Why AI Will Never Become Humanlike
Our minds which AI was modeled after, are really good guessing machines. They figure out how to take random neurons firing when we first open our eyes as babies and give us a more and more controlled hallucination of what reality is.
That’s why babies and younger children often struggle to identify people. They might confuse someone for their mother simply because that person looks similar or treats them in a certain way. It’s not a coping mechanism, it’s because of how our brains work.
If you could take a supercomputer with an infinite amount of neurons, put it into a skull and only let it understand its world through eyeballs, ears, and sensory touch abilities… it could very well resemble a human.
here’s why that won’t ever happen, at least not in our lifetime.
There are two reasons for this.
AI in its current form becomes extremely good at 1 task… and I mean really good. The longer the AI exists the better it gets.
For example, Tesla’s self-driving AI might get in a crash or two, but, after the crash, the AI will analyze what it did wrong and ensure that never happens again. An immortal learning machine doing this for decades? At some point, human drivers will become illegal, and self-driving cars will be the expected norm.
At some point, human drivers will become illegal, and self-driving cars will be the expected norm.
The second reason is the computational power required is not feasible with current technology. The human mind has more neuron connections than the universe has planets. To create an artificial mind with that level of power is impossible.
There is a way it could be possible, but exploration into that realm of science has too many ethical problems to be well studied.
