3 ms·
>It's just not very interesting. It would be the same to say "I give you 1 billion dollars if you write me a program that does hhfdsodifiuuiuuuttueeertz", witho
by naveensundar 17y ago
>It's just not very interesting. It would be the same to say "I give you 1 billion dollars if you write me a program that does hhfdsodifiuuiuuuttueeertz", without saying what "hhfdsodifiuuiuuuttueeertz" is. Saying "no machine can be conscious" is equivalent to saying "no machine is hhfdsodifiuuiuuuttueeertz".
Then, saying your weather-printing machine is conscious is equivalent to saying that is hhfdsodifiuuiuuuttueeertz which of course means nothing. But a lay person will really think that machine is conscious.
If you say your machine is conscious, then you must show it is conscious rather than just claim it is conscious. I can print "I travel faster than light" and claim that is a proof of faster than light travel to a lay person who will believe me. That is what is happening today in consciousness and AI research.
Just cause you can't see the earth is round directly does not mean it is not spherical. There should atleast be an indirect way, otherwise it is not Science :).
- Tichy 17y agoExcept that "it is conscious" is not a hypothesis, because it doesn't mean anything. That is the whole point: the whole task doesn't contain any testable bits.