You might wonder why “AI” is wrong so often. The easiest way to explain it is:
ChatGPT doesn’t actually understand what any of the words that it is saying. It’s more like a parrot than an actual human; a parrot might speak complete sentences in English, but has no idea what it’s saying, it just regurgitates back conversations it hears. Think of it like a fancier (and much more expensive) version of Google search: great for finding information already on the Internet, but hit or miss for solving problems that haven’t been posted on the Internet yet (well, it can also solve some math/physics problems, as long as similar problems have already been posted on the Internet somewhere, just with different numbers).
Or maybe think of it as a bigger version of your iPhone’s autocorrect suggestions. Try it: pull out your smartphone and just keep hitting the autocorrect’s next most suggested word 20 times, and it comes out with English sentences, but they are just things you’ve already typed in your phone; it cannot think for you, and it would be ludicrous to think it could.
So, despite having AI GPUs that can do 20 quadrillion calculations per second (just 5 of these GPUs are estimated to have the same number of calculations per second capabilities as a human brain, and Nvidia alone has sold millions of them), ChatGPT and DeepSeek still make lots of mistakes on even relatively “easy” problems (for a human)
So here’s the trillion-dollar project: the first person who figures out how to express in computer code, how to make those supercomputers think, and understand words, in the same way that humans understand words, should rush to the patent office, get an IPO, and make themselves very, very, rich.
Btw, for investing advice, I don’t recommend Nvidia stock. I recommend investing in AI software, not hardware, and before the AI boom, software companies (like Microsoft) have historically had a higher total market cap than hardware companies (like Intel, AMD, Nvidia, and ARM). Chinese tech companies seem like an even better investment than US stocks. Either way, starting to invest earlier is just as important as which specific stocks to pick.
Oh, and also, if you ever hear a viral news story about how an AI agent “started thinking on its own” or “started to disobey humans and wants to kill humanity like in certain movies”, you should confidently go right ahead and report those “news sources” as spreading misinformation and just farming views/scams.
