What I think of AI: A hot take

My relationship with AI is complicated.

First of all, generally speaking, I don’t like it. I think it is overhyped. I think AI companies have stolen copyrighted material for training their models without compensating the rights holders. I think AI companies are overvalued and are spending too much money and using too many resources (electricity, for instance). I think the value of AI isn’t anywhere near what the AI companies say it is. (If they really believe their own publicity, I don’t know).

But it’s also interesting, and I would be lying if I said I haven’t played with it a little. The ability of LLMs to summarize material is fascinating.

Mostly, though, I stay away from it. I don’t find myself drawn to it, though I have a couple of LLMs downloaded to my machine. The only time I’ve let AI touch my writing has been using the Rephrase function in ProWritingAid editing software. Almost every time I turned it loose on a questionable bit of my prose it has given me hilariously wrong alternatives. A few times I found its suggestions useful as a starting point for rephrases of my own. Only once has it given me the perfect rewrite.

I’ve also turned Google’s NotepadLM tool loose on the manuscript to one of my unpublished novels. It came back with a useful mindmap and an amusing two-person spoken analysis (somewhat like a podcast) of the work. Overall, it was a pretty accurate look at the novel, but somehow completely missed a critical characteristic of one of the main characters.

So I’ll continue to leave Apple Intelligence turned off, while reading blogs and newsletters from people like Ed Zitron and other AI skeptics. I’m not optimistic that all the money Microsoft, Google, OpenAI and others are pouring into this technology is going to be worth it, but I don’t trust the big tech companies anyway. Why should this be any different?