Folding towels is a skill that humans find easy, but robots need to learn it. Guess who gets to teach them? CC-licensed photo by arbyreed on Flickr.
You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.
A selection of 9 links for you. Turned out. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.
Inside the race to train AI robots how to act human in the real world • Los Angeles Times
Nilesh Christopher:
»
Now that artificial intelligence has mastered almost everything we do online, it needs help learning how we physically move around in the real world.
A growing global army of trainers is helping it escape our computers and enter our living rooms, offices and factories by teaching it how we move.
In an industrial town in southern India, Naveen Kumar, 28, stands at his desk and starts his job for the day: folding hand towels hundreds of times, as precisely as possible.
He doesn’t work at a hotel; he works for a startup that creates physical data used to train AI. He mounts a GoPro camera to his forehead and follows a regimented list of hand movements to capture exact point-of-view footage of how a human folds.
That day, he had to pick up each towel from a basket on the right side of his desk, using only his right hand, shake the towel straight using both hands, then fold it neatly three times. Then he had to put each folded towel in the left corner of the desk. If it takes more than a minute or he misses any steps, he has to start over.
His firm, a data labeling company called Objectways, sent 200 towel-folding videos to its client in the United States. The company has more than 2,000 employees; about half of them label sensor data from autonomous cars and robotics, and the rest work on generative AI. Most of them are engineers, and few are well-practiced in folding towels, so they take turns doing the physical labor.
“Sometimes we have to delete nearly 150 or 200 videos because of silly errors in how we’re folding or placing items,” said Kumar, an engineering graduate who has worked at Objectways for six years.
The carefully choreographed movements are to capture all the nuances of what humans do — arm reaching, fingers gripping, fabric sliding — to fold clothes. The captured videos are then annotated by Kumar and his team. They draw boxes around the different parts of the video, tag the towels, and label whether the arm moved left or right, and classify each gesture.
«
A recent episode of ATP discussed how computer programming is about spotting pitfalls, using the example of the schoolteacher who asks the class to instruct them, step by step but assuming no intuition, to make a peanut butter and jelly (jam, yes? Ugh) sandwich. It’s much harder than it seems. And so with folding towels, which we do almost reflexively once we’ve done it a couple of times.
Robots have so much to learn about the world. (Thanks Gregory B for the link.)
unique link to this extract
“I Deliver Parcels in Beijing” book author Hu Anyan on Chinese e-commerce, AI, and American readers • Rest of World
Viola Zhou:
»
Hu Anyan has held 19 jobs in six cities across China — selling bicycles, running a clothing store, working in a bakery, making 3D architectural renderings, doing night shifts at a logistics warehouse, and eventually delivering packages.
Hu, 46, wrote about these experiences in a memoir-style book, I Deliver Parcels in Beijing. An avid reader, Hu documents his encounters with abusive managers, irate customers, and sprawling residential complexes in casual language, and with colorful details and a touch of humor.
When published in 2023, Hu’s book became a bestseller in China. Readers lapped up anecdotes of the lives of some of the millions of couriers powering the country’s ultra-efficient e-commerce industry, which treats individual laborers as dispensable. Many also related to Hu’s experience with economic uncertainty, dwindling social mobility, unemployment, and unfulfilling work.
Ahead of the launch of the English-language translation of the book by Jack Hargreaves, Hu spoke to Rest of World about his literary journey, his views on whether couriers will be replaced by automation, and what he hopes Americans will learn from the book.
«
It’s a fun interview, if not particularly long.
unique link to this extract
Apple prepares to sell low-cost laptops for first time • Bloomberg via Financial Post
Mark Gurman:
»
Apple Inc. is preparing to enter the low-cost laptop market for the first time, developing a budget Mac aimed at luring away customers from Chromebooks and entry-level Windows PCs.
The new device — designed for students, businesses and casual users — will target people who primarily browse the web, work on documents or conduct light media editing, according to people familiar with the matter. Apple is also targeting would-be iPad buyers who might prefer a traditional laptop experience instead.
Code-named J700, the machine is currently in active testing at Apple and in early production with overseas suppliers. The Cupertino, California-based company plans to launch it in the first half of next year, said the people, who asked not to be identified because the product hasn’t been announced.
An Apple spokesperson declined to comment.
The move would represent a strategic shift for Apple, which has historically focused on premium devices with hefty profit margins. The company also has vowed not to chase market share with lower-end offerings.
But Apple is facing a growing threat from Chromebooks, the low-cost laptops that run Google’s operating system, Chrome OS. There’s also a potential opportunity to entice Windows customers. Microsoft Corp.’s shift to Windows 11 has rankled some users of the previous-generation software and left them without security updates.
Shares of personal computer maker HP Inc. briefly dipped to a session low on the news. The stock and that of fellow PC maker Dell Technologies Inc. were both down about two% as of 12:03 p.m. in New York. Apple gained less than one% to US$270.25.
Apple plans to sell the new machine for well under US$1,000 by using less-advanced components. The laptop will rely on an iPhone processor and a lower-end LCD display. The screen will also be the smallest of any current Mac, coming in at slightly below the 13.6in one used in the MacBook Air.
«
If Apple is looking to take significant share from Chromebooks, it’s years late – Google has taken huge chunks of the education market by not only having a simple user login but also the cloud system behind it. There’s always a market for cheaper laptops, of course.
unique link to this extract
A YouTube education • Cultural Capital
James Marriott:
»
YouTube is now second only to the BBC as the most popular broadcaster in the UK.
I often write disparagingly about the modern internet’s inexorable slide towards video but I also have to (grudgingly) concede that I have learned an awful lot on YouTube over the years.
In my teens it helped introduce me to poetry, philosophy music and art. I thought I would make a list of videos which form a kind of curriculum in the humanities (and a tiny bit in the sciences).
This list is partial and biased towards my own interests. I’m sure I’ve missed many things (I’m not into animal documentaries but there must be one of those good David Attenborough-style things with baboons etc somewhere) so please send me your favourite educational YouTube videos in the comments.
Here is a YouTube education. Or my version of one.
«
Marriott goes on to list multiple series which will educate and inform you – Kenneth Clark’s Civilisation, Ways of Seeing by John Berger, Dawkins and Pinker on evolution, language and the brain, and so on.
Except.. you can find those on the BBC iPlayer. Which, OK, isn’t available outside the UK, but Marriott’s article seems to imply that this is “YouTube as a replacement for what the BBC doesn’t offer”. The BBC does! Or, alternatively, it’s for people who don’t have TV licences. Which then raises the question of how long that makes sense: if YouTube is effectively taking revenue from the BBC (those lost licence fees; I doubt the income from people watching Civilisation makes up for it).
unique link to this extract
Chicago firm that resolves ransomware attacks had rogue workers carrying out their own hacks, FBI says • Chicago Sun-Times
Tom Schuba:
»
Rogue employees of a Chicago company that specializes in negotiating ransoms to mitigate cyber attacks were carrying out their own piracy in a plot to extort millions of dollars from a series of companies, prosecutors say.
Kevin Tyler Martin, a ransomware threat negotiator for River North-based DigitalMint at the time of the alleged conspiracy, was among two men indicted in the scheme. A suspected accomplice who wasn’t indicted was also employed at DigitalMint, court records show.
DigitalMint has denied any wrongdoing, fired both employees and cooperated with the investigation.
Also indicted was Ryan Clifford Goldberg, an incident response manager for the multinational company Sygnia Cybersecurity Services. Sygnia said Goldberg no longer works for the company and it “is not the target of this investigation, however we continue to work closely with law enforcement.”
According to an affidavit filed in September by an FBI agent, the three men began using malicious software in May 2023 “to conduct ransomware attacks against victims,” first hitting a medical company in Florida by locking its servers and demanding $10m to unlock the systems, court records say .
The FBI agent noted the men ultimately made off with $1.2m, although it was apparently the only successful attack.
«
Struggling to find a description here. Playing both ends against the middle? Man-in-the-middle attack? But it’s an amazing piece of exploitation.
unique link to this extract
Inside NORAD’s Cheyenne Mountain Combat Center, c.1966 • Flashbak
Paul Sorene:
»
In 1966, the North American Air Defense Command (NORAD) Cheyenne Mountain underground facility became America’s command and control centre for when the Cold War turned white and the nuclear apocalypse became real.
We can take a look around the bunker in pictures published in a 1966 official NORAD report by David W. Shircliffe, Directorate of Command History, Command Public Affairs Office, and the 1970s book NORAD Command Post: The City Inside Cheyenne Mountain by Henry W. Hough.
«
Dr Strangelove is just out of picture in the one with Nixon. I observe that seniority and rank are still expressed through the quality of chairs at monitoring desks.
unique link to this extract
Bending Spoons cofounders become billionaires after Italian startup raises at $11bn valuation • Forbes
Iain Martin:
»
Luca Ferrari bought his first app back in 2014 for just $10,000 with the hope that he and his three cofounders in Milan-based startup Bending Spoons could turn it around. A decade later, Ferrari has become one of tech’s biggest dealmakers, and a new billionaire, after a new funding round valued his startup at more than $11bn.
Forbes estimates that Ferrari’s stake in the startup, which is named after a scene from the Matrix movie, is worth $1.4bn, while his cofounders Matteo Danieli, Luca Querella and Francesco Patarnello each hold a $1.3bn stake based on shareholder data published by the Italian Business Register.
The valuation comes from Bending Spoon’s fresh funding round of $270m from investors including T. Rowe Price and existing investors Baillie Gifford, Cox Enterprises, Durable Capital Partners and Fidelity. There was also a $440m secondary share sale by existing shareholders in the company. It’s unclear whether any of the cofounders sold stock in the secondary transaction.
…“Our strategy is very clear and sharp but it doesn’t include any bias to any particular segment,’ says Ferrari.
That strategy involves Bending Spoons using debt to snap up apps, products and websites with healthy revenue but where growth has often stalled. In some deals, like Evernote, it seems to have followed a private equity playbook of cutting staff, and hiking prices, but Bending Spoons says it has also invested heavily to overhaul and expand other apps it has acquired, like Meetup.
Bending Spoons, which was largely self-funded until 2021, has drawn comparison to a buyout fund, and Canadian software consolidator Constellation ($52.2bn market cap), but neither label fits, said early investor Peter Singlehurst of British fund manager Baillie Gifford. “They own and operate digital applications and are great at growing them very profitably because of the level of talent in the organization,” Singlehurst told Forbes.
«
1) Bending spoons… so nothing to do with Uri Geller?
2) I still don’t get what BSpoons’s business model is to have such a gigantic valuation. OK, so they own AOL now, but they got that in a fire sale because, well, there was essentially no further value in it to a giant media company. But these guys will make it into a rocket?
unique link to this extract
Metaphors for biology: sizes • Asimov Magazine
Sam Clamons:
»
Biology can be hard to intuit, in part because it operates across vastly different scales, from single atoms all the way up to entire ecosystems. Students of biology therefore often first meet its agents and mechanisms through metaphors: molecules are charged balls connected by sticks! evolution designs organisms to maximize their fitness! mitochondria are the powerhouses of the cell! While metaphors give us qualitative handles to grasp, they often oversimplify complex ideas.
This is because most metaphors fail to address specifics — especially regarding numbers. Consider another common bio-metaphor: DNA is the blueprint of the cell. That’s useful for conceptual understanding, but how big is this blueprint? Is it as big as a novel or an encyclopedia? How much space does it take up? It’s possible to look up or calculate the answers to these questions; the human genome is 6.2 billion base pairs, which takes up about 10 cubic microns. But how big is that compared to the total volume of a cell? Is it most of it or just a tiny fraction?
To answer questions like these, you need more quantitative metaphors. Whereas a standard metaphor says that mitochondria are the powerhouses of the cell, a quantitative metaphor says how big a battery a mitochondrion would be, say, if a cell were a toaster. If qualitative metaphors are like containers, then quantitative metaphors are closer to yardsticks.
And these yardsticks are exceptionally useful down at the scales where biology operates. We all know cells are small. But so are proteins, nucleic acids, and water molecules. We often think of everything “small” as equally small, but that is not the case. Proteins are titanic, hulking machines compared to water molecules, and the mRNA that encodes a protein is orders of magnitude larger still!
To make the sizes and shapes of various biomolecules concrete, let’s imagine that each water molecule within a cell has been blown up to the size of a grain of sand. If this were the case, then…
«
The scale is still hard to grasp even with this sort of scale offered.
unique link to this extract
1979: Alpha particule-induced soft errors in dynamic memories
Timothy May and Murray Woods in 1979:
»
A new physical soft error mechanism in dynamic RAM and CCDs is the upset of stored data by the passage of alpha particles through the memory array area.
The alpha particles are emitted by the radioactive decay of uranium and thorium which are present in parts- per-rnillion levels in packaging materials.
When an alpha particle penetrates the die surface, it can create enough electron-hole pairs near a storage node to cause a random, single-bit error. Results of experiments and measurements of alpha activity of materials are reported and a physical model for the soft error is developed. Implications for the future of dynamic memories are also discussed.
«
This is how Intel and the rest of the semiconductor industry discovered “bit flipping”: they were using water from wells which were themselves polluted with uranium and thorium (sometimes from nuclear testing) to make the packaging for the chips. This paper figured out how much energy the alpha particles needed to flip bits, and so how much refining of their sources was needed. (Turned out it was 100-1000 times better than they had going.)
unique link to this extract
| • Why do social networks drive us a little mad? • Why does angry content seem to dominate what we see? • How much of a role do algorithms play in affecting what we see and do online? • What can we do about it? • Did Facebook have any inkling of what was coming in Myanmar in 2016? Read Social Warming, my latest book, and find answers – and more. |
Errata, corrigenda and ai no corrida: none notified
The Overspill: when there's more that I want to say 