(C) Daily Kos
This story was originally published by Daily Kos and is unaltered.
. . . . . . . . . .



Failed Writer's Journey: Art, Thinking, and Imitative AI [1]

['This Content Is Not Subject To Review Daily Kos Staff Prior To Publication.']

Date: 2025-03-07

Lincoln Michel (excellent writer with a new book Metallic Realms coming out soon. Hint. Hint. Hint.) had an interesting Bluesky thread a few days ago that gets to a significant problem with imitative AI:

There's an amusing gap between AI boosters who keep talking about how writers "must embrace new technology or be left behind!" and actual writers who spend half our time figuring out how to AVOID technology (locking cell phones away, Freedom app, pen and paper, etc.) to be able to actually write. Part of the gap I think are non-artists assuming that making art is the same as any other "job" and the goal is basically "good enough and as quick as possible" like it might be in most workplaces.

I think this is at least directionally correct (man, I have been in corporate America far too long ….) and I think it highlights imitative AI’s biggest problems. I am not so much worried about imitative AI replacing art. At this point, I don’t think that it can. Businesses will try, but in the medium term it’s just not going to be acceptable to enough of their clients and consumers. The material it produces is subpar compared to professional artists and there doesn’t seem to be an upward trajectory in it. Which makes sense — as I have spoken to before the math will tend toward the middle and hallucinations are impossible to avoid, so you aren’t going to get much in the way of good. That is not a problem in other areas, though, at least not at first blush.

A business email does not need to be a pleasure to read — it just needs to convey the information in the shortest amount of time. If you turn your email about TPS reports into an epic prose poem you are both awesome and about to get fired. If you need some boilerplate code, then it can seem like imitative AI is a faster solution than writing it yourself. But there are two significant problems.

First, the hallucination issue. You may not want your email about TPS reports to be an epic prose poem, but you do want to make sure that the email is correct with regards to the TPS reports. You cannot know that with an imitative AI solution. Same with the code. Imitative AI is known for introducing security issues and subtle bugs into the code that it produces. Without taking the time to really check its output, you cannot be assured that it is correct.

And that is the second, more serious problem. If you use imitative AI in your formative years or when learning a new skill, how will you pick up the necessary talent to check imitative AI? Writing, even crappy little emails about TPS reports, is thinking made manifest. Coding is similar — writing a program is an attempt to convey understanding of how a computer system can solve a specific problem. If you reply upon imitative AI, then you can, without even realizing it, turn your thinking over to the hallucination machine. And that is a much worse problem than the hallucinations and middling output.

Thinking is one of the things that makes us humans. The emphasis on Artificial General Intelligence highlights this, I think. The ultimate goal is something that can think in the way we do. By turning over your thinking to an imitative AI system, then, you are creating a situation where you never learn to think properly, at least about a given topic and possibly in general depending on how much you turn over. We live in a world where not quite good enough is acceptable in many businesses if its produced fast enough. And since imitative AI can do not quite good enough really fast, we risk creating a group of people who learn not to think but to rely on a hallucination machine for at least some of their thinking. And that is not good for anyone.

Weekly Word Count

About eight thousand, so a good week overall. I spent the last two days trying to re-think the characters and their motivations in anticipation of the rewrite, so no words per se but I think movement in the right direction. Anyway, 19 chapters into a planned 28 (not sure I will get to 28 in this draft. I tend to underwrite, especially fist drafts) so with luck I should be done in a few weeks.

Then comes rewriting. I actually find that process easier than working on a blank page. Reshaping something that already exists is to me easier than creating something from whole cloth. Not that I am good at either, see the sign above the door and all that, but rewriting flows easier for me.

Have a great weekend everyone!

[END]
---
[1] Url: https://www.dailykos.com/stories/2025/3/7/2308576/-Failed-Writer-s-Journey-Art-Thinking-and-Imitative-AI?pm_campaign=front_page&pm_source=more_community&pm_medium=web

Published and (C) by Daily Kos
Content appears here under this condition or license: Site content may be used for any purpose without permission unless otherwise specified.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/dailykos/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.