Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Binary file not shown.
15 changes: 15 additions & 0 deletions 2025/Whodunit? Writing Murder Mysteries with LLMs/README.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,15 @@
# Whodunit? Writing Murder Mysteries with LLMs - James Heller, Apartment 304

[Expanded Post](https://blog.apartment304.com/whodunit-llm-murder-mysteries)

[Slides](2025-whodunit-lightning.pdf)

[Play **Whodunit?**: Solve AI authored mysteries](https://whodunit.rip)

Can a large language model craft a brand-new murder mystery on demand? Explore a mystery-writing AI and the tips and tricks that help it generate coherent, logic-driven mysteries for players to solve. The system writes backwards from the solution, breaks the story into structured steps, and uses Temporal to manage retries when APIs fail. The result is a low-tech game powered by cutting-edge tools.

## Biography

James is a software engineer at Apartment 304, which he co-founded in 2016. He works primarily on cloud architecture, application, and data engineering projects, and enjoys using Go whenever he can. He automates Halloween decorations and recently enlisted AI’s help in writing murder mysteries for game night.

The End?
2 changes: 1 addition & 1 deletion README.md
Original file line number Diff line number Diff line change
Expand Up @@ -176,7 +176,7 @@ Panelists: David Soria Parra, Gari Singh, Ian Cottrell, Jaana Dogan

### Lightning Talks

#### Whodunit? Writing Murder Mysteries with LLMs - James Heller, Apartment 304
#### [Whodunit? Writing Murder Mysteries with LLMs - James Heller, Apartment 304](2025%2FWhodunit%3F%20Writing%20Murder%20Mysteries%20with%20LLMs)

Can a large language model craft a brand-new murder mystery on demand? Explore a mystery-writing AI and the tips and tricks that help it generate coherent, logic-driven mysteries for players to solve. The system writes backwards from the solution, breaks the story into structured steps, and uses Temporal to manage retries when APIs fail. The result is a low-tech game powered by cutting-edge tools.

Expand Down