site banner

Small-Scale Question Sunday for February 16, 2025

Do you have a dumb question that you're kind of embarrassed to ask in the main thread? Is there something you're just not sure about?

This is your opportunity to ask questions. No question too simple or too silly.

Culture war topics are accepted, and proposals for a better intro post are appreciated.

2
Jump in the discussion.

No email address required.

I agree whole-heartedly with @MathWizard .

I write my own novel(s) for two reasons, one being that I want to establish, for reasons not much more noble than "I was considered a good writer back when that was a rare commodity", the other being that since there's very little of the very niche genre of fiction I enjoy, I might as well write my own.

Some authors view the rise of AI writing with fear and panic. I don't, both because I never relied on it for an income (and of course I'm worried about my own job going away) and I read far more than I write.

My default existence is boredom and ennui, desperate trawling and doomscrolling for literature (or any text) that assuages that thirst. There are never enough of the novels I enjoy, the authors I admire never seem to churn out novels faster than I can read them. Some of them, like Banks, are churlish enough to die instead of keeping on writing indefinitely.

Thus, even if I lose a small amount of utility from my writing becoming obsolete, I gain a great deal more from just having a near endless buffet of good writing to read.

Some would deny that AI can ever do that. Those people are idiots.

I think current LLMs are about 70% of the way to replacing me as an author, assuming they're given an excerpt of text I've written and then told to go wild with it. It's only a matter of years before the matter is settled. A year, if I'm being both highly optimistic and pessimistic.

So it goes for most people. For those who aren't making music as a career and who don't reflexively hate AI made music, the endlessly availability of fresh new music in the genres they like is amazing. While surgeons wouldn't be happy about being replaced, someone who got their appendix removed cheaper and safer would be happy about it. At any rate, it's coming for us whether we like it or not.

I'm assaulted daily by anti-ai takes that sap my will to live, but one in particular sticks in my mind. It went something like comparing generative tools to a child throwing his toy box into a ceiling fan, because the end result is the same as him playing with them. If taken as anything other than babbling of a motivated partisan, that metaphor would reveal how little the poster thinks of his readers, and of himself as well. His writing might only have value for the fun he has while producing it, and nothing else. But that's not normal.

Obviously if one really cares about the output and not the craft itself, then it's different. But I don't feel like that is the central example of "I'm writing software for fun". But yes, if that's you (and @MathWizard), you might decide to hold off based on your personal evaluation of how quickly we will have AI tools available to do that for you.

I would say I'm not convinced that will be soon. Right now AI generated content sucks ass. I'm not reflexively against it (though you might not believe me), but it just is bad and not worth my time even if I want to Consoom (TM). And that's just talking about creative endeavors (not my area of expertise), where I can only really judge the output from a layman's perspective. If we are talking about programming (which is my area of expertise), AI is laughably bad. It's so bad at writing code that it's a drain on productivity, because you have to check everything it does to make sure it is correct. LLMs are a really neat party trick, but right now that's all they are. They still can't actually do anything useful.

Could it get better? Sure. But people have been saying "it's gonna get better in a year or two" for years now, and it still isn't there. At this point, there's not evidence to suggest that AI is going to be able to do these tasks competently within the next 5 years or so. That doesn't mean it can't happen, but I think that people predicting AI to do all these wonderful things (meaning no personal disrespect to you) are getting caught up in hype without any substance behind it at this time.