paraschopra launches crowdsourced study on whether LLMs got funnier at original jokes
paraschopra · x · 2026-09-29
Researcher paraschopra is running a small study probing LLM progress in non-verifiable domains, using humor as the test case: can frontier models write better original jokes than older ones?
To prevent memorization, every joke must include two randomly chosen words, forcing novelty. LLM-as-judge results already show a clear improvement trend over time, and he's recruiting human raters (10 minutes) to ground those judgments; participants get results first by email.
Related event: Crowdsourced Study Tests Whether LLMs Can Write Original Jokes(2 posts)→
More from Research
- MIT builds AI-powered Raman 'barcode' to noninvasively detect senescent 'zombie cells' — MacrinePhD · 2026-09-30
- Anil Seth's biological naturalism essay against AI consciousness draws mass commentary collection — anilkseth · 2026-09-30
- Causal Inference Assumptions: A Practical Checklist Before You Run the Model — Nasereliver · 2026-09-30
- Sparse-View Gaussian Splatting Demo Shows 3D Reconstruction From Few Images — enndeeee · 2026-09-30
- Open-source self-driving guide adds VFH: binning LiDAR into 5° obstacle histograms — 4310sy · 2026-09-30
- Meta researchers show Qwen 3.x can run as general asynchronous agents without task-specific training — yresearch · 2026-09-30