next »

[sams]

I have a feeling this world were a man give 25% of his estate to a machine might have some very twisted culture ... if I were one of the nephew I would conclude that Master Po might have ''forbidden'' fun with the droid.

Man-machine love right ?

I hope we get some culture war angle on the subject, plus the whole nonsensical hilarity of the fat guy and spoiled brat, it will be hysterical.

What are you feelings, will the droid be accused of ''seducing'' and indulging master Po to get a share of the inheritance ?

[ZeissIkon]

I have a feeling this world were a man give 25% of his estate to a machine might have some very twisted culture ... if I were one of the nephew I would conclude that Master Po might have ''forbidden'' fun with the droid.

Man-machine love right ?

I hope we get some culture war angle on the subject, plus the whole nonsensical hilarity of the fat guy and spoiled brat, it will be hysterical.

What are you feelings, will the droid be accused of ''seducing'' and indulging master Po to get a share of the inheritance ?

I doubt either the relationship you're imputing or the terms of the will are any more unusual or scandalous than they were in Thomas Jefferson's day -- he freed a number of slaves in his will (at least a couple of which are believed to have been his own children by Sally Hemmings), and included a complete set of joiner's tools in the inheritance of one of those (the tools were the son/slave's, but as such legally belonged to the master, and the value of the tool chest and contents was, at the time, greater than the value of a skilled adult male slave).  In fact, given androids (at least on Huŏxīng) in the role slaves played in America in the 18th and early 19th century (as opposed to the "farm machinery and livestock" role they were placed in after the cotton gin made them tens of times more productive, by economic measures), I'd expect there's a lot of both going on, though both practices may be outside socially accepted norms.

[terry_freeman]

In the last panel, there were some shadowy figures listening to Po and Veronica's conversation - although I wonder if the conversation is "telepathic" or "electromagnetic".

These might be the nephews, who are perhaps not inclined to give up any part of "their" inheritance to a "mere" robot.

500 years from now, I expect that robots will be as capable as humans, if not more so. Like the fictional Commander Data, or Watson of recent Jeopardy fame, they'll have access to vast stores of information. Unlike Watson, it is to be hoped that they'll deal with human ambiguity.

Kurzweil posits a Singularity, when computer intelligence will vastly exceed that of humans, and will be directed toward improving its own capabilities. I can readily see computer hardware becoming faster, more compact, and more complex, to the point where it rivals the human brain on those terms.

However, I've been watching AI for about 35 years now, and one crucial element will be needed for the Singularity to occur: computer software will need to become introspective, able to make fundamental changes to the way it solves problems. The present course for programs - whether they play Chess, Go, or Jeopardy - is for humans to create a program; test it; find some cases where it breaks down and behaves as if it were incompetent; the humans then tweak the program and repeat the cycle.

Computer-driven learning presently is unable to learn from failure in the manner that humans do.

[sams]

Computer-driven learning presently is unable to learn from failure in the manner that humans do.

I think the computers will always have the advantage of knowing specific stuff, have large memories and processing fast. But there is still a huge gap to being humans and learning ... not just store data but being able to think it throug.

[Plane]

I like the idea behind" The Ghost in the Shell" that as time goes on the diffrence between augmented Humans and intelligent machines will become a fine line.

It might be handy to have a computer or two built in to my system , a few titanium bones, ...

But would my computer ever want to have some "wetware" installed?

[terry_freeman]

I like the idea behind" The Ghost in the Shell" that as time goes on the diffrence between augmented Humans and intelligent machines will become a fine line.

It might be handy to have a computer or two built in to my system , a few titanium bones, ...

But would my computer ever want to have some "wetware" installed?

Maybe. Think about Commander Data's quest for humanity; or consider the "ability to make radical leaps without going nutso" ability which (some) humans have - the thing which allows human programmers to improve today's programs by making radical changes.

Some argue that computers can't be creative. I don't buy that argument, but I will say that our track record at devising computers which can be creative while still retaining a good sense of direction is poor, so far. So, maybe a fusion of computer and wetware would have some utility. 

One of my hobbies is the game of Go, which still represents a serious challenge to computers. So far, the best computers are perhaps an amateur 3 dan - if playing blitz against humans. This is a very respectable accomplishment, but when humans have more time to think, computer programs do not fare so well. When today's programs are losing, their ability falls over a cliff; they play very, very badly. At the moment, this fragile behavior seems to be built in to the algorithms used.

People who are very good at writing such programs are (so far) at a loss to add the top-level analysis and introspection which humans normally bring to the game, which allows humans to improve their performance and to deal gracefully with adverse situations.

[ZeissIkon]

I wonder if the conversation is "telepathic" or "electromagnetic".

I figured the caret-brackets were there to denote translated conversation in some dialect of Chinese.  Nothing to say the equivalent of verbal telepathy is out of reach for someone with Seamus's resources -- I figure we're less than twenty years from being able to "type" by thought alone (there are already demonstrations around of monkeys controlling a robotic arm directly, with their brains, and humans performing the equivalent of flipping a switch or moving a joystick solely with an EEG hookup); do that quickly, or upgrade to intercepting and transmitting intent to verbalize, and you have a form of telepathy -- I think there's a better than even chance of seeing this in my lifetime (which is most likely more than half over -- I'm past fifty).  Not much distinction, at this point, between "telepathy" and "digital", since the data is almost certain to be transferred digitally (think Bluetooth 5.1).

Think about Commander Data's quest for humanity; or consider the "ability to make radical leaps without going nutso" ability which (some) humans have - the thing which allows human programmers to improve today's programs by making radical changes.

Some argue that computers can't be creative. I don't buy that argument, but I will say that our track record at devising computers which can be creative while still retaining a good sense of direction is poor, so far. So, maybe a fusion of computer and wetware would have some utility.

Worth remembering that Data is a fictional character; we don't have any real idea what an android might actually be in another three or four hundred years (for TOS and TNG/DS9/Voyager eras, respectively) -- mentally, I suspect they might be an example of what John W. Campbell used to ask his authors to make their aliens: beings that "think as well as humans, but differently from humans."  Watson actually seems encouraging to me; one of the hardest problems in making computers think "like" humans has been the slippery flexibility of language, where one word can mean half a dozen different things, and multiple words may sound so much alike that it's only possible to determine which was intended from context (and not always then, even for humans who are native speakers of the language in use).  It appears Watson is capable of interpreting natural language about as well as a first grader -- which is a huge step forward from speech recognition (or even text parsing) as it existed as little as five years ago.  Connect this language interpretation with a database optimized toward interactive conversation rather than a trivia quiz with an odd answering requirement, and we could well have a machine capable of passing the Turing test in a matter of a couple more years.  There's already, today, a contract in place to repurpose Watson to provide advice to physicians -- more along the lines of giving them a nudge in case they forget to ask an important question or order a useful test, than actual diagnosis, but making this widespread could improve the quality of medical care significantly (for those who actually manage to get to see a doctor when they need one, at least).

[terry_freeman]

I believe that Watson is did not actually do speech recognition, but parsed the electronic ( typed ) form of the question. It did an impressive job, however. The really interesting part was getting information from a massive free-form database. I believe it worked from encyclopedia entries and dictionaries and so forth, not from hand-coded translations thereof; therefore it had to translate immense amounts of text on the fly. I could be wrong - Watson may have pre-translated those databases in advance of the show. Impressive either way - how do you decide what is important in such a mass of text?

Kurzweil is noted for his work on speech recognition, which is getting quite good nowadays. Within my lifetime, I expect a competent secretary to provide good transcripts and captions of whatever conversations or audio/video recordings I may be interested in. Right now, the electronic translations are sort-of-there, of about the same quality as the googlish one gets for an automatic translation of another language. Sometimes they're perfect, sometimes just perfectly hilariously wrong.

What do human translators bring to the table, that machine efforts don't? Perhaps it's that "this translation makes no sense, therefore I must learn something" detector.

[spudit]

Hopefully future translator software wilt bee smarter that today's spill czech software.

[ZeissIkon]

Hopefully future translator software wilt bee smarter that today's spill czech software.

I've said for years that a spell checker is useless without a grammar checker -- otherwise, it turns minor typing errors into heinous grammatical faux pas.  Lots of folks have said this, but I said it before 1990 (yes, spell checkers existed then; I had one on my Color Computer III).  Wrong Word Spelled Right makes you look stupid much faster than an error that anyone familiar with a keyboard will recognize as a typing mistake.

As for machine speech recognition, it still has trouble with even minor speech impediments (my own speech, without teeth, gives some speech recognition systems fits even though it's perfectly understandable to a human -- and to better speech interpreters), though I suppose that will continue to improve.  Hook up a top quality speech-to-text system to a grammar checker (to insert at least a primitive form of "that doesn't make sense"), then to a natural speech interpreter like Watson's, then put a text-to-speech on the output, and you'll have a computer that can carry on a voice conversation intelligibly; given some money, this could exist by the end of this year -- and if you give Watson a more general database than a couple encyclopedias and a thick dictionary, you might manage a Turing pass with just that (though I doubt it -- Watson seems optimized to give short answers to short questions).

[terry_freeman]

Watson was designed to give Jeopardy-style answers to Jeopardy-style questions.

It would completely botch a Turing Test.

[Scott]

Indeed, I think we're going to be seeing a lot of task-specific AIs before a truly generalized AI shows up.

[spudit]

Lots of stories about task specific AIs out there.

I remember seeing these here movie once where they put one in charge of a spaceship called Discovery and sent it off to Jupiter to check out some kind of shiny black intersteller cell tower. The AI, really nice voice but it had some issues, got kinda moody and things got very bad then very, very wierd.

Seemed to have a happy ending though.

[ZeissIkon]

Seemed to have a happy ending though.

Did we see the same movie?  I thought the ending was completely impenetrable, even though I'd finished reading the novel literally hours before seeing the movie for the first time and knew what was supposed to be happening.  Kubrick being arty, I guess.  Still, I won't let the last five or ten minutes ruin the movie for me -- watching the Orion shuttle dock with the space station to the Blue Danube was sufficient to make the whole movie for me (and, except for that last ten minutes, it was actually a very good movie, IMO).

[quadibloc]

I thought that the one who sang to him, and was to recieve money and manumission papers, was human - and a slave. But I fear she is about to meet a sad fate from the people who were about to kill Po had he not died naturally. (Or who did kill him, if he died from something like poisoning.)

I don't think the eastern half of Earth is free, although it hasn't taken slavery to the levels of perversion the western half has. People are enslaved, but they do not have their humanity taken away.