badukJr wrote:lemmata wrote:An alternate
view.
Its really hard to take any commentary seriously
where they refute a mathematical argument with a couple of handwavey sentences. "Oh, everyone knows a pole doesn't balance on a cart, so this is wrong"
It seems like these author's articles are always like that, they attack mathematically heavy work with words only - as if it will somehow prove the math wrong. Its very strange.
Perhaps I shall offer a criticism myself, if only to prevent the thread from turning into an argument about vaccines.
There is no mathematical argument to refute. The so-called "equation for intelligence" is actually just an old thermodynamics equation for entropic force. It accurately predicts certain important physical processes. TED Video Guy (TVG) did not come up with this equation. In fact, we might even say that all the interesting claims made by TVG are really hand-waving arguments.
What TVG does is essentially this: He takes a system (like the three discs of different sizes, the smallest of which is in a tube). He defines some state/control variables that determine the transition to future states. Changing the control variables moves the large disc. He then lets these state/control variables be relabeled as variables in the thermodynamics equation and uses it to generate the value of the control variable (move the largest disk) in the next state.
When he does this, the large disc is moved to hit the medium disc until it dislodges the smallest disc from the tube. He says that the discs represent a monkey, a tool, and a piece of food in a tight space.
But there is no math that says this is what it is supposed to represent. Indeed, he doesn't even mathematically prove that this equation will necessarily generate such behavior. The computer simulations all use very specific parameter values. It is unclear how robust they are to changes in those values. Essentially, in academic terms, all he did, even in the published paper, was tell a few fascinating
anecdotes.
The only scientifically valid conclusion we can draw from TVG's original paper is that this equation can be used to generate behavior for three different situations that is usually achieved with three separate algorithms.
TVG only jumps into the realm of pseudoscience hokum typical of many TED talks when he starts claiming that he has found any scientifically valid evidence that this equation has anything to do with intelligence. That's wonderful for a cocktail party, which perhaps describes the prevailing mood at these TED talks, but not appropriate for a serious conversation about science.
