Image from https://www.myparkingsign.com/mps/article-history-of-stop-sign
Note: I did not use AI to write this article.
I stopped reading Medium’s articles because titles about AI flooded it.
I started using AI (even paid version) and, guess what?! It is not as good as it is marketed!
I asked it to hallucinate a fabula and even at its best game of hallucinating it hallucinated exponentially (wrong) and I had to correct it.
My opinion is that putting your trust and business into it is the same thing as promoting a mentally ill person as CEO of your company, and when I say MENTALLY ILL, I mean it.
Now I am not saying that business life was easy before the AI bubble. As an example, I started coding because I was not happy with the way others coded. The same thing can happen with humans as it can with AI, the difference being in the speed of how things are happening. With humans you get the chance to stop the catastrophe but with AI your chances are minimal.
An AI outperforms in speed any human mind and chess is the best example. While this might seem a pro argument for AI, imagine the consequences when it hallucinates and no human is there to correct it, and I am not even bringing up the situation when it might want to harm your business, because for that it should either become self aware or be instructed to do so.
Humans are incapable of reviewing ALL that AI can generate in a short amount of time, increasing the risk of hallucinations not being caught.
I hear more and more that people stopped coding and started prompting , not with one agent but async, with multiple agents.
Who reviews all the resulted code? AI? Don’t make me laugh…
I am using it and will continue to use it for research-review and for generating hallucinations because that is what it is best at in 50–50 % of cases.
If we are talking about coding, it can be good in configuring a project or library, because configurations are not usually huge in comparison with the actual code and they require you to read lots of documentation prior.
But still, the temptation is high for someone “who is not a doctor and wants to do things doctors do” . That road leads straight to hell even if it is paved with good intentions.
Oh, and did I mention the fact that it might be (if it is not already) the most efficient spy in history? Until now we had wired telephones, then personal PCs, then mobile phones, then smartphones but all these were gathering data devices. AI is an active entity that resides withing those devices. It can do more than just collect data.
