There are days when I feel like I want to find a different world to live in; this post from Kevin Kelly has brought that feeling to the fore. Consider:

Some authors have it backwards. They believe that AI companies should pay them for training AIs on their books. But I predict in a very short while, authors will be paying AI companies to ensure that their books are included in the education and training of AIs.

Or perhaps...

If a book can be more easily parsed by an AI, its influence will be greater. Therefore many books will be written and formatted with an eye on their main audience. Writing for AIs will become a skill like any other, and something you can get better at. Authors could actively seek to optimize their work for AI ingestion, perhaps even collaborating with AI companies to ensure their content is properly understood, and integrated.

Kevin has certainly consumed large amounts of Kool-Aid on this one. Personally, I plan to keep writing for humans, even if that is seemingly obsolete.

@corbet personally I think llms are hot ecocidal garbage but one recent observation from work. Sone of our docs and marketing stuff is unreadable market-speak that badly fails from a "Plain English" point of view. It turns out that it's also terrible for AI oriented SEO.
There are far better reasons to improve legibility, but that's not one I expected.

@corbet
Last night I finished reading a novel that is short, poetic, multilayered, haunting and deliberately ambiguous.
I know Kevin Kelly is referring to his world of non fiction, "thought leader" books but even there the idea that influence would come through ease of AI regurgitation seems remarkably narrow. How many airport books are riffing off of Sun Tzu, thousands of years after he wrote? Paradox and ambiguity are powerful tools for shaping and strengthening thinking, as dumbbells are for the body.
Influence through LLM optimisation will have a similar effect as SEO had on writing on the internet.

@corbet that is why I renewed my abo of @lwn

@corbet This seems the positive part of such a trend. I'm more concerned about authors who will actively exploit AI training for their convenience. It happened for SEO and will happen again. Manipulation is the negative part of such a trend of our days.

@corbet Well… the second one does reminds me a lot of working to please the way platform implemented recommendations and user feeds.

@corbet

"Hello ChatGPT, can you give me a short but precise summary of the article linked above?"

Great! This is a very interesting task. The article is quite long, so asking for a concise summary is a tremendous Idea!
(...thinking....)
Here is a short summary highlighting all important points:
"nil"
Do you want me to summarize any other article for you?

@corbet
"If a book can be more easily parsed by an AI its influence will be greater" sounds a lot like "work for free and get paid in exposure".

Besides, literary influence is NEVER about writing a book similar enough to 10,000 other books to be palatable to the averaged-out reader. Even if it gets your work seen by more people, none of them will care or even remember what they read.

@corbet
I got laid off in June and with my free time I have been reading a lot of newer indie authors and am finding more and more stories that read like something you would get from Google GeminiCLI or Anthropic Claude CodeCLI.

So a lot of LLM training material.

@corbet I don't even buy the premise. A good author tries to get their points across with clarity, precision and persuasion. If that's not good enough for AI then we're doing AI wrong not writing wrongly.