Posts

Comments

Comment by mclaren on Artificial Addition · 2007-11-22T01:28:11.000Z · LW · GW

The issue with AI has nothing to do with ignorance or arrogance. The basic problem is that intelligence can't be meaningfully defined or meaningfully quantified. Documented fact: Richard Feynman had a measured I.Q. of 120. Documented fact: Marilyn Vos Savant had a measured I.Q. of 180 or 200, depending on which test you place more faith in. Documented fact: Feynman made a huge breakthrough in physics, Vos Savant has accomplished nothing worth mentioning in her life. I.Q. measurements fail to measure intelligence in any meaningful way.

Here's another fact for you. Louis Terman collected a group of so-called "geniuses" sieved by their high I.Q. scores. Two future nobel prize winners, Shockley and Alvarez, got tested but discarded by Terman's I.Q. tests and weren't part of the group.

Question: What does this tell you about current methods for measuring intelligence?

There is no evidence that people can meaningfully define or objectively measure intelligence. Rule of thumb: if you can't define it and you can't measure it objectively, you can't do science about it.

[Remainder of gigantic comment truncated by editor.]