I never get coupons for what I actually buy, I only get coupons for things I don't buy. Those things are sometimes related, but if I wanted the other brand, I'd already be buying the other brand.

I take an iterative process where I get to experiment with the technology, building agent harnesses, building workflows and processes, adversarial critiquing, and of course I spend a good amount of time hand editing.

Everything I've published I find sincerely good and to have value. My most recent book (~200 pages), published last week, has been read exactly one time and has one 5 star review. So right now it's not a commercial success, but I both enjoyed making the book and I think I made a reader happy. That's joyful to me.

That book took probably 12-16 hours of work from me. Writing it by hand would have taken me weeks, if I had the skill to do it, which I don't. But I do have the skill to shape a draft into something nice to read.

So, why? It's fun to create. It's fun to publish. I think I'm providing value. I'm learning about writing and editing and I'm learning about how to best use AI.

My experiments with AI writing are additive. Nothing I do with AI prevents human writers from writing.

I do think there's an argument that AI floods the field and makes human writing harder to find, and overall diminishes the commons.

But I want to provide value and am trying to with my experiments. The market will decide over time.

Human review is the bottleneck at my office. I think software will be more and more AI written but also more and more AI validated.

For my nonfiction series, I have attempted adversarial agent fact checking, and I rerun it with newer models as they come out.

So: little human review of citations. I didn't expect this thread to shine on me favorably, but perhaps it's interesting to some.

It was another outlet to explore using AI outside of my 9-5 programming. I've experimented with building custom agents to write stories, custom memory storage methods, custom workflows, etc. I've also used AI to write content management systems and experiment with that.

Like with software, even if "slop", there's a particular joy to having an idea come to life very very quickly. I can start with a premise, a character, a few scene ideas, and get a workable draft.

For some of my ideas (outside of erotica), there are no books that I can find that cover what I'm interested in. So in another sense it's like using AI to do research for my own enjoyment, but I'm also tightening it up and publishing it.

> What sense of ownership/pride do you feel when you consider one of your works as "done"?

It's diminished compared to things I've done completely by myself. But it's non-zero and it's positive. I've actually shipped products rather than sitting on them, and there's joy and pride in that.

But I do still put a lot of work into it, both on the software side managing it, but also manually hand-editing it. I rewrite things until it doesn't feel like AI to me, even if I also use AI to make some of those edits.

> What would you do differently, if you were back at the starting block and had a budget of $800 to spend on writing nonfiction?

The lessons are very similar to other lessons on here around software projects.

Like with game development, don't spend your time on the engine. I'd spend less time building scaffolding and structure software.

I'd also be more patient and wait for weekly resets instead of bumping up to the higher tier.

I'd also stick to fewer projects more closely until completion -- I have a queue of about 80 some books to go through before I can publish them.

And then I'd say spend more time on the marketing side of things.

> Because romance sells, the professors thought it would be the genre most susceptible to A.I. intervention, but instead it was nonfiction...

I've been working mostly in erotica but have shifted to some nonfiction series lately.

I stopped for a while, but Fable and GPT 5.6 Sol have reinvigorated me.

I use Claude Code and other CLIs to manage the process, create first drafts, covers, research, etc. I heavily edit most things before I publish it.

The first question is probably one of profitability. So far I have made a little over $40, compared to probably $800 in AI subscription fees. So from that point of view it's been an absolute disaster.

If anyone is curious, I'd be happy to talk about any of it.

As in, the puzzle does not have a unique solution using sudoku logic alone. You must use one of two forms of hidden word logic to finish the puzzle. 1) Deduce where a word must be placed and add letters, or 2) deduce where a word may not be placed and restrict letters. To keep it more sudoku like and aid in solving, every row, column, and box has at least one letter of a found word.

In playtesting I thought it was fun, although at times very difficult for me (I'm not an expert sudoku player), so I thought I'd build it out to share.

The puzzles are generated by creating racks of 9 unique letters, finding all words that can be made with those letters, generating sample grids, finding the words in them, throwing out candidates that don't fit the word placement criteria, then carving the board by removing letters until two constraints hold: we have >=2 solutions by sudoku logic and exactly 1 solution when the word search logic is added. Overall about 60000 puzzles are generated and thrown out for every sound one.

I used Claude pretty extensively for the code, which includes:

- Puzzle generator - Puzzle solver - Puzzle explainer - Static website (Cloudflare) - Online puzzle player (local storage) - Online static puzzle walkthrough - Print book layouts - PDF layouts

Overall, it's been a lot of fun to focus on for a few weeks.

https://www.sudokuwordsearch.com

Comments URL: https://news.ycombinator.com/item?id=48585510

Points: 3

# Comments: 0