In Denmark, falling letter volumes mean deliveries will cease this year. How long might the UK’s letter service have left? CC-licensed photo by Sergey 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. Not going postal. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.
Greenhouse gases reduce the satellite carrying capacity of low Earth orbit • Nature Sustainability
William Parker, Matthew Brown and Richard Linares:
»
Anthropogenic contributions of greenhouse gases in Earth’s atmosphere have been observed to cause cooling and contraction in the thermosphere, which is projected to continue for many decades. This contraction results in a secular reduction in atmospheric mass density where most satellites operate in low Earth orbit.
Decreasing density reduces drag on debris objects and extends their lifetime in orbit, posing a persistent collision hazard to other satellites and risking the cascading generation of more debris. This work uses projected CO2 emissions from the shared socio-economic pathways to investigate the impact of greenhouse gas emissions on the satellite carrying capacity of low Earth orbit. The instantaneous Kessler capacity is introduced to compute the maximum number and optimal distribution of characteristic satellites that keep debris populations in stable equilibrium.
Modelled CO2 emissions scenarios from years 2000–2100 indicate a potential 50–66% reduction in satellite carrying capacity between the altitudes of 200 and 1,000 km.
«
Starlink orbits in low Earth orbit, at about 500km up. Contradictory result: warming of the planet goes together with cooling of the upper atmosphere. And fewer satellites, of course.
unique link to this extract
Volkswagen reintroducing physical controls for vital functions • Autocar
James Attwood:
»
All future Volkswagen models will feature physical controls for the most important functions, design chief Andreas Mindt has said.
The German firm has been criticised over the past few years for moving many of the vital controls in its cars from physical buttons and dials to the infotainment touchscreen. Volkswagen also introduced haptic ‘sliders’ below the touchscreen for the heating and volume and it started using haptic panels instead of buttons for controls mounted on the steering wheel.
More recently, the firm has reintroduced physical steering wheel buttons and Mindt said it is committed to reintroducing physical buttons, starting with the production version of the ID 2all concept that will arrive next year.
“From the ID 2all onwards, we will have physical buttons for the five most important functions – the volume, the heating on each side of the car, the fans and the hazard light – below the screen,” said Mindt. “They will be in every car that we make from now on. We understood this.
“We will never, ever make this mistake any more. On the steering wheel, we will have physical buttons. No guessing any more. There’s feedback, it’s real, and people love this. Honestly, it’s a car. It’s not a phone: it’s a car.”
Mindt said VW will continue to offer cars with touchscreens, in part due to new legal requirements that, as in the US, will require all cars to feature a reversing camera.
«
The obvious question – which someone who first linked to this pointed out – is, why didn’t you notice these problems before? How long did it take driving the car to recognise that this wasn’t ideal? Or were touchscreens just so much cheaper and simpler you closed your eyes to the disbenefits for the driver?
unique link to this extract
How plants are responding to a warming world — and why it matters • The MIT Press Reader
Theresa Crimmins:
»
in India, the flowering period for saffron has been substantially shortened in recent years. Saffron “threads,” highly valued for the flavor and rich color they bring to dishes and drinks, are actually the female reproductive parts of the saffron crocus flower. In parts of India, saffron flowers now open when temperatures are too warm for their development. This leads to a high rate of flower death, and with no flowers, there is no saffron. Between 2013 and 2017, saffron production in Kashmir declined by 90 percent. Consequently, many saffron farmers are shifting their plantings to higher elevations with cooler temperatures.
The northeastern United States is a major fruit production region, as are southern states. As temperatures in these regions have warmed, leaf out and flower bud development as well as the last spring frost have shifted earlier in the season. In many locations, however, the date of the last frost has not shifted earlier to the same degree as plant activity. Consequently, tender plant tissues are at greater risk of exposure to damaging frosts. Many of the plants that produce fruits we enjoy, including blueberries, apples, and cherries, open their flower buds early in the season, sometimes even before they break leaf buds. Once flower buds begin to open, they become sensitive to cold temperatures.
As with saffron, if flower buds are killed by frost, there are no fruits. So advancing phenology is expected to worsen the risk of frost damage in the coming decades. The start of springtime biological activity in the United States is projected to advance by up to three weeks by the end of the century. One set of predictions indicates that we can expect to experience early warm springs followed by damaging freeze events in nearly one out of every three years by the mid-21st century. The same is predicted for Europe and Asia, with up to a third of Europe and Asia’s forests predicted to be threatened by frost damage in future decades.
«
Crimmins has a book out on “phenology” – the timing of seasonal plant activity.
unique link to this extract
Why extracting data from PDFs is still a nightmare for data experts • Ars Technica
Benj Edwards:
»
The PDF challenge also represents a significant bottleneck in the world of data analysis and machine learning at large. According to several studies, approximately 80–90% of the world’s organizational data is stored as unstructured data in documents, much of it locked away in formats that resist easy extraction. The problem worsens with two-column layouts, tables, charts, and scanned documents with poor image quality.
The inability to reliably extract data from PDFs affects numerous sectors but hits hardest in areas that rely heavily on documentation and legacy records, including digitizing scientific research, preserving historical documents, streamlining customer service, and making technical literature more accessible to AI systems.
“It is a very real problem for almost anything published more than 20 years ago and in particular for government records,” Willis says. “That impacts not just the operation of public agencies like the courts, police, and social services but also journalists, who rely on those records for stories. It also forces some industries that depend on information, like insurance and banking, to invest time and resources in converting PDFs into data.”
…Unlike traditional OCR methods that follow a rigid sequence of identifying characters based on pixel patterns, multimodal LLMs that can read documents are trained on text and images that have been translated into chunks of data called tokens and fed into large neural networks. Vision-capable LLMs from companies like OpenAI, Google, and Meta analyze documents by recognizing relationships between visual elements and understanding contextual cues.
The “visual” image-based method is how ChatGPT reads a PDF file, for example, if you upload it through the AI assistant interface. It’s a fundamentally different approach than standard OCR that allows them to potentially process documents more holistically, considering both visual layouts and text content simultaneously.
And as it turns out, some LLMs from certain vendors are better at this task than others.
“The LLMs that do well on these tasks tend to behave in ways that are more consistent with how I would do it manually,” Willis said. He noted that some traditional OCR methods are quite good, particularly Amazon’s Textract, but that “they also are bound by the rules of their software and limitations on how much text they can refer to when attempting to recognize an unusual pattern.” Willis added, “With LLMs, I think you trade that for an expanded context that seems to help them make better predictions about whether a digit is a three or an eight, for example.”
This context-based approach enables these models to better handle complex layouts, interpret tables, and distinguish between document elements like headers, captions, and body text—all tasks that traditional OCR solutions struggle with.
«
What really happened with the DDoS attacks that took down X • WIRED
Lily Hay Newman and Zoe Schiffer:
»
Multiple researchers tell WIRED that they observed five distinct attacks of varying length against X’s infrastructure, the first beginning early Monday morning with the final burst on Monday afternoon.
The internet intelligence team at Cisco’s ThousandEyes tells WIRED in a statement, “During the disruptions, ThousandEyes observed network conditions that are characteristic of a DDoS attack, including significant traffic loss conditions which would have hindered users from reaching the application.”
DDoS attacks are common, and virtually all modern internet services experience them regularly and must proactively defend themselves. As Musk himself put it on Monday, “We get attacked every day.” Why, then, did these DDoS attacks cause outages for X? Musk said it was because “this was done with a lot of resources,” but independent security researcher Kevin Beaumont and other analysts see evidence that some X origin servers, which respond to web requests, weren’t properly secured behind the company’s Cloudflare DDoS protection and were publicly visible. As a result, attackers could target them directly. X has since secured the servers.
“The botnet was directly attacking the IP and a bunch more on that X subnet yesterday. It’s a botnet of cameras and DVRs,” Beaumont says.
A few hours after the final attack concluded, Musk told Fox Business host Larry Kudlow in an interview, “We’re not sure exactly what happened, but there was a massive cyberattack to try to bring down the X system with IP addresses originating in the Ukraine area.”
«
As usual, Musk is talking complete and utter bollocks. And in this case he certainly knows it. Ukraine’s hacker groups have far better things to do with their time.
unique link to this extract
North Korean Lazarus hackers infect hundreds via npm packages • Bleeping Computer
Bill Toulas:
»
Six malicious packages have been identified on npm (Node package manager) linked to the notorious North Korean hacking group Lazarus.
The packages, which have been downloaded 330 times, are designed to steal account credentials, deploy backdoors on compromised systems, and extract sensitive cryptocurrency information.
The Socket Research Team discovered the campaign, which linked it to previously known Lazarus supply chain operations. The threat group is known for pushing malicious packages into software registries like npm, which is used by millions of JavaScript developers, and compromising systems passively.
Similar campaigns attributed to the same threat actors have been spotted on GitHub and the Python Package Index (PyPI). This tactic often allows them to gain initial access to valuable networks and conduct massive record-breaking attacks, like the recent $1.5bn crypto heist from the Bybit exchange.
The six Lazarus packages discovered in npm all employ typosquatting tactics to trick developers into accidental installations.
The packages contain malicious code designed to steal sensitive information, such as cryptocurrency wallets and browser data that contains stored passwords, cookies, and browsing history. They also load the BeaverTail malware and the InvisibleFerret backdoor, which North Koreans previously deployed in fake job offers that led to the installation of malware.
…Software developers are advised to double-check the packages they use for their projects and constantly scrutinize code in open-source software to find suspicious signs like obfuscated code and calls to external servers.
«
Good idea! Perhaps though we could get machines that could do this? We could call it something like “intelligence”. Stumped for a good name though.
unique link to this extract
TikTok’s owner hasn’t negotiated with prospective buyers as deadline looms • Axios
Dan Primack:
»
TikTok has less than one month until its U.S. lifeline expires on April 5, but sources say there still haven’t been negotiations between its Chinese owner and prospective buyers.
Suitors are increasingly frustrated by their inability to get under TikTok’s hood, in terms of both finances and technology. There also has been some confusion as to who is actually in charge of Trump administration negotiations for the deal. Early expectations were that it would be Treasury Secretary Scott Bessent, given his role as chairman of CFIUS, but Vice President Vance seems to have taken the baton.
Beijing, meanwhile, continues to be publicly silent; one source suggested the situation is like juggling four balls when one is invisible.
Everyone has ideas of how a deal could be structured, but no confidence on how it needs to be structured.
The April 5 deadline isn’t really set in stone. President Trump had no legal authority to extend the U.S. ban by 75 days, so there’s no reason to think he wouldn’t extend it out further (or even indefinitely). [ie “He already broke the law, so why not break it further?” – Overspill Ed.]
But his hand would be helped if at least a term sheet were submitted, since that could trigger a legal extension under the law passed last year. There’s also a proposal from Sen. Markey (D-Mass.) to push back the deadine by another 270 days, although it’s stuck in legislative limbo.
«
JD Vance, the meme-Veep, is also said to be involved in the discussions. To no effect, apparently.
unique link to this extract
Denmark’s postal service to stop delivering letters – BBC News
Adrienne Murray and Paul Kirby:
»
Denmark’s state-run postal service, PostNord, is to end all letter deliveries at the end of 2025, citing a 90% decline in letter volumes since the start of the century.
The decision brings to an end 400 years of the company’s letter service. Denmark’s 1,500 post boxes will start to disappear from the start of June.
Transport Minister Thomas Danielsen sought to reassure Danes, saying letters would still be sent and received as “there is a free market for both letters and parcels”.
Postal services across Europe are grappling with the decline in letter volumes. Germany’s Deutsche Post said on Thursday it was axing 8,000 jobs, in what it called a “socially responsible manner”.
Deutsche Post has 187,000 employees and staff representatives said they feared more cuts were to come.
…The decision will affect elderly people most. Although 95% of Danes use the Digital Post service, a reported 271,000 people still rely on physical mail.
«
Denmark has a population of about 6 million; letter volumes have fallen from just over 1.4bn in 2000 to less than 200m last year. That’s an average of over 230 per person in 2000 to 33 last year.
In the UK, 2024 volumes were 6.6 billion for about 69 million people, or 95 per person per year. The peak was 2004-5, with 20.1bn letters – a bit less than 300 per person.
The trend is the same, and now Royal Mail is in private hands, when will the tipping point come?
unique link to this extract
DOJ: Google must sell Chrome, Android could be next • Ars Technica
Ryan Whitwam:
»
Google has gotten its first taste of remedies that Donald Trump’s Department of Justice plans to pursue to break up the tech giant’s monopoly in search. In the first filing since Trump allies took over the department, government lawyers backed off a key proposal submitted by the Biden DOJ. The government won’t ask the court to force Google to sell off its AI investments, and the way it intends to handle Android is changing. However, the most serious penalty is intact—Google’s popular Chrome browser is still on the chopping block.
“Google’s illegal conduct has created an economic goliath, one that wreaks havoc over the marketplace to ensure that—no matter what occurs—Google always wins,” the DOJ filing says. To that end, the government maintains that Chrome must go if the playing field is to be made level again.
The DOJ is asking the court to force Google to promptly and fully divest itself of Chrome, along with any data or other assets required for its continued operation. It is essentially aiming to take the Chrome user base—consisting of some 3.4 billion people—away from Google and hand it to a competitor. The government will vet any potential buyers to ensure the sale does not pose a national security threat. During the term of the judgment, Google would not be allowed to release any new browsers. However, it may continue to contribute to the open source Chromium project.
This filing includes some changes from the initial remedy filings of 2024, but more changes could be coming. The case is currently under the purview of Omeed Assefi, who is leading the DOJ’s Antitrust Division until Trump nominee Gail Slater gets Senate confirmation. Slater expressed support for increased scrutiny of Big Tech in her confirmation hearings, suggesting she could seek to turn the screws on Google after taking charge.
…There are two notable changes in the government’s position regarding Android and AI investment. Neither area will completely escape the government’s grasp under the proposed remedies, but Google won’t have to sell anything immediately.
«
Still awaiting the new Trump-appointed chief, and the decision of the judge in the case.
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 
Who would anyone buy Chrome?
The browser itself is largely open source. You can build your “own” easily. The only reason for anyone to buy it is to use it for pushing your own services.
Microsoft and Amazon have such services (Bing, Amazon store, Prime etc.), but they will not be allowed to buy it and use it for pushing their services. Of course the browser could collect all sorts of usage data, but they get plenty of that already.
So who’s it going to be, why and who sets the price? What if nobody wants to buy?
Maybe they’ll give Chrome to Musk.
Related, EU forced Apple to allow native Chrome on the iPhone and iPad. This is fantastic – for Google. Their tech’s market share increases.
(Google hasn’t bothered to port Chrome to BrowserEngineKit yet though)