AI & Self-Awareness

QUESTION: You do not see AI as actually becoming conscious? There are so many claiming that is the future. Are you hiring programmers in machine language?

LK

ANSWER: Let me explain something. Most generative AI models today are being trained and run in the cloud. These models are language-oriented, generating text. They are often at least 10 times to 100 times bigger than older AI models. ChatGPT is learning from the questions people are asking. While this is impressive to the average person, there is no real economic value other than adding to the search function. This has resulted in a boom along with an insatiable appetite for running large language models at this point in time.

Yet there is something mysterious that we do not yet comprehend – individualism that economists try to eradicate with equality. Even dogs have individual personalities. My little one will take a pill covered in peanut butter. The older one takes the peanut butter and spits out the pill. Just like having two children, they are not the same. What causes one to have a personality that is different from the other?

I’m afraid I have to disagree with this theory that if you throw in enough data to a neural net, suddenly, the computer will become self-aware. My little dog was just 11 weeks old. She is still exploring her environment, displaying curiosity, so she has a distinct personality BEFORE acquiring knowledge of her environment. Anyone who has children will see the same. This PROVES beyond a shadow of a doubt that this theory of a computer becoming self-aware if we throw in enough data is just nonsense. We do not teach our children how to be self-aware. They are born that way. Apparently, so are dogs. Both of mine recognize themselves in a mirror.

There is something in there that creates the personality, and it appears from birth in humans as well as dogs. My dogs clearly think dynamically, as humans are capable of, but not all do. If I get up with a coffee cup, they know I am going to the kitchen and head there. They make connections through observation. Not all animals have that ability. So why are dogs capable of looking for patterns and anticipating my next move, and a hippo, snake, or alligator is not?

I do not believe we are anywhere close to comprehending those differences, and as such, we cannot create a true cognitive machine that is self-aware when we do not understand what makes us self-aware. We are born that way – computers are not.

To build Socrates, I had to study intensely how we actually think. Let’s say you met the person who is your soul mate. You went to dinner. Your mind is actually recording every aspect of that evening, and you are totally unaware of what it is recording. Years later, you return to that same restaurant, and that memory involuntarily comes rushing to the forefront of your mind. Or there was music playing that night, and suddenly you hear that same song, and that memory again involuntarily comes to mind. Perhaps it is the food. The point is that memory is stored, but any of our senses can access it. The complexity is enormous.

My oldest dog does not like going to the Vet. She even knows her surroundings. Miles from the place, she knows where I am going and begins to tremble. She is aware of her surroundings miles from home, which astonished me. It was not a capability I assumed a dog would have. A dog stayed with a two-year-old girl who went missing and always protected her.

There is a lot more hidden within us and many animals that cannot be explained or recreated by a computer.

Just be careful with claims suddenly of people coming up with AI trading programs. It took me 17 years to build Socrates. You cannot create an AI trading program in a few weeks. No AI system will ever forecast the future, looking at a single market. The wildcard ALWAYS comes externally.

As for hiring AI programmers interested in real work, yes. We always have projects under development.