Recently, I was talking to someone about it, and I realized that while Artificial Intelligence “seems” quite intelligent, it is essentially just a very complex statistical formula. Neural networks and other AI techniques are really just advanced ways of doing statistics.

If it can, it will likely become much more intelligent than we are, and that is quite scary. On the other hand, if AI cannot learn from itself, it is only as good as the data set it is given.

For example, chatbots like GPT can provide inaccurate information if the data set they are using is outdated. The problem with AI is that it is limited by the quality of the data set it uses.

I believe we will heavily rely on AI in the next few years, but at some point, we will hit a plateau where AI becomes useless because the data set is too old.

People may stop putting information into the internet because they rely too heavily on AI. While AI is good at generating content like computer code, essays, and emails, it struggles with skills that are not highly represented on the internet, such as plumbing or electrical work.

The reason for this is that people want cutting-edge technology and entertainment, not necessarily practical skills. AI works well with this type of content, but it will struggle with other skills unless it can learn from itself.

In summary, AI is a powerful tool, and we should embrace it, but its limitations will eventually become apparent. Its success depends on the quality of the data set it uses, and it may struggle with skills that are not highly represented on the internet (that is, if it doesn’t learn how to learn from itself).