Skip to content

AI – The Reality of Complexity

Spread the love

AI Computing

COMMENT: You are always ahead of the curve. Today AI is the buzzword – with most of the AI being machine learning, where you were developing Socrates with real AI in the 70s and 80s.


Bulldogs playing Poker

REPLY: I fooled around with neural nets when they first began. The problem with this approach is that you expect a machine to develop a human instinct as if you are playing a poker game. It is a gut feeling you might have about a person to alert you if he is bluffing or real. That cannot be coded, nor will a computer with machine learning be able to acquire such a “gut feeling,” which is an entirely different game than chess.

Raven chimpanzee_on_Wall_Street_buys stocks

Socrates is NOT a neutral net. I had to teach it how to trade. I put my instincts into the system. Creating a neural net, throwing in all the data, and praying it will learn how to trade is more or less like a monkey throwing darts at the Wall Street Journal regarding what stock to buy in a bull market. Raven, a six-year-old chimpanzee, became the 22nd most successful money manager in the USA after choosing her stocks by throwing darts at a list of 133 internet companies. The chimp created her own index, dubbed MonkeyDex, and in 1999 delivered a 213 percent gain, outperforming more than 6,000 professional brokers on Wall Street.

Complexity Sychronization

Complexity 1

Attempts to use a neutral net with machine learning have not beaten Raven in funds management. The level of complexity is monumental. On top of all of that, it will never discover the nonlinear structure of the world by dumping in a chunk of data and praying for the best.