#authors beware

LLMs are plagiarism machines, and beyond that make the Internet worse, but even if you ignore all of that, there is still the massive harm data centers cause to individuals.

Look, I love technology, I understand the risks of any technology. But I also know that one has to draw the line ethically when one can.

I've been going to Barnes and Nobles since I was a tyke. I remember my dad driving my sibling and me to a Barnes and Nobles to try to convince us that reading was good actually and we should be addicted to reading. I bought most of my book collection from there and independent bookshops.

I've been doing a book store report on Barnes and Nobles and can tell you their stocking by year and the diversity report since I was 13. (Also I can tell you the now dead Borders and Walden too)

I can tell you which books I bought from Barnes and Nobles and the stories of how I found them. Book of Alta, I bought that in London Barnes and Nobles just before the stairs as I was turning and the cover caught my attention and I was drawn by the fact I'd never seen Jane Yolen in the United States. I still have that book and have fond memories of buying it and associate it with my trip to London with my parents when I was 9. I think my sibling bought yet another Terry Brooks book...

But this is a good way to make sure no one wants to stock their book in your store.

So writers and authors, beware this is happening and think harder how you are going to change your strategies. Discuss it with your agents and editors too. Because you fighting AI slop will make your life worse. More accusations will happen towards you if Barnes and Nobles lets this happen.

And fight like hell to show all of the effects the data centers will have on individuals to the CEO himself so he can feel guilty about his support of data centers. Note: I had to update my Gen AI post with this info. I hate this. I'm about to write a protest letter to break their hearts. I encourage you to do the same:

In August I posted an update to a long-running story of mine. A few days later, I received a detailed, engaged comment. It was easily the most elaborate comment I ever received, and I was delighted.

It praised a visual sci-fi description I was rather proud of, some subtle character work, a gut-punch moment and the visceral closing scene. It felt like this reader was seeing precisely the features of the text where I put the most care.

I replied warmly, thanking them and offering some mild spoilers.

A week ago or so, I found out that the comment had been deleted. That stroke me as odd, and I went back to the conversation (I still had the original comment in my inbox).

I've become convinced the comment was AI generated

I have a rather in-depth knowledge of current AI tools from my job, and with the right prompt these tools are perfectly capable of reading a story and then generate three paragraphs of warm, engaged, natural sounding content highlighting the best crafted parts.

Why?

I think the purpose was training. The AI was prompted to generate the comment, I reacted positively to it, my reply was filed away as a "success" data point, the comment was deleted to cover their tracks.

Unfortunately there's nothing to do with this realization; I played into their game and cannot take it back. YMMV, but for me the biggest damage done was the break of trust. Now I have added an additional layer of wariness and distrust towards readers' comments, and that sucks.

The takeaway

As I described above, the comment was extremely sophisticated. It didn't read off or stiff; it read perfectly human, even though in hindsight I could identify some mannerisms which lean AI.

But I think overall the real smoking gun was the content. The comment was "too good to be true": it insisted on the parts I put more craft into, i.e. into the parts that were (1) objectively better from a narrative/prose standpoint, because more curated (2) the parts I was more eager to see validated/recognized.

A human reader will bring their own perspective to the table; they might latch onto a minor detail you didn't, misinterpret a theme in an interesting way, or be completely silent on the thing you worked hardest on. Their comment is a record of their unique interaction with your story, not a mirror reflecting your intentions back at you.

That I think is the strongest predictor: if the comment says exactly what you wanted to hear, it's AI (and that is true of any kind of feedback in general).

Wow.

So, Scribd dumped the romance/erotica genre a coupla’ years back because...people were reading more??

And the nonsense doesn’t stop there, alas...

Romance Novel Reviews - All of the Romance, None of the Bullshit

"On the other hand... is she levitating? Is her leg merged at the knee joint into the standing woman's hip? Is this a yoga pose from the advanced partner yoga class I've never had the gumption to try? When you meet Reynolda, is there gravity? And whose shoe is that behind the "g" in "Meeting?" 

I may have hit my quota of question marks. Which is a problem, because I have a lot of questions. 

Concerned with work-related matters, I awoke at 4:30 a.m. today from a stress nightmare. I couldn't get back to sleep, so I decided to just go to work. Which is what I did. At 4:30 a.m. When normal human beings were sleeping.

Now you all WANT THINGS from me. You have QUESTIONS and NEEDS, and you JUST KEEP COMING. You don't understand why I haven't finished reading your manuscript yet. You want to know why the copyeditor won't get back to you. You want to know what that harpy of a marketing manager is doing instead of getting your precious book on the NYT bestseller list. And you all sound so damn perky on the phone.

Authors, to date I have never committed murder, but there is a first time for everything and I have an unnecessarily sharp letter opener on my desk. What I mean by this is

I

AM

IN

NO

FUCKING

MOOD.

Sincerely,