What is something a human could do...
Posted: Wed Jul 30, 2008 2:00 am
On a computer, that could not be programmed?
I'm just curious.
I'm just curious.
Page 1 of 1
Posted: Wed Jul 30, 2008 3:49 am
Pee on it.
Posted: Wed Jul 30, 2008 4:12 am
poke the keys to see what happens.
Posted: Wed Jul 30, 2008 6:32 am
Draw on the case
disassembly
disassembly
Posted: Wed Jul 30, 2008 10:21 am
hug your wife and kids.
er and throw the cat out the back door.
er and throw the cat out the back door.
Posted: Fri Aug 01, 2008 3:10 am
Personally I'd take out in the desert and shoot it.
Posted: Fri Aug 01, 2008 8:04 pm
Recognize a joke
Re:
Posted: Fri Aug 01, 2008 8:20 pm
Lol, this from the guy who couldn’t tell I was joking when I said “I have only one word to say” then said 3ccb056 wrote:Recognize a joke
Posted: Sat Aug 02, 2008 4:48 am
That's because he is a computer. Grats.
On a more serious note, there are tasks that computers can't do well, and also tasks that computers probably could never do. An example of the former is recognising places, people and objects from photographs. An example of the latter is discovering new laws of physics.
On a more serious note, there are tasks that computers can't do well, and also tasks that computers probably could never do. An example of the former is recognising places, people and objects from photographs. An example of the latter is discovering new laws of physics.
Posted: Sun Aug 03, 2008 9:35 pm
There is a long list of items that we humans do that would take a very long time to teach a computer to do... as well as a long list of things a computer could never do.
Here's just a few things we humans can do very well.
- Understand or Misunderstand Emotions.
- Pray.
- Say \"I love you\" and really mean it.
- Hate.
- Love.
- Ignore.
- Jump to illogical conclusions based on partial information.
- Prejudice.
- Peer Pressure.
- Understand non-verbal clues such as body language.
- Discern truth from lies by observation.
- Procreation.
- Bear Children.
- Carry on multiple conversations in a chat room at the same time.
... etc ...
Funny side note:
Many years ago they had a contest using some very simple computer \"artificial intelligence\" using a maze and a mechanical \"mouse\" to go through a random maze to see which teams' mouse could do it the quickest time.
One team had a mouse that had a logic error in it... it ran the maze extremely quickly... the other teams, not to be outdone, reacted by quickly dumbing down the AI in their mice... so the dumbest \"mouse\" actually ran the maze fastest.
Hehehehehehehe!
Here's just a few things we humans can do very well.
- Understand or Misunderstand Emotions.
- Pray.
- Say \"I love you\" and really mean it.
- Hate.
- Love.
- Ignore.
- Jump to illogical conclusions based on partial information.
- Prejudice.
- Peer Pressure.
- Understand non-verbal clues such as body language.
- Discern truth from lies by observation.
- Procreation.
- Bear Children.
- Carry on multiple conversations in a chat room at the same time.
... etc ...
Funny side note:
Many years ago they had a contest using some very simple computer \"artificial intelligence\" using a maze and a mechanical \"mouse\" to go through a random maze to see which teams' mouse could do it the quickest time.
One team had a mouse that had a logic error in it... it ran the maze extremely quickly... the other teams, not to be outdone, reacted by quickly dumbing down the AI in their mice... so the dumbest \"mouse\" actually ran the maze fastest.
Hehehehehehehe!
Posted: Sun Aug 03, 2008 11:06 pm
Getting a computer to love, understand jokes, etc. are problems of complexity AI-complete, i.e. the Strong AI problem can be reduced to any of these problems, so if you can prove that a computer can do one of them, then you've proved that a sufficiently advanced computer can do any of them.
Re:
Posted: Mon Aug 04, 2008 8:41 pm
Something interesting to think about, a really great way to demonstrate this same concept is by looking at computer games.SirWinner wrote:Funny side note:
Many years ago they had a contest using some very simple computer "artificial intelligence" using a maze and a mechanical "mouse" to go through a random maze to see which teams' mouse could do it the quickest time.
One team had a mouse that had a logic error in it... it ran the maze extremely quickly... the other teams, not to be outdone, reacted by quickly dumbing down the AI in their mice... so the dumbest "mouse" actually ran the maze fastest.
Hehehehehehehe!
Many games with dumber AI appear to actually be smarter to the player because they can fulfill an objective faster and the faster they can do this, the tougher the enemy.
Decision making within the computer realm revolves around ratios, probability, and random selection.
Humans tend to have a more complex thought process including morality, observance of patterns, greed, and strategy.
Enemies in Crysis and Splinter Cell may appear to be dumber because they have a limited detection distance and they lose sight of you very easily but they are programed to have more thought than say a sniper in Halo where they can spot you the second you walk into an area. Most games, upon entering an area, the AI will immediately attack you and will appear to be somewhat psychic. In a game like Oblivion, guards AI functions on the basis that if a player commits a crime the guard will simply go to wherever your location within a city is and attack you. In-fact, if you steal say, a potato, a guard will suddenly appear or spawn right behind you the second that you take it, even though, you managed to sneak into the house. Not only this, but the guard would then proceed to chase you clear across the continent (or whatever you call it) of Cyrodil. No matter, if you try to hide or fool the AI in one way or another, they will always find you. Why? Because the AI is retarded.
