I was firmly against AI years ago when it appeared. As much as I like technology I couldn’t understand how LLM would help me.
Now that I have used them they have helped me, but critically they haven’t done anything that I couldn’t have done myself. They have allowed me to lazier, and conversely to work harder.
What do I mean when I say that? Well the casual user of AI thinks it is magic. You put in a few words and you have a document. The downside is that you can’t trust that document, and its just a starting point not an ending one.
So that means that I have to double check every fact and assertion. I have to verify it with my experience and knowledge. I have to think with its logic and then compare it to my own. Often its logic is subtle and tricks you and that is one way that you have to work harder. Then after all of that is done, I am left with the feeling often and more increasingly as I use it, that I could have done it faster and more reliably without AI.
So how did I get sucked into using it? A friend asked me to use it and suggested it helped them do all kinds of things. Ok I gruelingly tried it and it did say a few interesting things. Still I wasn’t impressed. Then I found that it could find software or understand things better than a google web search and I was like “well I’ll just use it for this use case”. Then slowly I started trying it for other things and most of the time it didn’t work but a ton of social media was saying how useful it was so I thought I’d try some of those suggestions. Sure some of them worked, but I didn’t need that particular thing so it was little more than a toy.
Then it started to do more things like image creation (not interested) and other stuff. Sure this might help others but it wasn’t really going to help me.
Then I had clients ask about it and I figured I better get ahead of the curve and try things out and be able to provide some guidance. So I tested the heck out of it. Meh I’ve had better software.
AI is good at summarizing and giving uncreative people ideas. It can correct text, and make pictures into copyrighted styles, but what can it really do? It guesses at meaning, and I’ve been caught up in 3 things it made up and I didn’t realize it until later. AI has a very low trust factor with me, which is why even though I use it, I don’t trust it.
Its hard to praise something that you can’t trust.