TLDR: The answer is no.
Before you read, I want to state emphatically that I am not an AI researcher. I'm not a professor of any sorts...
A few moments ago, I saw a notification on Notion - new AI capabilities in the famous note-taking application:
According to this headline, you can now suggest to Notion what you'd like to see, and it will be generated for you. Computers have now inched closer to the "let there be this or that" point. Basically, you can now tell a computer what you want to see, and it will be done for you.
This isn't exactly new, of course. Many of you have already experimented with ChatGPT at this point. It can solve homework, act as a terminal, tell jokes, write cover letters, and even defend wrong answers with solid arguments.
So you already know that we are at such a point in time as has never been in the history of mankind.
However, with computers gaining such strong abilities even with our currently limited classical computing power, one can only ponder on that headline, and begin to think about it seriously. Is AI infinite?
Well, for starters, let's look at our own biological makeup as humans. We aren't infinite. Not even close.
What does it mean to be finite? Oxford dictionary tells us it is to be "limited in size or extent". This is not only talking about physical limitations in our growth or size but also about our mental capabilities. We forget. We make silly mistakes. We fall sick. We die.
But are we really finite? This would need to be considered across two main areas: physiologically and psychologically.
Physiologically, we are plagued with a billion diseases. We cannot lift a car with a single finger. We are severely limited. We cannot breathe underwater. We cannot fly. We cannot go to space without special suits. We cannot survive extreme temperatures. The list is endless, really.
Psychologically, we cannot handle heavy mental processing. We cannot think about two things at once. The multitaskers among you are definitely rolling their eyes right now, but it's true. At each given split-second, you are focused on only one thing at a time. You may be able to switch between topics so fast that you experience the illusion of multitasking, but in reality, I do not believe it is possible. As you age, it gets even worse. You forget several things. You take more and more time to figure out solutions to problems. You begin to wonder where the brain that could once understand complex algebra has hidden itself. Need I say more? The truth is right before us.
Now that we have established the fact that we aren't infinite, let us ask: can finite beings create infinite beings?
The short answer is no. A big and resounding NO.
Can any machine be infinite?
The reason we have AI in the first place is that we realized there were many things we couldn't trust our minds to do. I must emphasize that we are talking about our minds here. We needed something that could think faster, learn faster, and know so much more than we could ever keep in our individual brains. This is the same reason we began to build mechanical machines. It's the same pattern. Are machines infinite? Not a chance.
Machines are more efficient. That's true. But they are not infinite. They themselves require maintenance, frequent oiling here and there, occasional replacement of parts and the like. Therefore, man is still greater than machine.
But what about AI? It doesn't necessarily need maintenance. It could be trained to train itself. It can even be trained to train and create other AI. That's something mechanical machines cannot do. AI is an attempt to create a machine that can reason like a human being. But is it infinite?
Let's look at a single case. We have seen reports of ChatGPT giving incorrect answers. It's so bad in fact, that StackOverflow quickly jumped up to ban answers given by it. Let me quote a section of the official notice given by StackOverflow:
Overall, because the average rate of getting correct answers from ChatGPT is too low, the posting of answers created by ChatGPT is substantially harmful to the site and to users who are asking or looking for correct answers.
The primary problem is that while the answers which ChatGPT produces have a high rate of being incorrect, they typically look like they might be good and the answers are very easy to produce.
And this is not only a ChatGPT problem. If any of you tried out Github's Copilot, as I did, then you know that out of the 10 auto-generated code blocks, most were incorrect. This was due to various reasons, but I personally observed that sometimes, the answers were given based on an older version of the language or framework, with slightly different syntax. It's basically GIGO. Garbage in, garbage out. If we fed it with pure garbage code, it would generate pure garbage code, because it had no way of autonomously checking the code for correctness.
This is the fundamental issue! AI cannot be said to be infinite if it gives 99% correct answers and still has 1% false answers. In order to be infinite, it must prove itself to be the ultimate source of truth. You should be able to ask it a deliberately extreme question like "what did the president tell his wife last afternoon", and it would search the database and tell you exactly what it was. It should be able to predict, with super accuracy, the outcome of a soccer match.
For it to be infinite, it MUST KNOW EVERYTHING.
If it does not know EVERYTHING, then it is obviously limited in knowledge. This is common logic. This is common sense.
For deep thinkers only:
To be infinite is to be "limitless or endless in space, extent, or size". That includes mental capability. It must not be in doubt about any subject matter. It must not be ignorant about any detail. If you begin to think of the universe as a set of mathematical and physical laws, then it stands to reason that if we destroy this universe and recreate it with the exact same laws, we would arrive at the same conclusion every single time, because 1 + 1 can never be anything other than 2. Therefore, your AI, if it is truly infinite, should be able to understand the laws of the universe with such precision that it knows everything. I may have lost 80% of you here, but all I'm saying is that your AI needs to have a mental capacity that is out of this world, and that is able to encompass all the knowledge ever known to this universe.
Remember that we are dealing with the mental faculty here. We cannot judge it on the basis of strength, health, wealth, and so on. We can only judge based on mental indicators: speed of thought, accuracy of information, reliability, conscience, ability to predict future occurrences, and so on. AI was created to assist the human mind. However, to call it infinite in your marketing is too strong of a claim. The technology still has a long way to go.
Will it ever be infinite?
As far as I am concerned, AI will never be infinite. Why is this so? Very simply put, AI relies on existing human knowledge in order to be useful. Even if you train it with 100% correct data, you will still require it to learn from its interaction with users. And from there, you and I can clearly see how flaws will be introduced. If the beings providing the input are not 100% correct all the time, there is no way on this planet that the machine learning from these beings will be 100% correct.
Now that you've read this, you can rest easy, knowing that while it may take some of your jobs, just as mechanical and electronic machines took away many manual jobs, it will never end up becoming your ruler.
Thank you for reading, and I'd love to hear your comments.