In the past, I have written about a guy named Greg, and I thought I should revisit the character and flesh out his backstory to give him more depth. Coincidentally, I’ve been thinking about trying out ChatGPT from OpenAI.
To be clear, I am not a fan of either Greg or artificial intelligence, whether combined or singularly. Artificial intelligence belongs in the same category as artificial meat: it’s awful in theory, but then you try it, and discover it’s worse than you expected. Nope, I was born a regressionist, and I will die a regressionist. Also, a snake handler.
Greg is the name I have given to a pervasive recurring fictional character that I sometimes include in my mental excursions. Greg is always the answer to the question, What was that guy’s name, the guy who did the thing with the stuff?
ChatGPT is browser-based demonstration application for natural language AI. You can sign in and chat with it, ask it questions, get advice on how to tell your boss he’s a dickhead, or even get it to write a blog post for you.
While ChatGPT does a neat job of understanding your questions and keeping track of the context thread of the conversation, it doesn’t take long to notice that many of its responses are based on fill in the blank templates, with a little bit of random variation thrown in to the mix so as to appear intelligent, or at the very least, nondeterministic.
AI is more artificial than intelligent, but the hypesters of the industry are regularly featured in the media warning the masses of the coming singularity that will put an end to all humanity. The trick, for the unevolved lifeforms out there in the galaxy, is to pull the power cable, and then bang some rocks together.