homunculus has appeared //
The person who controls the system, who is the little person, who watches what the retina shows, who makes the decisions, who speaks out…
This is the dualist answer to the problem of free will-yes, my body has a human in it, but that little person can make independent decisions. …)
Anthropomorphism can be a powerful tool. We imagine that a person, who is like us, controls the complex thing. This person would be the one to do the same things as we would if we were in the control panel.
We try to predict the behavior of a tiger or an lion by imagining a small person at their controls (perhaps a feline person who is more wild, less intelligent and more feline than we are). The stories we tell about life on Earth are a collection of tales that describe the little people within everyone.
Artificial intelligence is problematic because we can read the code, and therefore prove that there is no human inside.
When computers beat us in chess we said “that’s just dumb code, it can solve problems.”
We did the same when computers began to “compose music” or “draw images”. Quotes are crucial, as the computer could not possibly be a person.
LLM, and other things such as ChatGPT, have turned this on its head. It’s impossible for AI researchers to use these tools without picturing the person inside.
This insight might be useful: There is no little person in us.
No one of us does.
All we do is code, just as ChatGPT.
We’re not discovering a brand new form of magic. The old type of magic was an illusion.
