arXiv Fool's Day 2023, now with more chickens

Bookmark and Share

I had a feeling I’d be writing this post. Last week I only found six joke papers to add to my list, but today’s arXiv feed delivered a very strong update, with eight more, bringing the 2023 total to fourteen, tying last year’s record.

I did not, however, have a feeling I’d be reading the phrase “large overdensity in the overall distribution of chickens”. One of the highlights of today’s feed is this extremely well-researched survey of bounds on the chicken density function (CDF) of the universe. While a lower bound is well established, the paper makes the first case for an upper bound as well, \(10^{23}\) chickens per cubic parsec. Fortunately (for the chickens, if perhaps not for science), that probably falls short of the gravitational binding threshold for small animals, but if the density fluctuations are large enough, who knows. I definitely want to see followup studies on this one.

(Loosely) speaking of birds, we also have the creatively titled “ChatGPT scores a bad birdie in counting gravitational-wave chirps”, which pretty firmly establishes that ChatGPT is not as good as actual scientists at counting gravitational wave events. (With some neat visualizations, too!) However, ChatGPT may be good at other things… for instance, is it any better at writing astrophysics papers than counting? We find that topic explored in “Can AI Put Gamma-Ray Astrophysicists Out of a Job?”, which is actually really well put together: it shows a series of prompts structured to produce the sections of a paper, along with their responses from ChatGPT and commentary by the authors. The AI does quite a good approximation of sounding like a scientist (including recognizing the importance of beer), but the authors conclude that their jobs are probably safe for a while.

Another paper suggests a more sobering perspective, though. It tackles the difficult problem of why, despite the thousands of exoplanet observations, no exomoons have been observed. Following up on recent work by Emmerich et al, we have to consider the credible hypothesis that the moons have been deorbited into their host planets by out-of-control AIs. It may only be a matter of time....

Moons were actually a fairly popular topic this year. For those who celebrate the Lunar New Year but get irritated that it jumps around to different points in the common solar calendar, never fear: you just have to wait 252 million years until the Earth’s rotation aligns with an integer multiple of the moon’s revolution period. Don’t wait to start planning your True New Year parties! Of course, by then, if the new naming scheme suggested by another paper catches on, we might just be calling it “Sun Db”. Or “moonmoon rock star Sun Db” for the pedants. No really. It makes sense.

There are even more where those came from, though! Check out the entire list of joke papers this year, and for previous years. And as usual, please let me know of any papers I’ve missed.

P.S. Even though it appears to be a legitimate project and not a joke, I have to give a shoutout to the Python library ronswanson which builds table models for the astrophysical computer modeling framework 3ML.