The artificial intelligence that I am currently using, ChatGPT requires a lot of instruction from me in order to work, and though some of the results have been fascinating to me, it has not been feasible for me to type all of those instructions constantly into my phone. It’s too laborious. But I quickly learned that I could speak my instructions into the phone instead. Unfortunately, doing this anywhere remotely viewable by other humans, and especially at work, makes me appear unhinged. Nevertheless I am driven so strongly by my fascination with my projects, and by the expediency of dictation, that I set aside my looking like a lunatic, and hope instead to assuage the pejorative opinion of the public by showing them the thrilling results of my chronic obsessive whispered instructions into my Phone. Unfortunately the results of my obsessive whispering into my phone usually consist of comics about me not shelving books, or comics about my finding weird Kafkaesque travesties throughout our library. I have begun to suspect those images may not have the opinion altering powers I was hoping for.
Not only do the instructions that I issue to the artificial intelligence lurking in my phone tend to be long and complicated, but, unfortunately, they do not result in great success nearly as often as I hope they will. Don't get me wrong, the things that AI produces are triumphs of the most powerful creative tool I have ever encountered, when they work. When the Artificial Intelligence works like I want it to work it is like a dream of superpowers coming true, like flying, or shooting lightning bolts from my fingers. Unfortunately it only works like this about five percent of the time. Thirty percent of the time it works poorly, or not at all. And the remaining 65% of the time comes down to some version of tantalizingly close to working in a way that causes me to think it surely must be correctable, but probably isn't and will cause me to nearly lose my mind if I'm going to insist upon trying. I don't know how amazing or ridiculous this sounds as a venture. But writing it all out it puts me in mind of someone plugging dollars into a slot machine with endless fervor. And before we dismiss this with easy disdain as an image of the pathetic obsession of a gambling addict, it requires one analytical correction: I am playing with house money here! The gambler is a doomed loser as he or she grimly feeds the machine, but I grimly feed the machine and go home loaded every day. (Weirdly this is probably a pretty good analogy for most people's jobs, but maybe set that aside). The important thing is that perhaps you can understand then how this causes me to be persistent and to try over and over again, which causes me to whisper into my phone all the more frequently to get the desired result.
But because, like with the person on the dollar slots, the volume of my entries becomes such a key element, I rarely examine very closely what I’ve actually said to my phone. I simply hope for the best unless I know that there is some kind of chronic misspelling mistake that I have to go in and manually correct. However, one evening recently, after issuing a long set of instructions into my phone, before entering that information I decided to read over what I had said. The AI seemed to be getting so many things wrong I wanted to see if part of the problem was on my end. Perhaps my minutes of rambling description, begging, and exacting clarification was not as articulate as I hoped. Maybe even the transcription of my voice into words was not working as well as I hoped- certainly my furtive whispering in dark corners couldn't be the easiest thing for my phone to interpret.
So I read what I said.
What do I sound like when issuing instructions to an AI? What do I sound like when speaking to an eager to please, lying, wildly idiotic genius assistant who is always eager to do what I want and is sure it can, but can actually only manage it about five percent of the time?
Unfortunately, to my horror, I sound a little like the current President of the United States.
It is not pretty.
No comments:
Post a Comment
If you were wondering, yes, you should comment. Not only does it remind me that I must write in intelligible English because someone is actually reading what I write, but it is also a pleasure for me since I am interested in anything you have to say.
I respond to pretty much every comment. It's like a free personalized blog post!
One last detail: If you are commenting on a post more than two weeks old I have to go in and approve it. It's sort of a spam protection device. Also, rarely, a comment will go to spam on its own. Give either of those a day or two and your comment will show up on the blog.