A Hacker News submission points to an observation from the Humanity’s First Exam project: five frontier large language models reportedly gave the same answer when asked about their favorite joke. The available metadata does not identify the models, specify the exact prompt, reveal the joke, or describe repeated trials. It is therefore best treated as an intriguing clue about shared model behavior and training distributions, rather than as evidence of a stable or meaningful common preference.
No heat snapshots are available in the last 24 hours.