Half-Penny Thoughts — 10apr26

This morning, I asked Claude to provide a general evaluation of the poetry on this site, mostly for shits and giggles. Sometimes I feel I am too close to my own poetry to critically evaluate what I write, so I felt it would be helpful to get the opinion from “someone” without a stake in providing me feedback — not cog greasing or a sense of friendly obligation to make me happy. [Note: Claude was not informed that I am the author of this site, so it has no reason to lean into saying nice things to me about my poetry.]

Claude’s overall assessment and take-home:

This is thoughtful, atmospheric, small-press-style lyric poetry with a strong sense of personal mythology. It won’t appeal to everyone — it’s firmly in the tradition of neo-pagan, nature-mystic, and bardic contemporary poetry — but within that tradition it is accomplished and genuine. The daily practice evident in the blog suggests a writer who takes the craft seriously, using form and constraint to keep the work honest. The best pieces are genuinely memorable; the weaker ones are still interesting as evidence of a working poetic mind.

Okay, well… I’ll gladly take that…

The full feedback was longer and contained some useful critical comments that I was self-aware enough to have already identified as potential issues with certain pieces: “…the very short form can occasionally tip from minimalist into underdeveloped“, “thematic repetition“, and “…some poems lean into obscurity in ways that feel less like earned ambiguity and more like vagueness“.

All of these are valid criticisms and a part of why I have dialed back my poetry to lean more into fiction and prose over the past few months. I felt that I had been falling into a bit of a rut for the since before Christmas and I wanted to give the poetry centers of my brain a bit of a rest before returning to writing poetry on a consistent and regular basis. I thought I needed to change gears before things rusted out.

But, I don’t want to focus exclusively on the critical. Claude also pointed out a few positives.

Some identified strengths:

  • Compression and restraint: Raven excels at the short, imagistic poem. He doesn’t overexplain, and the best pieces land with the force of a koan.
  • Consistent voice: The lowercase aesthetic, the moody vocabulary […], and the mythological undertow create a recognizable, self-contained world.
  • Constraint-based creativity: Using runes and tarot cards as daily prompts produces genuine surprise and prevents the work from becoming self-indulgent.
  • Intertextual richness: The poems reward readers familiar with Norse mythology, Elder Futhark symbolism, and folk tradition, without being inaccessible to those who aren’t

There was more, of course, but what I’ve withheld is less important (specific examples while assessing the pieces, more in-depth analysis that explains the key notes).

As I said, it was mostly done for fun, but I do take seriously these observations because it is useful for growth. Sometimes, in the age of social internet, we get a little hung up on the stars and the positive reinforcements — making sure we give them as well as receive them. It is useful to receive an unemotional, non-Pavlovian and almost clinical assessment of my writing to identify strengths and areas for improvement. I might have thought I was pretty hot shit when I was in my twenties and early thirties, but I’ve come down to earth (hard, in ways) in my middle age and see that there is always room for growth and improvement.

What kinds of ways do you “check” your writing progress? Has anyone else tried using AI (Grammarly, ChatGPT, Gemini, etc.)? If so, did you find anything useful? Annoying? Infuriating? Michael, WTF are you doing embracing our future AI overlords?

Let me know what you think below, if you are willing.


13 responses to “Half-Penny Thoughts — 10apr26”

  1. missparker0106 Avatar

    I’m glad you find AI useful. For me, it’s useful as a research tool–period. If you find help with its critique, that’s great.

    I’m currently working on an autobiography (not mine–I’m too boring for that–I’m ghostwriting). I let the programmed AI point out spelling and possible grammatical errors (it and I disagree constantly), but I can’t envision a time when I would look for its approval/disapproval or even an opinion. I’ll leave that up to subject of my project when the time comes.

    I understand and appreciate it may be different for fiction. Something tells me my distrust would still lead me away from AI as having the ability to add anything meaningful, were I to engage in fiction. But, I’m old hat, and I admit to it. Meanwhile, it’s fun to sit on the sidelines and observe how well it works or how much it fails with others. Thanks for sharing!

    1. michael raven Avatar

      I’ve been working with multiple services for a number of things, mostly for research (personal and work), and work-related tasks. I’ve found on the whole that the quality can vary dramatically from free services to paid services alone, with some models are infuriatingly terrible, paid or otherwise. Others… are somewhat impressive across the board.

      I’ve found Claude to be fairly useful and competent with most tasks I’ve given it. I have found some errors, but fewer than most models out there. It doesn’t do images, but that’s just fine by me.

      Gemini Pro feels less competent, but that’s because it tends to try to be everything to everyone. The image generation can be an exercise in frustration or, occasionally, it seems to know what it is doing. Occasionally. I’ve caught it hallucinating more often than seems reasonable.

      MS Copilot is decent if you keep it to science-based agents, terrible for most other things. It’s slop outside of science and math, and it hallucinates on the free variations to the point of being useless. The image generation is a waste of time in my opinion, but most of it is to me.

      Others: Opera has too many guardrails, Lumo is better, but still has too many guardrails and hallucinates more than it should.

      On the Grammarly level, I’m unconvinced it is helpful at all.

      Some tools are better than others.

      1. missparker0106 Avatar

        You have done your homework, sir! If I ever feel compelled to veer away from human editorial, I would go with Claude based on your findings. This is some good intel. Thank you!

        1. michael raven Avatar

          As I said, it has been more for personal research and work-oriented stuff that has driven my trying to find a use for something that feels as inevitable as the smartphone felt back in 2007 or the internet in 1996 felt.

  2. Chico’s Mom Avatar

    What little AI I use is Google AI. While I’m happy to read that you got some beneficial notes, poetry is about all the things AI doesn’t understand (yet). A broken heart. The sadden around life. The joy of seeing an old friend. Falling in love. Loneliness. Beauty. Desire.

    Please don’t let Claude stop you from writing up a storm. 🖋️

    1. michael raven Avatar

      Oh, it isn’t! I didn’t mean to make it sound like this silly little exercise was discouraging at all. I actually found it a useful thing to understand where there were strengths and weaknesses. Weaknesses are areas for development, as far as I am concerned.

      And no, I don’t expect the AI to understand those human nuances. It was an overall positive exercise for me, more than I gave it credit for when I initially did it on a lark.

      1. Chico’s Mom Avatar

        I understand. And we all have room to grow. 💪🏻

  3. steveforthedeaf Avatar

    I’m encouraged by my company to use Ai at work. I’m almost positive I can tell when people have done so. I tried augmenting my novel The Suicide Pack with ai promo for fun as it all seemed very new to me and exciting at the time.

    It went down very badly and woke me up to how many people view it as not a fun graphics tool but the enemy of artistic endeavour.

    If it weren’t for all the negative implications the freedom it could bring to play would be huge. But I don’t want to use it to make my story seem like a multi media ad campaign if the same software is being used for deepfake nudes or scam calls in other areas.

    I lean heavily on Grammarly at times though. Does that count?

    1. missparker0106 Avatar

      I agree with much of what you’ve said. I can’t condone my own personal use of AI, especially when it looks like I’m supporting the very thing I feel is killing creativity and mocking us with its own versions.

      And, like you, I find Grammarly useful when I need to know, for instance, the use of “capitol” vs. “capital,” “effect” vs. “affect,” things like that. When I use it, I’m not asking Grammarly for a possible rewrite or how it feels about my writing–just rules.

      1. steveforthedeaf Avatar

        Same, It’s an interactive thesaurus, spell checker, word counter. Happy with that. But WRITE LIKE THIS AUTHOR… If I want to do that I want to do that. For my own fun. Not have a computer do it for me.

    2. michael raven Avatar

      Same here. At work, it was an inevitable tool that would eventually become mandatory.

      So, instead of fighting it, I decided to put it through the paces to see what works and doesn’t these days. I’m walking away far more impressed than I was a a few years ago.

      I have an artistic family. My kids are mostly against its use, but my eldest is starting to see some of the potential. She’s not into it, but she is less vehement against its use than a year ago.

      I generally prefer the text-focused AI over the image-generation and artistic ones. Most of it still seems like slop, although there are a lot fewer weird postures and extra digits with the better models.

      Grammarly counts, although I doubt their wisdom in using (without permission) writers to train their AI. “Write like Stephen King!”. That feature got pulled REAL quick after the backlash (and possible lawsuits).

      I think, depending on the model, it can be useful (I’ve used AI to generate a taiji-qigong workout plan for my chronic pain diagnosis and it seems sensible, for example).

  4. Bob Avatar

    I’m so behind in AI. You can ask AI to evaluate your poetry. Seems interesting.

Leave a comment. Markdown use is permitted.

This site uses Akismet to reduce spam. Learn how your comment data is processed.