Guess which technology the US Federal Aviation Authority uses for backups? Yes, floppy disks. CC-licensed photo by Ivan Radic 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.
There’s another post coming this week at the Social Warming Substack on Friday at 0845 UK time. Free signup.
A selection of 9 links for you. Megabyte! I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.
US air traffic control still uses floppy disks for backup • The Register
Richard Speed:
»
The Federal Aviation Administration (FAA) has confirmed that the US air traffic control system still runs on somewhat antiquated bits of technology, including floppy disks and paper strips.
This came during last week’s Budget Hearing for the US House Appropriation Committee, in which the current FAA boss, Chris Rocheleau, explained to the committee that a new system would mean “no more floppy disks or paper strips.”
Asked by Congressman Mike Quigley how the FAA plans to make up for the “12% of aeronautical information specialists” – who update charts, maps and key data – that had either left the FAA or were planning to leave, Rocheleau said “first and foremost we’re assessing how we’re doing that and what what can we do better – so for instance going from a paperbased process to an electronic based process, that’s one of them.”
A few staffers should probably expect their job descriptions to enlarge, it seems. Rocheleau said the FAA would keep hiring for critical safety positions but would also be “leveraging the talent that we do have that is staying and making sure that they can both do the critical safety functions as well as those support functions.”
Asked by Kentucky representative Hal Rogers whether the FAA planned to “build a new system separate and apart from the present system” where it would simply switch the one system to “on and the other one to off,” Rocheleau described the transition as “a little more complicated than that” while committee chair Tom Cole quipped “They’ll be doing it while you’re in the air, Mr Rogers.”
The issue of outdated technology has troubled agencies for years. The recent outage at Newark Liberty International Airport – where a copper cable knocked out multiple systems including radar and comms and disrupted hundreds of flights – has thrown the problems faced by the FAA into sharp focus.
Rocheleau described a network systems refresh in which various units, including the ATC at Newark Liberty, would switch from the “copper wires” of the “old-fashioned telephone lines” over to fiber optic cables, as well as new modern systems for “radars and facilities,” promising “intentional deliberate testing to make sure the redundancy and the resiliency is there to ensure the safety of the traveling public.”
But this won’t happen any time soon.
«
Always lovely to know that systems on which thousands of lives depend are reliant on ancient technology.
unique link to this extract
Apple WWDC 2025: the 13 biggest announcements • The Verge
Emma Roth:
»
…Apple is revamping the design across its operating systems. Now they will have a new Liquid Glass theme that adds more transparency to buttons, switches, sliders, text, and media controls throughout their interfaces.
The date, time, notification previews on iOS 26’s lockscreen will also adopt the new look, allowing you to get a clear view of your wallpaper behind all the text.
…There’s a new Apple Intelligence feature coming to iOS 26 that lets you use the same buttons you would to take a screenshot to call upon the AI assistant. From there, you can ask additional questions about what you’re seeing on your screen with ChatGPT or search for a particular object on Google or Etsy to find similar images.
…Apple is building AI-powered live translation into the Messages, FaceTime, and Phone apps. You can use the integration to automatically translate texts in the Messages app, while you’ll hear speech translated aloud on the Phone app. In FaceTime, Apple will display translated live captions while you listen to your conversation partner.
…Though Apple’s keynote was a bit light on AI news, the company announced that it will let third-party app developers access the on-device large language model used by Apple Intelligence, allowing them to create tools of their own.
«
These all look straightforward enough, though you know that when people start installing the new look on their phones it’s going to be a Marmite bomb – love it or hate it. There’s new windowing coming to the iPad as well – once more – but we’ll see how that goes. Only available in developer betas at present.
unique link to this extract
Apple research questions AI reasoning models just days before WWDC • MacRumors
Tim Hardwick:
»
A newly published Apple Machine Learning Research study has challenged the prevailing narrative around AI “reasoning” large-language models like OpenAI’s o1 and Claude’s thinking variants, revealing fundamental limitations that suggest these systems aren’t truly reasoning at all.
For the study, rather than using standard math benchmarks that are prone to data contamination, Apple researchers designed controllable puzzle environments including Tower of Hanoi [moving rings between pegs] and River Crossing [moving missionaries and cannibals, occasionally]. This allowed a precise analysis of both the final answers and the internal reasoning traces across varying complexity levels, according to the researchers.
The results are striking, to say the least. All tested reasoning models – including o3-mini, DeepSeek-R1, and Claude 3.7 Sonnet – experienced complete accuracy collapse beyond certain complexity thresholds, and dropped to zero success rates despite having adequate computational resources. Counterintuitively, the models actually reduce their thinking effort as problems become more complex, suggesting fundamental scaling limitations rather than resource constraints.
Perhaps most damning, even when researchers provided complete solution algorithms, the models still failed at the same complexity points. Researchers say this indicates the limitation isn’t in problem-solving strategy, but in basic logical step execution.
Models also showed puzzling inconsistencies – succeeding on problems requiring 100+ moves while failing on simpler puzzles needing only 11 moves.
…The take-home of Apple’s findings is that current “reasoning” models rely on sophisticated pattern matching rather than genuine reasoning capabilities. It suggests that LLMs don’t scale reasoning like humans do, overthinking easy problems and thinking less for harder ones.
«
Sure, fine, LLMs don’t “reason” like we do. If we encountered aliens, would we decide they didn’t “reason” if they didn’t solve problems in the same way as us? Alternatively, would they think we didn’t reason if we couldn’t solve their problems consistently? (There’s a Substack making this point.)
It fits with Apple’s incoherent position on LLMs, though, and rumours that it thought they were rubbish from the start.
unique link to this extract
Cambridge mapping project solves a medieval murder • Ars Technica
Jennifer Ouellette:
»
In 2019, we told you about a new interactive digital “murder map” of London compiled by University of Cambridge criminologist Manuel Eisner. Drawing on data catalogued in the city coroners’ rolls, the map showed the approximate location of 142 homicide cases in late medieval London. The Medieval Murder Maps project has since expanded to include maps of York and Oxford homicides, as well as podcast episodes focusing on individual cases.
It’s easy to lose oneself down the rabbit hole of medieval murder for hours, filtering the killings by year, choice of weapon, and location. Think of it as a kind of 14th-century version of Clue: It was the noblewoman’s hired assassins armed with daggers in the streets of Cheapside near St. Paul’s Cathedral. And that’s just the juiciest of the various cases described in a new paper published in the journal Criminal Law Forum.
The noblewoman was Ela Fitzpayne, wife of a knight named Sir Robert Fitzpayne, lord of Stogursey. The victim was a priest and her erstwhile lover, John Forde, who was stabbed to death in the streets of Cheapside on May 3, 1337. “We are looking at a murder commissioned by a leading figure of the English aristocracy,” said University of Cambridge criminologist Manuel Eisner, who heads the Medieval Murder Maps project. “It is planned and cold-blooded, with a family member and close associates carrying it out, all of which suggests a revenge motive.”
Members of the mapping project geocoded all the cases after determining approximate locations for the crime scenes. Written in Latin, the coroners’ rolls are records of sudden or suspicious deaths as investigated by a jury of local men, called together by the coroner to establish facts and reach a verdict. Those records contain such relevant information as where the body was found and by whom; the nature of the wounds; the jury’s verdict on cause of death; the weapon used and how much it was worth; the time, location, and witness accounts; whether the perpetrator was arrested, escaped, or sought sanctuary; and any legal measures taken.
«
Chernobyl dogs are (not?) experiencing rapid evolution, biologists say • Popular Mechanics
Darren ORf:
»
we can cross radiation off the list of explanations for the current state of the Chernobyl canine population. Published in the journal PLOS One by scientists from North Carolina State University and Columbia University Mailman School of Public Health in New York, this new genetic analysis looked at the chromosomal level, the genome level, and even the nucleotides of the Chernobyl dogs, and found no abnormalities indicative of radiation-induced mutation.
To establish a baseline for comparison, the team compared the genome of [feral] Chernobyl City dogs located 10 miles from the Chernobyl Exclusion Zone (CEZ) to dogs found in regions of Russia, Poland, and other nearby countries. Once they determined that the populations were genetically similar, they then used the Chernobyl City dogs as a representative control for their study. Of course, the task wasn’t simple, as more than a couple dozen dog generations have past since the original pups that witnessed to the Chernobyl meltdown in 1986.
“We know that, for example, exposure to high doses of radiation can introduce instability from the chromosomal level on down,” Matthew Breen, senior author of the study from North Carolina State University, said in a press statement. “While this dog population is 30 or more generations removed from the one present during the 1986 disaster, mutations would likely still be detectable if they conferred a survival advantage to those original dogs. But we didn’t find any such evidence in these dogs.”
That said, the 2023 study [which suggested there were changes] still provides a template for further investigation into the effects of radiation on larger mammals, as the DNA of dogs roaming the Chernobyl Power Plant and nearby Chernobyl City can be compared to dogs living in non-irradiated areas. Despite a current lack of firm conclusions, the study has shown once again that an area that—by all rights—should be a wasteland has become an unparalleled scientific opportunity to understand radiation and its impact on natural evolution.
«
There have been various studies around Chernobyl – wolves, boar – and it’s a bit unclear what’s going on, overall.
unique link to this extract
Nigerian engineer sets Guinness World Record with world’s smallest GPS tracking device • Technology Times
Iretomiwa Balogun:
»
Nigerian engineer Oluwatobi “Tobi” Oyinlola, based at the Massachusetts Institute of Technology (MIT), is earning global recognition after developing a GPS tracking device smaller than a human thumbprint, a feat that sets a new Guinness World Record.
The device, measuring just 22.93 mm by 11.92 mm, is officially certified by Guinness World Records as the smallest GPS tracking device prototype in the world. Designed and built in Cambridge, Massachusetts, it integrates a custom-printed circuit board, an embedded microcontroller, and Bluetooth capabilities, enabling it to receive GPS signals, log location data, and transmit it via Bluetooth – all without the need for an external antenna.
“This project started as a personal challenge,” Oyinlola says. “My goal was to create a compact GPS device that integrates everything into one unit – microcontroller, Bluetooth, and antenna – without external components.”
According to the MIT-based researcher, the innovation holds potential for applications across wearables, medical microchips, wildlife monitoring, and personal safety devices, offering a pathway to new use cases where minimal form factor and high functionality are critical.
«
Impressive. Though the best bet is that it’s not going to be used for “personal safety” so much as “tracking suspicious people”.
unique link to this extract
Rat breaches bank ATM in India, eats $18,000 worth of cash • Reuters
Zarir Hussain:
»
When bank technicians in India were finally summoned to investigate why an ATM (automated teller machine) had not been working for days, they began to smell a rat.
What they found inside the ATM was almost $18,000 worth of shredded Indian rupee notes and one dead rodent that had somehow eluded the machine’s security camera for its next, and last, meal, a State Bank of India (SBI) official said on Thursday.
“The ATM was out of order for a few days and when our technicians opened the kiosk we were shocked to find shredded notes and a dead rat,” said Chandan Sharma, SBI branch manager in the town of Tinsukia in the northeastern state of Assam.
“We have started an investigation into this rare incident and will take measures to prevent a recurrence.”
SBI is India’s largest bank with more than 50,000 ATMs spread across the country. Most ATMs in India have a closed-circuit camera installed for enhanced security.But an inspection of the camera footage at the ATM in Tinsukia turned up no rat entering it, Sharma said. Of the 2.9 million rupees ($42,685) in the ATM, 1.7 million rupees ($25,022) were recovered intact. But banknotes worth 1.2 million rupees ($17,662) were destroyed.
«
Not much fun for the rat either, though.
unique link to this extract
Millions of low-cost Android devices turn home networks into crime platforms • Ars Technica
Dan Goodin:
»
Millions of low-cost devices for media streaming, in-vehicle entertainment, and video projection are infected with malware that turns consumer networks into platforms for distributing malware, concealing nefarious communications, and performing other illicit activities, the FBI has warned.
The malware infecting these devices, known as BadBox, is based on Triada, a malware strain discovered in 2016 by Kaspersky Lab, which called it “one of the most advanced mobile Trojans” the security firm’s analysts had ever encountered. It employed an impressive kit of tools, including rooting exploits that bypassed security protections built into Android and functions for modifying the Android OS’s all-powerful Zygote process. Google eventually updated Android to block the methods Triada used to infect devices.
…In 2023, security firm Human Security reported on BigBox, a Triada-derived backdoor it found preinstalled on thousands of devices manufactured in China. The malware, which Human Security estimated was installed on 74,000 devices around the world, facilitated a range of illicit activities, including advertising fraud, residential proxy services, the creation of fake Gmail and WhatsApp accounts, and infecting other Internet-connected devices.
In March, Google and a consortium of other Internet organizations took part in a coordinated action to disrupt BadBox 2.0, a new campaign affecting more than 1 million low-priced, off-brand Android devices. The infected devices were based on the Android Open Source Project, not the Android TV OS. They also weren’t certified under Google’s Play Protect security program. Human Security identified more than a dozen TV models that were impacted. It was the second BadBox disruption action in as many years.
On Thursday, the FBI warned that the BadBox threat remained and urged consumers to look for signs their devices may be infected.
«
Problem: there are few visible signs of infection. Simpler move: check for a model known to be bad, replace it. Thanks China! And Android!
unique link to this extract
Bluesky’s decline stems from never hearing from other side – The Washington Post
Megan McArdle:
»
It’s not surprising that progressives want to return to the good old days. But it’s not working, and I’m skeptical it ever will.
The people who have migrated to Bluesky tend to be those who feel the most visceral disgust for Musk and Trump, plus a smattering of those who are merely curious and another smattering who are tired of the AI slop and unregenerate racism that increasingly pollutes their X feeds. Because the Musk and Trump haters are the largest and most passionate group, the result is something of an echo chamber where it’s hard to get positive engagement unless you’re saying things progressives want to hear — and where the negative engagement on things they don’t want to hear can be intense.
That’s true even for content that isn’t obviously political: Ethan Mollick, a professor at the University of Pennsylvania’s Wharton School who studies AI, recently announced that he’ll be limiting his Bluesky posting because AI discussions on the platform are too “fraught.”
All this is pretty off-putting for folks who aren’t already rather progressive, and that creates a threefold problem for the ones who dream of getting the old band back together. Most obviously, it makes it hard for the platform to build a large enough userbase for the company to become financially self-sustaining, or for liberals to amass the influence they wielded on old Twitter. There, they accumulated power by shaping the contours of a conversation that included a lot of non-progressives. On Bluesky, they’re mostly talking among themselves.
One can say the same about Truth Social, of course, but that’s not an example the left should be eager to emulate. Segregating yourself in a political silo amplifies any political movement’s worst tendencies, giving free rein to your most toxic adherents and cutting you off from vital feedback about, say, your unpopular tariff policies.
«
The numbers for Bluesky keep going in the wrong direction. Accounts: 36.5 million. Daily posters: 1.5m on November 19, 0.67m now. That’s a fall from 4% to less than 2%. That page also has another relevant statistic showing the power law of social networks: 95% of users have 85 or fewer followers. It’s only when you get to the top 0.3% – about 100,000 people – that people have more than 1,000 followers. And only 3,600 have more than 11,000 followers.
That’s when you need a really good algorithm to show people new and interesting content from across the network. Otherwise, it dies.
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 