geekhack
geekhack Community => Off Topic => Topic started by: cgoldberg on Mon, 21 February 2011, 10:14:28
-
very interesting stuff.
"Building Watson: An Overview of the DeepQA Project"
https://www.stanford.edu/class/cs124/AIMagzine-DeepQA.pdf (PDF)
-
I want one for my company's call center =)
-
Watson is not actually "intelligent".
I'm impressed but only mildly. While computers can be programmed to make decisions on very large amounts of data, the decisions are programmed by people. The computer is only as intelligent therefore as the programmer who designs it.
-
Watson is not actually "intelligent".
I'm impressed but only mildly. While computers can be programmed to make decisions on very large amounts of data, the decisions are programmed by people. The computer is only as intelligent therefore as the programmer who designs it.
Good thing I don't program, then...it'd be iRobot in real life!
-
Watson is not actually "intelligent".
I'm impressed but only mildly. While computers can be programmed to make decisions on very large amounts of data, the decisions are programmed by people. The computer is only as intelligent therefore as the programmer who designs it.
Funny that. The goal posts defining intelligence have moved quite a bit in the last half century or so... It used to be about things that humans find hard - but a lot of that turned out to be stuff that wasn't impossible to program a computer to do. Nowadays AI researchers tend to be more interested in getting computers to do things that humans find easy to do, but would find it hard going to program a computer to do.
It doesn't matter what breakthrough comes next, the goalposts will shift again.
-
I want one for my company's call center =)
Yeah! but get one to hound a rival company's call center with difficult questions.
I'm sure Watson can be put to evil uses.
-
Watson is not actually "intelligent".
This ^ contradicts...
The computer is only as intelligent therefore as the programmer who designs it.
This ^ doncha think?
I am quite impressed with what they were able to get this computer/program to do. Lets say for the heck of it that this computer is "only" as intelligent as it's programer(s) as you put it, then I think it's safe to say it's far more intelligent than the majority of people. LoL XD
=)
-
I meant that its not "intelligent" in the same sense that humans are. It's impossible for computers to be, IMHO.
-
Just remember it was humans who built Watson. :)
-
Watson is not actually "intelligent".
I'm impressed but only mildly. While computers can be programmed to make decisions on very large amounts of data, the decisions are programmed by people. The computer is only as intelligent therefore as the programmer who designs it.
Computers can be programmed to make decisions that the programmer would not know how to make a correct decision about. In other words, computers can be programmed to learn how to make their own decisions (see neural nets.) As an example, a programmer could program a chess computer that could consistently beat the programmer. The programmer only sets guidelines on how to learn to win at chess and not explicitly what move to make in each situation. Whether you consider that intelligence or not is up to you, ultimately.
-
I DO think its interesting that while computers can hear, see and speak, they cannot smell or feel. That makes me feel vastly superior.
Also computers know nothing of expressing emotions. Emoticons do NOT count people!
No matter HOW cute they are.
(http://geekhack.org/picture.php?pictureid=653&albumid=112&dl=1297990076&thumb=1)
-
It doesn't matter what breakthrough comes next, the goalposts will shift again.
For a while, but I don't see that as being dishonest.
Computers still can't do a lot of things that people can do. And, after the fact, things we thought would be hard for us to get computers to do have turned out to be achievable... in trivial ways that didn't give us much of a clue to how to get computers to do lots of other things that require "intelligence".
Some tasks do call for more sophisticated types of programming than the ordinary things computers are usually used for. They make less efficient use of CPU cycles as well.
But the goal posts won't move if something based on computer hardware has wants, desires, and feelings.
Lets say for the heck of it that this computer is "only" as intelligent as it's programer(s) as you put it, then I think it's safe to say it's far more intelligent than the majority of people.
No. It's not intelligent at all.
It is good at answering trivia questions.
However, if someone told it that the budget for paying the electrical power bill for the computers on which it was running was running out, would it try to think of a way to find a paying job? Would it do anything? Would it even care?
No. It has none of that sort of thing. It parses text questions fed to it (it didn't do speech recognition, but was treated as a deaf contestant) in a somewhat open-ended manner, but it doesn't come close to dealing with reality like an intelligent organism.
This is not to say that intelligent computers are not possible. The neurons of which our brains are composed are physical objects which obey physical laws. But we are still hugely distant from being able to design true intelligence in silicon.
-
Funny thing is, I dont think it was all that much better at answering questions (than a jeopardy-calibre human). I think it's main strength was the same thing that has always been the forte of computers, and the reason we use them daily: It can do what we can do FASTER and more consistently. Namely, in this case, "buzz in". I find it hard to believe that either of the human contestants would know less than 80% of those answers, but for the most part, the only time they got to buzz in before Watson was when his certainty of an answer was relatively low.
This is basically the same reason we have used calculators for ages, not because we cant do the math, just that we are relatively SLOW at it.
-
Well this topic may not be about what it says but anyway..
I'm seriously considering getting out of the IT business alltogether. I've been coding for a living for the past 7 years and I'm getting fed up. The main issue is that Microsoft keeps producing new sh*t every 2-3 years and it's virtually impossible to keep up. In my current job we're still doing most of the coding in .NET 2.0 which brings me to the second issue - having too many (idiot) bosses. I've changed three jobs now and it's been the same everywhere. I'm beginning to lose hope that there are normal companies out there.
The alternative is trade. My father has a small shop that deals mostly in numismatics as well as other stuff. If I decide to take over I plan to increase international visibility and sales. Current level is well almost non-existent. I'm also planning to expand the business to include Magic The Gathering for example as well as vintage games and stuff.
Any other coders/devs here thinking about changing jobs?
Cheers,
SainT
-
I think programmers have the least to worry about. Watson is already poised to start working in the medical field. Law seems like the next step, along with simpler tasks such as customer service and data mining.
-
Wow, I just noticed how unfunny and/or vague my sense of humor apparently is.
-
Well this topic may not be about what it says but anyway..
I'm seriously considering getting out of the IT business alltogether. I've been coding for a living for the past 7 years and I'm getting fed up. The main issue is that Microsoft keeps producing new sh*t every 2-3 years and it's virtually impossible to keep up. In my current job we're still doing most of the coding in .NET 2.0 which brings me to the second issue - having too many (idiot) bosses. I've changed three jobs now and it's been the same everywhere. I'm beginning to lose hope that there are normal companies out there.
The alternative is trade. My father has a small shop that deals mostly in numismatics as well as other stuff. If I decide to take over I plan to increase international visibility and sales. Current level is well almost non-existent. I'm also planning to expand the business to include Magic The Gathering for example as well as vintage games and stuff.
Any other coders/devs here thinking about changing jobs?
Cheers,
SainT
Stop coding for/with MS products...
-
My place does .net but tech currency is a big priority. We are already on vs 2010 and Sql 2008.
-
The Double Helix (http://www.amazon.com/Double-Helix-Personal-Discovery-Structure/dp/074321630X) is strongly recommended.
As in Watson and Crick, I take it. As I suspected from the lab coat...
The alternative is trade. My father has a small shop that deals mostly in numismatics as well as other stuff. If I decide to take over I plan to increase international visibility and sales. Current level is well almost non-existent. I'm also planning to expand the business to include Magic The Gathering for example as well as vintage games and stuff.
Remember that new Magic: the Gathering product can't be sold internationally.
A coin and stamp store in a large shopping mall in my area added Magic: the Gathering as well as sports cards in its general "collectibles" theme, and is doing quite well.
-
The alternative is trade. My father has a small shop that deals mostly in numismatics as well as other stuff. If I decide to take over I plan to increase international visibility and sales. Current level is well almost non-existent. I'm also planning to expand the business to include Magic The Gathering for example as well as vintage games and stuff.
Remember that new Magic: the Gathering product can't be sold internationally.
A coin and stamp store in a large shopping mall in my area added Magic: the Gathering as well as sports cards in its general "collectibles" theme, and is doing quite well.
-
As in Watson and Crick, I take it. As I suspected from the lab coat...
Remember that new Magic: the Gathering product can't be sold internationally.
A coin and stamp store in a large shopping mall in my area added Magic: the Gathering as well as sports cards in its general "collectibles" theme, and is doing quite well.
Don't get too hung up on gold right now. A lot of coin dealers, during these rushes, get flooded with gold-centric business and can end up pretty dependent on the market bubble.
In the US, in 1980, gold spiked to like USD900 per ounce (apparently close to USD2000 when adjusted for inflation) but the bubble burst fast.
When the price of gold collapses again-- and odds are, it will, given that a lot of it was a runup due to fears about other markets-- the last guy holding onto the gold loses his shirt.
/afk to hoard foreign banknotes.
-
Good thing I do .net now. M$ FTW!
-
RE: Watson's buzz speed, they are basically saying it does, in fact, have a super-human ability to buzz. 5-10 milliseconds is many times faster than human avg. Also, it is CONSISTENT.
Their argument that Watson does not win ALL the buzz-races is silly, since Watson's decision to buzz was based on it's certainty (much the same as a human would decided whether to buzz or not).
If you dont agree, read the response from Ken himself, who admits that is an ADVANTAGE of Watson, but did not consider it "unfair", as Watson did suffer certain disadvantages as well.
-
What argument? And who is a sore loser?
-
Something only humans care about.
What are "sufficiently vague posts to avoid any direct discussion and possible conflict"?
-
Well this topic may not be about what it says but anyway..
I'm seriously considering getting out of the IT business alltogether. I've been coding for a living for the past 7 years and I'm getting fed up. The main issue is that Microsoft keeps producing new sh*t every 2-3 years and it's virtually impossible to keep up. In my current job we're still doing most of the coding in .NET 2.0 which brings me to the second issue - having too many (idiot) bosses. I've changed three jobs now and it's been the same everywhere. I'm beginning to lose hope that there are normal companies out there.
The alternative is trade. My father has a small shop that deals mostly in numismatics as well as other stuff. If I decide to take over I plan to increase international visibility and sales. Current level is well almost non-existent. I'm also planning to expand the business to include Magic The Gathering for example as well as vintage games and stuff.
Any other coders/devs here thinking about changing jobs?
Cheers,
SainT
I feel your pain. I switched to security for a while, then back to programming again. The painful fact about programming is that their is very little "intellectual capital" that can be carried over from year after year. Knowing a product today isn't going to help you tomorrow. Lawyers / Doctors / Engineers build intellectual capital every year and are rewarded for that intellectual capital late in their career. Programming is such a young industry that it's hard to build capital with the ever changing technology.
-
As discussed in Irobot, imo computers won't be fully recognized as intelligent, even if they are, until they act on their own and disobey humans in their own self interest, just like a 2 year old child does.
Just hope it's not hooked up to weaponry or skynet when it does.
-
I am not afraid of Watson. Any computer designed by man can be destroyed by man.
Here's how:
Me: "Watson, please explain the logic behind Cherry Corp. product codes."
Watson:
(http://www.allamericanpatriots.com/files/images/mushroom-cloud.jpg)
-
Last time I checked, Watson wasn't responsible for programming itself, so how exactly do programmers need a new profession?
-
okay, once again I am apparently out of the loop. What happened? A computer went on Jeopardy? And why does the title of the thread make me think of the Terminator skeleton rising from the smoking ruins?
-
Will Watson be forced to sing as well?
-
Just imagine the havoc if muggle studies covered autohotkey... None of that tedious wand waving...
-
The painful fact about programming is that their is very little "intellectual capital" that can be carried over from year after year.
I disagree. Once you master essential programming concepts you can use and build on them indefinitely, so the intellectual capital is definitely being carried over. Sure, you have to learn new things on a regular basis but mastering different languages and platforms is trivial in comparison to learning how to program well in the first place.
-
I disagree. Once you master essential programming concepts you can use and build on them indefinitely, so the intellectual capital is definitely being carried over. Sure, you have to learn new things on a regular basis but mastering different languages and platforms is trivial in comparison to learning how to program well in the first place.
Keyword was "very little". Relative to a profession like law.
-
I agree with pfink - new languages don't come around that often, and rarely offer a huge paradigm shift. Frameworks and such are more of a pain, but there's such a wealth of material available these days it really doesn't take long to feel familiar with something.
Law keeps changing as well - actually, I'd imagine it's probably quite similar to programming in the amount of time that 'change' takes up.
Maybe it's because I grew through doing electronics, then assembly, then C etc, then various OO languages - from my perspective it really hasn't changed that much! On the other hand, I wouldn't like to go back to programming assembly with hardly any libraries available: progress is good :)
I'm still hoping to find time to get a handle on Lisp... preferably before I retire!
-
I think that frameworks/APIs are getting larger and there is much more reuse than ever before. Especially with frameworks like Rails you can put up a very feature rich site crazy fast. Languages are becoming more feature rich too - as soon as Java automated creation of getters and setters (in Java 6), .Net did the same thing in Framework 4.0. As far as I'm concerned, the less code I have to write, the better. Gives me more time to do what's really valuable in software development - working closely with analysts to design the solution.
-
Are you suggesting I'm learning C99 and MIPS for nothing?
-
Me: "Watson, please explain the logic behind Cherry Corp. product codes."
You've been watching too much Star Trek.
-
To be clear you OFTEN are a sore loser.
Care to cite sources?
-
Not really.
That's what I thought.
-
He's lazy.
-
Thanks for taking the high ground, we could all learn a thing or two from you...