Feedback is our weekly column of bizarre stories, implausible advertising claims, confusing instructions and more
Pussycats on parade
Some topics appear to be catnip to Feedback鈥檚 readers, and one of them is cats. In our 23 September edition, we described a study in which pet cats were fitted with accelerometers. This revealed that they spent the lion鈥檚 share of their time doing absolutely nothing, with 20.1 to 41.4 per cent of their time spent sitting down and a further 34.8 to 52.9 per cent devoted instead to lying down.
We suggested that these cats were a tad slothful, and might even be edging towards idleness.
Advertisement
We received some feedback on this opinion. Reader Jack Haley sent us a lengthy text 鈥渃ourtesy of Google AI鈥� (put a pin in that; we鈥檙e coming back to it), recounting the experiences of , a geologist in Australia who took part in a in the 2010s.
The results, as outlined in Haley鈥檚 text, were startling: 鈥淪quid, a heavier ginger tabby, was mapped taking regular nighttime strolls to the local high school and a nearby shopping center. Semi, a black- and-white cat, went much further, venturing a staggering three kilometers away from home.鈥�
Furthermore, 鈥淪tudies like the Cat Tracker project have shown that roughly 40 per cent of cats whose owners believe they are indoors or close by at night actually range over a hectare or more, frequently engaging in secret, long-distance neighbourhood explorations.鈥�
Curious, Feedback put our exploring paws on and eventually found that the was launched from the University of North Carolina before being in Australia (and there鈥檚 now a Cat Tracker 2.0). Its final results were in 2020 in Animal Conservation. 鈥淲e worked with citizen scientists to track 925 pet cats from six countries, finding remarkably small home ranges,鈥� the feline watchers wrote. The average range was 鈥�3.6 hectares鈥� (0.036 square kilometres) and in the sample only three cats ranged over more than 1 square kilometre.
Readers may notice some subtle differences between how the AI framed the results and how the scientists did. Technically, if you think about it, the AI did get the core facts right, and yet, Feedback still thinks cats are quite lazy.
Meanwhile, assistant news editor Sam Wong sent us a study titled 鈥淲ealth, health, cats and Toxoplasma gondii: An ecological correlational study of subjective well-being across 93 countries鈥�. This found that 鈥渘ational happiness was positively associated with cat density鈥�.
They mean 鈥減opulation density鈥�, of course, but Feedback is also pretty sure that cats with greater physical density have significant effects on the happiness of their human attendants. One of ours can contract until she has a density approximating that of a neutron star when she wants to sleep on one of her humans.
Raging hellfire
A lot of lectures suffer from a dire lack of relatability. What, a prospective audience might wonder, does the evolutionary divergence of Cambrian lophotrochozoans or a drag-induced skin effect in a Bose-Fermi mixture have to do with my life?
No such relatability limitations affected a on 8 October at Utah State University by Miranda Judson, a graduate student in philosophy at Arizona State University. The organisers noted that a announcing it had unexpectedly gone viral 鈥� and it鈥檚 no wonder.
The talk was titled 鈥淲elcome to Hell: A theodicy鈥�. The subtitle went on to clarify that 鈥淭his talk argues that the evidence of our world is such that the probability we鈥檙e in hell is surprisingly high. If we鈥檙e in hell, we鈥檇 expect to see the evidence we actually do see鈥�.
Alas, the talk wasn鈥檛 streamed online, and Judson doesn鈥檛 appear to have yet published a paper on it. But honestly, given the general state of things, we feel it wouldn鈥檛 be too difficult to fill in the blanks. We鈥檒l report back if we learn more, hopefully not via the medium of a demonic voice in our head.
Bedtime tales
Speaking of evidence that we鈥檙e in hell, readers may have heard of the AI engineer telling his wife that he doesn鈥檛 need to read to his kids at bedtime because he can just train an AI to do it for him. If you haven鈥檛 heard of this, well, now you have.
The engineer in question is Alvin Sng, who works for a company called Factory. Sng the following on X on 7 October: 鈥淢y wife: Can you read our baby a book? Me: I can just play an audiobook from my phone. My wife: But then it won鈥檛 be in your voice. Me: I can train a voice model.鈥� His conclusion was: 鈥淧arenting hits different when you work in AI.鈥�
Feedback isn鈥檛 going to dunk on this, because a large proportion of the people with access to the internet already have, and at this point it鈥檚 too easy. However, we will note that Sng has since deleted the post, and on 8 October : 鈥淥k, enough jokes for a while. I鈥檒l go back to reading bedtime stories for now.鈥� The replies to this post suggest that people don鈥檛 believe he was joking.
The trouble is, of course, that the claims routinely made by AI bros are so over the top that it is hard to differentiate a parody from the real thing. The internet adage Poe鈥檚 law describes this phenomenon for extreme views, which are frequently so ridiculous that any parody must be clearly marked as such.
Perhaps we need a corollary, something like: 鈥淲ithout a blatant display of humour, it is utterly impossible to parody antisocial AI hype in such a way that it cannot be mistaken for the genuine article鈥�. We could call it Sng鈥檚 law.
Got a story for Feedback?
You can send stories to Feedback by email at feedback@newscientist.com. Please include your home address. This week鈥檚 and past Feedbacks can be seen on our website.
