The Watson super computer...Skynet? (1 Viewer)

[Ruz-El]

Originally Posted by DaveF
Jeopardy-playing computer Watson: (yawn) I've got "Siri". IBM spends its fortunes & man-years creating 1960s-esque, house-filling computer for the useless task of playing a specialized game show. But for a year, I've had an app on my *phone* that "understands" natural language and does useful things like finding restaurants. While IBM's doing expensive tech demos, other people are making amazing tools for daily life.

You phone can understand and answer jeopardy like questions? You should put it on the show and win a million dollars!

(your phone wouldn't exist if tech guys didn't attempt to create computers that do dumb stuff like play chess, answer trivia... just so you know. This is how things advance, and a fun way.)

 

[DaveF]

Originally Posted by Russell G http://www.hometheaterforum.com/forum/thread/308878/the-watson-super-computer-skynet#post_3782340
You phone can understand and answer jeopardy like questions? You should put it on the show and win a million dollars!

Uhm...no...I said nothing even suggesting that. What post are you reading?

What I said is that my phone can process natural language in a way that can appear almost human and provides useful results for daily interests. In contrast to "Watson" which does nothing useful from a practical perspective.

R&D is useful and necessary for the creation tools like Siri. But it's not apparent that IBM's machine-vs-man stunts actually further useful technology, so much as it's a giant ad campaign for their corporate IT services.

 

[DaveF]

I'll put it another way: Wolfram is developing a similar sort of technology, which they've made available online and in various computer programs. It's called "Alpha". You ask it a math or physics question in and it attempts to give the answer. It's available in various free and for purchase versions. Presumably they will continue to advance their technology and products, to the benefit of individuals and their profit.

Google has a similar tool. It's called "Google". You ask it questions (in a terse shorthand, typically) and it finds tries to find the answers. It's not limited to Jeopardy-style questions, and can't give a single "correct" answer. But it does a pretty good job and happens to be the most popular and profitable tool of its sort.

If IBM were interested in something more than a publicity stunt, "Watson" wouldn't be made to parse a very specific and quirky type of question and to produce answers in a similarly specific style for a TV gameshow. If IBM had a use for this, if it were generalizable, they'd have launched a product "Watson" to rival Wolfram|Alpha and/or Google.com. If they had a clue what they were doing, beyond outsource IT services and R&D boondoggles, they'd wouldn't be showing it off in one-time TV ratings-grabbers, they'd be creating the next great computer knowledge system to compete with the global giants.

The possible uses are staggering. "Answer.com" with a vengeance. A self-writing "Wikipedia". A remarkably accurate health diagnosis service to knock the socks off of "WebMD.com". And these could be fantastically profitable.

Instead we get a Jeopardy stunt.

I'm unimpressed.

 

[Ruz-El]

Alpha is a complex calculating program. Computers for years where basically giant adding machines, I'm surprised such a program doesn't already exist.

Google is a search engine. It looks at keywords and then presents links to things that might be applicable to those keywords, based on popularity.

Watson looks at keywords, tries to understand what the actual question is based on their relation to each other, and then come up with an answer to that question. Similar to how a human brain works. It's hugely more complex then a google search engine, and is why it's so impressive.

Google puts blocks into predetermined rows, Watson looks at the blocks and figures out the best pattern for those blocks to go in. It also "realizes" what type of information it needs more of to improved, based on getting false answers. Google and search engines, not so much.

 

[Inspector Hammer!]

I just spent the last hour reading various material on Watson and I keep seeing this disturbing pattern...people referring to it as 'he' and 'him'...damn.

In my reading rarely did I come across anything possibly negative about this creation save for the occasional off the cuff reference to Skynet etc, almost everything was hopeful and optimistic even though we can't know what this will hold for us for bad or for good and that, to me, is a tad, I don't know, short sighted?

As I get older I also am becoming increasingly more aware of how much we humans put stock in our own magnificence, congratulating ourselves for creating these amazing things for no other reason than we can, never stopping to consider whether or not we should or what it could mean for us. Hell I'm surprised that Cloning isn't allowed, I'm surprised because for once we had the forethought to stop and ask "is this a good thing? Probably not so we shouldn't do it." and that's not usually in our nature.

Now, and allow me to express this as I, like you, have feelings about this and it is my right, I personally feel that technology such as this can possibly become dangerous just as you feel that it cannot, that's just my viewpoint. Feel free to come back at me with facts and even more optimism but I'm sorry, I must dismiss them simply because it's impossible for any of us to know, including myself, with 100% certainty that this will be harmless to us in 30, 40, 50 or even 100 years.

Sure you may hope that it will help us and be perfectly safe...but you can't know and that's is precisely why we have to proceed more cautiously and with a great degree of reverence of this technology's possibilities. If I'm coming off sounding a lot like Ian Malcolm in Jurassic Park there's a reason for that...he was absolutely right lol. It's very easy for us to sit here right now and say things like "It will never be connected to the internet and it will never be put in charge of anything without human involvement" but if our history has taught us anything it's that our curiosity always gets the best of us. All it will take to change our minds and try to use it that way is a few years of flawless performance before we say "hey, let's try to use it for this or that, it's been great so far".

With AI technology it's as if we're watching a Chimp trying to start a jet plane, we see him in the cockpit fumbling around and am absolutely sure that he won't be able to figure out how to start it and take off and suddenly it does. Now I don't only mean to suggest that such technology could become physically destructive but that it could impact us negatively in other significant ways, in one article I read they talk about how one day we may develop AI so advanced that we'll have androids that think, act and look just like we do and that they could perform jobs more efficiently than we can...and that's a good thing because?

That would mean that machines will replace human workers, putting them out of a job, correct? Explain to me how that could possibly be viewed as good or an advancement of any kind.

The older I get the more spiritual I get and can see how much our advancements in technology actually hurts us a species rather than helps, we become consumed with the blind need to make things faster, better, smarter and we never stop to ask this very simple question...why? Isn't the wonder of the human brain enough for us that we need to create something that outdoes it simply because we have the capacity to?

I don't know, I personally am going to remain wary of this Watson and it's further development and not because some movie told me I should be, I just know that we have a track record of creating incredibly powerful things that end up being used for more harm than good. So I guess you could say that I, too, am fascinated with Watson...just not for the same reasons you are.

Also, from what I understand **IT'S** victory on Jeopardy wasn't all that impressive when you consider that it all came down to how quickly it responded with the buzzer, it didn't win because it "knew" more but because it could buzz in faster than the humans and that, IMO, gave it an unfair advantage.

Okay, rant over, if you read the whole thing thank you and if you think I'm a paranoid nutjob that assumes the worst than you'd be right. I've never been neither a glass is half empty or half full type of person, I've always been one to question the stability and integrity of the glass that holds the liquid.

 

[nolesrule]

Also, from what I understand **IT'S** victory on Jeopardy wasn't all that impressive when you consider that it all came down to how quickly it responded with the buzzer, it didn't win because it "knew" more but because it could buzz in faster than the humans and that, IMO, gave it an unfair advantage.

Watson still had to return the correct answer. If you've ever seen a Bing commercial, you know that the first result is not always the most relevant, which was a key to the experiment.

 

[Inspector Hammer!]

It's a personal preference to do so because they gave the thing a name, most wouldn't call their toaster or microwave he or she.

I'd still be very curious to know the outcome if Watson's programmers designed it to mimic human synaptic speed to hand response, if we had a fighting chance to buzz in more would it still have won?

 

[Greg_S_H]

No. If it hadn't had a speed advantage, it would have been anybody's game and the money would have been more evenly distributed. The questions weren't that tough--especially for Rutter and Jennings. They just couldn't get in, despite what that article says. Yes, Watson had to wait for the light and wasn't pre-clicking like human competitors do, but it received the entire question instantly, where the humans had to speed read it. And then there's the slight delay between thought and action, which is a factor even for the best, fastest players.

 

[RobertR]

Originally Posted by Inspector Hammer!
in one article I read they talk about how one day we may develop AI so advanced that we'll have androids that think, act and look just like we do and that they could perform jobs more efficiently than we can...and that's a good thing because?

That would mean that machines will replace human workers, putting them out of a job, correct? Explain to me how that could possibly be viewed as good or an advancement of any kind.

Google "Luddite". Some people have protested the use of machines as a "replacement" for humans since the beginning of the Industrial Revolution, right on up to using robots in auto factories and replacing typesetters with computers. According to the logic of such people, the increasing use of machines should be making us steadily poorer. Instead, no matter how much some try to romanticize our preindustrial existence (Avatar being a recent example), the undeniable fact is that our use of machines has made us better fed, housed, clothed, medically better off, more comfortable, etc. etc. In short, machines have made us immensely wealthier. We could kill unemployment overnight by decreeing a reversion to manual labor. Of course it would make us incalculably poorer....Eking out an existence sweating in a loincloth all day chasing game with a sharp stick (oops, that's a machine) isn't my idea of fun, nor would it make me or anyone else more noble or "spiritual".

If machines "replace humans", it's because it makes sense to do so economically. Suppose you had two auto factories: One makes cars using manufacturing methods from the 1920s--plenty of workers, plenty of jobs. The other is so completely automated that it uses very few workers. It also makes cars at one fifth the cost of the other factory. In other words, it performs the function of making cars much, much better. Which car would you say people are going to buy? Which car would car buyers say makes them much better off, because they have to pay so much less for it? "But what about the workers?", you say. I say, what is the purpose of your car factory? Is it to make cars (in other words, satisfy human wants and needs) or is it to "provide jobs"? We need FAR fewer farmers to grow our food than we used to (because of the use of machines). It seems to me that the only way you could say that's a bad thing is by saying "I don't CARE about the end purpose of growing food, farming's purpose is to employ people." Never mind that all those former farmers are now free to do other things that satisfy human wants,such as making cell phones, medical equipment, and (gasp!) designing and making all those "evil" machines that "replaced" them....

 

[Ruz-El]

Originally Posted by Greg_S_H
Yes, Watson had to wait for the light and wasn't pre-clicking like human competitors do, but it received the entire question instantly, where the humans had to speed read it.

Speed read or not, that read light doesn't come on until Alex finishes reading the questions, so they get the info as close as you can get it. Jennings and the other guy didn't beat Watson on questions were Watson struggled, I think it was as fair as you could get it. They should put a speech recognizer in Watson to eliminate the feeding of text and have another tourney!

And since Watson had a mechanical button presser thing, I think that's enough to label him a ROBOT! Bring on the robot apocalypse!

 

[Jason Charlton]

Rest easy, my friends.

We'll all be protected.

 

[Jeff Gatie]

Gee, a thread in which John states his distrust of science and scientists. In other breaking news; water is wet, the sky is blue . . .

 

[Inspector Hammer!]

Is that the reputation I've earned?

 

[Sam Posten]

ROFLWAFFLES! http://www.theatlantic.com/technology/archive/2013/01/ibms-watson-memorized-the-entire-urban-dictionary-then-his-overlords-had-to-delete-it/267047/