Podcast Episode

Gigaom AI Minute – June 26

:: ::
In this episode, Byron talks about computers and nuance.

Gigaom brings you our unique analysis and commentary on the present and future of AI.

Transcript

I'm spending some episodes of AI Minute talking about common sense. How it is that humans know not to walk off a cliff, but it's hard to teach a computer why that would be a bad idea. I want to read you four funny headlines that I found on the web, and I want you to think about why each one of them is funny.

"City unsure why sewer smells"
"Five stabbed at welcome home from jail party"
"Man found dead in graveyard"
"Army truck vanished after camo paint job"

Now, in each of those cases you were probably able to understand why it is that those were funny. Why would a sewer smell? Well, why wouldn't a sewer smell? Graveyards are full of dead people, so why was it a news story that a dead person was found in a graveyard? Well, it was probably a dead body, and so forth. Now, how are you able to parse those out? How would you explain that to someone else? How would you teach a computer that "army truck vanishes after camo paint job" is a humorous headline?

It's that nuance and that interplay between different concepts, that you probably didn't have to stop and analyze those to see what they were worth a chuckle or not. You, almost intuitively, are able to realize why those are funny. And whatever that intuitive element is, is something we haven't figured out how to teach computers and it's an impediment on our creation of a general intelligence.

Share your thoughts on this topic.

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.