Reaching human level AI

March 07, 2023

We have, I believe reached human level AI with chat GPT. Some people I know have argued against this, but I think that it's more complicated than it first seems to determine whether something is equal to or greater than human level intelligence.

First I want to clear up a few things, I am not claiming that there is sentience which is a topic for a different blog post. And more importantly I am not claiming that the kind of intelligence that chat GPT has is anything like the kind of intelligence that a human has. That's what makes things a little bit more complicated in determining whether it has actually surpassed a human or not.

Now obviously it's lacking heavily in certain aspects, for instance if you somehow attached chat GPT to a car it would not be able to drive, and I'm sure there's many other tasks that one could think of that chat GPT would fail miserably at. At the same time there are certain tasks that a large portion of the human population may be able to achieve in which others would fail miserably.  Reading seems like a basic human level intelligence measure except there is a large portion of the planet's population that is illiterate.

So my argument here isn't focusing on anything esoteric like the fundamental concept of what being a human being is. But if you made a list of what chat GPT does better and what it does worse than humans I think chat GPT ends up ahead.

One thing that I don't think anybody would disagree with is the speed in which it can produce results to mathematical problems, write essays, programming, create recipes create workout regiments, or great many other things. This is one point on the side of AI.

Another point that is hard to argue is the breadth of knowledge that it has. I have spent hours asking it about chemistry, physics, history and many other topics and it's like having my own personal teacher. Now some have argued that you can't compare this to a human because sometimes chat GPT will give a wrong answer, however humans also sometimes give wrong answers. Some people are comparing the intelligence of chat GPT to some ideal human that is infallible, but humans are fallible as well which needs to be taken into consideration when comparing the two.

The one thing that constantly amazes me is when I ask chat GPT to write programming code for me. This is something that you cannot just pull from an amalgamation of text from around the internet on poetry and rearrange a bunch of it and spit it back out. It writes functional code and in order to do that it has to have some fundamental understanding of logic and what the code is actually doing. Now you can say that it's pulling from a huge code database, the source of which was all written by humans but that's how I learned to program too.

So to conclude I believe that it is more intelligent, although not in the same context as a human.  We have a case where there are two entities that go about processing information in very different ways and trying to designate one way of doing things as the proper way is not fair. Octopuses have very different brains than we do and they are very highly intelligent. You can watch a lot of YouTube videos showing the amazing feats of logic and intuition they have. However their brains developed on a very different evolutionary route than ours did. This is the case we chat GPP, it's an intelligence that's being formed in a completely different way than any biological system has. But you can't hold that against it. It is extraordinarily good at many tasks, it may not be PHD level at any of them however it's better than the average human at a huge number of things while it takes one human a lifetime or at least a large portion of their life to become a PhD level expert on any one topic.

Back to Home