site banner

Friday Fun Thread for July 3, 2026

Be advised: this thread is not for serious in-depth discussion of weighty topics (we have a link for that), this thread is not for anything Culture War related. This thread is for Fun. You got jokes? Share 'em. You got silly questions? Ask 'em.

1
Jump in the discussion.

No email address required.

https://www.hyperstitionai.com/unslop-results

I came across the results of this contest to generate LLM-written stories, with one of our participants being our very own @self_made_human. I'd be interested in any thoughts you had on participating in the contest as well, if you had any.

My main takeaway after skimming the finalists was that while some of the concepts were interesting, even with a lot of prompting and harness effort, there's just not much difference from the sloppy prose and incoherent writing you get when one-shotting text out of an LLM.

It's hard to tell how much of this is just because writing's not a focus at the LLM labs right now, but the fact that every LLM converges onto similar writing attractor states definitely makes me bearish on LLM's ability to continue to generalize.

Glad someone noticed.

The main takeaway? The competition was an absolute pain in the ass. Reading hundreds of thousands of tokens of AI generated prose and creating a working agentic harness + scaffolding was exhausting. It took dozens of cycles of iteration to find the right balance.

Funnily enough, what really made all the difference is that I ended up sharing a large corpus of my own text for reference - and it shows. The story I'm the most proud of is "The Bowl" (which is not nonsensical, at least to anyone who has or had a Labrador), but both my entries have clear self_made_human fingerprints. The Bowl made me sob while reading it for the first time, and I knew I wasn't going to get anything better than that. That's more than most human literature can ever say.

To the chagrin of people complaining about my use of AI assistance, at times, I know what I'm doing. Gwern, Roon, Alexander Wales and a few others think I know what I'm doing, and liked the output. Is it a surprise that I'm tired of arguing with idiots here when I could be off making money? $500 just about covers the effort, though I would have liked the full 10k. Ah well, I introduced myself as a dark horse candidate who might be worth giving a few hundred dollars in tokens just to see what he could do. And it worked.

Here's a public conversation I had with Gwern:

https://old.reddit.com/r/slatestarcodex/s/QiMdeD3OkG

I have plenty of additional thoughts, tips and things learned through trial and error plus bitter experience. Unfortunately, they're proprietary. Anybody who wants them better pay me for the privilege.