at Tanagra" no one knows what it means.

[DrP]

52 minutes ago, Prometheus said:

Check out Arrival too - focuses on a linguist trying to decipher a non-lingual alien language.

Thanks I'll check it out...  erm, well actually that was the film I was thinking of and thought it was called Contact. lol.    Yea - is a good film.

33 minutes ago, Strange said:

"Darmok and Jalad at Tanagra" no one knows what it means.

"Shaka, when the walls fell"!

 

[Endy0816]

10 hours ago, Strange said:

There was at least one Star Trek episode (Darmok) where the universal translator failed because, even hough it could translate the words that were spoken, no one could understand the translation.

The Tamarians communicate entirely in quotations and references to mythology, so when one of them says something that is translated as "Darmok and Jalad at Tanagra" no one knows what it means.

Something interesting I found in that episode was that the crew of the Enterprise were seen drawing from an existing cultural database of the surrounding region.

This could also explain how the translator always seem to work so fast. The federation, trade partners and allies could all be exchanging information amongst themselves. Logically an Alien species could even be putting their own language database out there to help facilitate contact.

Edited April 10, 2019 by Endy0816