Friday, April 5, 2019

Good, Bad, Or Indifferent: MDU Was Lucky To Shed Denver When It Did -- April 5, 2019

Link here to MDU and Denver, November 5, 2015.

The end of the shale "boom" in Colorado, April 5, 2019.

It will be interesting to see how fast the SEC requires operators in Colorado to start writing off their investments.

My dad always said it was better to be lucky than to be smart.

****************************
The Book Page

My son-in-law is fascinated by artificial intelligence. Me? Not so much.

Today's buzzword is transhumanism.

There is a lot of talk about how engineers have now figured out how to "wire" computers the "same way" that human brains are wired.

I'll believe it when I see it.

I can't even get Alexa to understand me some days, and my relationship with Alexa is incredibly superficial, pretty much limited to finding music to play and occasionally, for the fun of it, generating grocery lists (to which I never refer anyway).

Alexa's "pre-programmed" phrases are finite, and the order in which she can put those phrases are even more finite. In other words, the number of sentences Alexa can come up with "on her own," is incredibly finite.

The human brain can come up with an infinite number of sentences (source: The Language Instinct: How the Mind Creates Language, Steven Pinker, c. 1994, pages 85 - 86). Example:
  • first sentence: I am eating a hot dog.
  • second sentence: Bruce said I am eating a hot dog.
  • third sentence: why does it matter that Bruce said I am eating a hot dog?
  • fourth sentence: who asked why does it matter that Bruce said I am eating a hot dog 
  • fifth sentence: Jim asked who asked why does it matter that Bruce said I am eating a hot dog
In other words, give me any sentence, a human can make it longer. One can program a computer do do that, of course, but that's not what Alexa does, nor does any other commonly used AI language. Human language is vast (infinitely vast); Alexa's language incredibly finite.

Something tells me Chomsky will be willing to work with AI but knows that AI will never  be able to match the human brain when it comes to "human language."

This book was first copyright in 1994. It is interesting to read it in light of current articles on AI. Also, having just read about the code talkers in WWII it is even more interesting.

It is also incredibly fun to read just at the time Sophia is learning to read, and having watched her closely while learning to speak English. I still love to hear her use new words, and her ability to improve her grammar literally overnight. It is amazing.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.