Twenty Questions, at least

You wouldn't think of the game of a site devoted to the game of Twenty Questions as the home of "the neural-net on the Internet," and yet... As the site's author explains: "20Q.net is an experiment in artificial intelligence. The program is very simple but its behavior is complex. Everything that it knows and all questions that it asks were entered by people playing this game. 20Q.net is a learning system; the more it is played, the smarter it gets." So, to play it is to make it "smarter." I wouldn't be at all surprised that it will also make you smarter - at least about the nature of artificial intelligence.

For another perspective on the artificial intelligence - twenty questions connection (henceforward to be referred to as AI-20Q) , I have to share a page from the DeepFUN website, coincidentally called "Twenty Questions."

The article consists of two parts. The first is about an artificial intelligence lesson. A friend introduced me to a computer version of Twenty Questions (this was maybe twenty years ago) that seemed uncannily sophisticated. And yet was written in only five lines of code! The sophistication was an illusion, supplied, not by the technology, but by human naivity!

The second, about a variation of Twenty Questions, called "Plenty Questions," where the answer guy can only respond in terms of "hot" or "cold" and variations thereof. A good game. A different game, actually. Requiring the development of quite different strategies. Demonstrating that the vast intelligence gathered by AI technology can be rendered virtually useless with a simple rule change. Try your AI on this - a 20 Questions game where all the questions can only be answered in numbers!

Demonstrating the following truth: no matter how artificial it gets, the intelligence of Artificial Intelligence lies not in its complexity, but in how well it engages the minds of the users.