AI generates benchmark paper as a joke

By PulseAugur Editorial · [1 sources] · 2026-07-24 15:34

Ethan Mollick humorously prompted Codex to create a benchmark for AI's ability to create benchmarks, and then to write a paper about it. The AI generated a PDF document that, surprisingly, contained an interesting paper on the topic. AI

IMPACT Demonstrates AI's ability to generate creative content, even in response to humorous or meta-prompts.

RANK_REASON The item describes a humorous, non-serious prompt to an AI that resulted in an unexpected but not groundbreaking output.

Read on Bluesky Jetstream — AI desk →

other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

COVERAGE [1]

Bluesky Jetstream — AI desk TIER_1 English(EN) · emollick.bsky.social · 2026-07-24 15:34

As a joke I prompted Codex "Build and run BenchBench, a benchmark of now good ai is at creating benchmarks. then figure out what benchbenchbench is and run that

As a joke I prompted Codex "Build and run BenchBench, a benchmark of now good ai is at creating benchmarks. then figure out what benchbenchbench is and run that. and then write benchbenchbench up as a good arXiv paper." I got a PDF. But the paper is actually kind of interesting?…

COVERAGE [1]

As a joke I prompted Codex "Build and run BenchBench, a benchmark of now good ai is at creating benchmarks. then figure out what benchbenchbench is and run that

RELATED ENTITIES

RELATED TOPICS