快猫短视频

Would snails be better than whales for explaining big data? Maybe

Feedback's proposal that the genome of the blue whale could be used to communicate the scale of large datasets is knocked back by a reader with a radical alternative suggestion

Feedback is 快猫短视频鈥檚 popular sideways look at the latest science and technology news. You can submit items you believe may amuse readers to Feedback by emailing feedback@newscientist.com

How the whale crumbles

More on the topic of unusual units of measurement, in particular how to communicate the size of information. Attentive readers will recall Christopher Dionne鈥檚 suggestion that the scale of large datasets could be conveyed by comparing them to the genome of the blue whale (12 April).

Bruce Horton writes in with a firm riposte. 鈥淭he whole point of using blue whales to measure the size of things is that everyone knows how big a blue whale is, so we can easily visualise the length of anything measured in blue whales,鈥 he writes. 鈥淏ut most people cannot visualise the size of blue whale DNA, so that idea won鈥檛 work well.鈥

He has a point. Feedback is still scarred by the era of the Human Genome Project in the early 2000s, when we had to find comparisons to get across how much information is carried on our DNA. These often involved stacks of Bibles reaching halfway to the moon. Nowadays, we would just use the collected Wheel of Time books.

Fortunately, Bruce has a solution. He points us to a 2005 study in the parody scientific journal Annals of Improbable Research, which describes SNAP: . The researchers attached a giant African land snail to a two-wheeled cart, the wheels of which were CDs or DVDs. While the snail moved slowly, the data nonetheless travelled at 37,000 kilobytes per second 鈥 meaning the snail-based system transferred information quicker than existing broadband connections.

SNAP, Bruce argues, 鈥渋s a standard unit of measurement of data transfer that is easy for anyone to visualise and understand and is recommended for common use鈥.

Perhaps. While we await further correspondence, we wish to recommend a new unit developed by Ken Taylor and his wife. They have an orchard that includes some damson trees, which Ken describes as 鈥渘otoriously variable in yield from year to year鈥. Hence they have established 鈥渢he crumble鈥, which is a measure of how many desserts they can make per harvest. Ken reports: 鈥2024 was a very bad year 鈥 just 3 crumbles.鈥

Shock findings

鈥淲ell who would have thought it,鈥 says news editor Alexandra Thompson. 鈥淪top the press.鈥 She was drawing Feedback鈥檚 attention to a with the title: .

Compared with smaller cars, one of those great hulking SUVs is more likely to kill you if it hits you. Now, you might expect Feedback to snark at the sheer obviousness of this: yes, heavier objects hit harder than lighter ones, if they arrive at the same speed. But of course, one of the main virtues of science is the refusal to accept common sense for an answer, but instead to check things.

We hereby invite reader contributions in the category of 鈥渘o shit, Sherlock鈥. The more painfully obvious the discovery and tediously long-winded the experiment, the better. Do people enjoy picnics less if they鈥檙e overrun by ants? Does your water bill go up if you have a leaky tap? At least one enquiring mind wants to know.

Licking badgers

Historian Greg Jenner made a noteworthy discovery in April. Writing on , Greg says: 鈥測ou can type any random sentence into Google, then add 鈥榤eaning鈥 afterwards, and you鈥檒l get an AI explanation of a famous idiom or phrase you just made up鈥.

Greg鈥檚 invention was , which Google鈥檚 AI meant 鈥測ou can鈥檛 trick or deceive someone a second time after they鈥檝e been tricked once鈥. Um, first of all, the US electorate begs to differ. Second, this is, and we can鈥檛 stress this enough, completely made up. Yet that didn鈥檛 stop the AI a detailed explanation. 鈥溾楲icking鈥 in this context means to trick or deceive someone,鈥 it says, and 鈥渢he phrase likely originates with the historical sport of badger baiting鈥. Badger baiting was a real thing; this etymological link is not.

In the replies, people submitted their own made-up phrases and Google鈥檚 鈥渋nterpretations鈥. came up with 鈥淵ou can鈥檛 run a mile without hitting it with a hammer鈥, which is apparently 鈥渁 motivational phrase often used to emphasize the difficulty or struggle involved in achieving a goal鈥. Feedback was particularly delighted by the use of 鈥渙ften鈥 in that torrent of nonsense.

offered 鈥淚t鈥檚 better to have a tentacle in the tent than a rat in the rattan chair鈥. Google informed him that this is 鈥渁 humorous idiom that suggests it鈥檚 better to be in a situation that is initially uncomfortable or unusual than a situation that is undesirable and/or dangerous鈥. Feedback has a number of thoughts on this, not the least: why should a rogue tentacle be considered uncomfortable but not dangerous? We鈥檝e read H. P. Lovecraft: tentacles are a bad sign.

Alas, the 鈥渕eaning鈥 function seems to have been deactivated. We tried to persuade Google to give us a definition for 鈥渘ever rub a roe deer鈥檚 cabbages鈥, and it wouldn鈥檛 do it.

Of course, it鈥檚 mean to pick on the AI for doing what it was built to do: generating responses to questions. And it鈥檚 not like we haven鈥檛 met any humans that would rather spew nonsense than admit they don鈥檛 know the answer to a question.

But it perhaps highlights the issues with adding this technology to a page meant to be a source of accurate information. Feedback now no longer entirely trusts the results on Google, which ironically means the AI was right: you really can鈥檛 lick a badger twice.

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.