In an experiment, Google has quantified how sending drivers headed for the same destination by different routes can reduce congestion and emissions. CC-licensed photo by Shiyang Huang 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. Rerouting. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.
Pro-Palantir data has litany of errors, English hospitals admit • Financial Times
Amy Borrett and Laura Hughes:
»
Hospital trusts have admitted there are a litany of errors in the data behind one of NHS England’s headline claims about the benefits of Palantir’s technology after an FT investigation identified irregularities.
NHS England has claimed that hospitals have seen a 15% decline in delays to discharging hospital patients after adopting a tool built using the US company’s software.
However, four trusts confirmed there are errors in the underlying hospital data after an analysis by the FT found they had reported a sudden drop in discharge delays from hundreds or thousands to zero before increasing again sharply — a scenario that would be highly improbable.
The errors cast doubt on NHSE’s headline claim about the overall fall in discharge delays, a figure widely cited as evidence of the success of Palantir’s Federated Data Platform, which aims to bring together disparate data in order to boost efficiency in the health service.
Palantir’s role has become an increasing source of controversy, given its ties to the US defence sector.
The evidence for the impact of the FDP on discharge delays looked “increasingly flimsy”, said Charles Tallack, former head of operational research and evaluation at NHS England. “The delayed discharge dataset may be suitable for day-to-day management purposes, but not for evaluation,” he said.
The benefits data is calculated using NHSE’s national, mandatory dataset, which tracks performance on discharge delays across all trusts and of which Palantir has no oversight. The NHS England website itself acknowledges the limitations of the data, noting there are only “minimal” quality checks.
Across the full four years of acute discharge data, the FT identified irregularities for 42% of all NHS hospital trusts.
«
That feels like a big number? And anyone who has bounced off the NHS won’t be surprised that there are data errors. It seems to be an inescapable attribute of the system, which no amount of effort can remove.
unique link to this extract
The newest way to go analogue • The Atlantic
Nancy Walecki:
»
When the iPod Nano was first released, in 2005, it cost $199 and was sold on the promise of limitlessness. This year, Celeste Stange bought her magenta 8-gigabyte model on eBay for $69 and the exact opposite reason. Every time she’d pick a song on Spotify to stream, she’d think about the millions of other songs she could listen to instead and get paralyzed by musical FOMO, she told me. “Now I only have what’s on here.”
Stange, who is 29, is part of a recent “analog” movement in which people—usually those in Gen Z—opt for less distracting alternatives to their “everything” device. Those alternatives are not always, in fact, analog, but are in many cases older digital devices: an iPod, a digital Canon PowerShot, a DVD player. Tiffany Ng, the 25-year-old author of the newsletter Cyber Celibate, now has a first-generation iPhone, an 11-year-old iPod, two CD players, a Walkman radio, and a 1986 Macintosh Plus that takes 45 seconds to load the “Welcome to Macintosh” screen.
Ng lives in Williamsburg, Brooklyn, where 20 years ago, the word analog would’ve conjured hipsters getting into vinyl. The argument then was that a record player produced real, textured music in a way an iPod’s digital files could not. But today, people are “going analog,” as they call it, to escape devices that offer endless options, in favor of those that offer a modicum of constraint.
The iPod’s original tagline—“1,000 songs in your pocket”—suggested that you could dance like the carefree silhouettes in the ads because you didn’t have to lug around CDs. This was the era when technology was “convenient, but not connected,” Tony Fadell, a former senior vice president at Apple who is known as the “father of the iPod,” told me; then the iPod became the iPhone, which soon had an app store, and convenience and connectivity became inextricably linked. People had their library of songs, but now they also had app notifications and Angry Birds and Instagram posts from that girl they went to middle school with. Eventually, streamable Apple Music replaced iTunes, and tens of thousands of songs became 100 million songs. “I can’t go through all of it,” Stange said.
…modern “analog” enthusiasts still have backups. Everyone I talked with had kept their smartphone. When Stange and I met, she took out her magenta Nano and her laptop, to show me how she downloaded her music onto the iPod. The Nano promptly died. She plugged it into the laptop to charge, but still, no dice. “I’m pretty sure it broke on the way here,” she said, and told me she’d need to figure out how to fix it. She sounded a little excited about the challenge. In the meantime, she still had her Spotify subscription.
«
Thief posed as Wi-Fi fixing hero, then stole priceless trophy • The Register
Avram Piltch:
»
On one [red teaming] assignment, [David] Schloss was asked to test the physical and network security of a company that was near the top of the Fortune 500 and had a reputation for sponsoring and providing the trophy for an international sporting competition. According to him, there were three copies made of the trophy: one for the winner, one for the host nation itself, and one for the sponsor.
When Schloss was conducting his audit, the location he visited was undergoing construction, creating problems with the office Wi-Fi that all the employees noticed and hated. So when Schloss and his team invaded the place and started probing the wireless network, no one questioned them.
“So, you got three of us that are kind of walking through this campus with antennas sticking out of our laptops. We were not being secretive at all, but we figured this is California and there’s plenty of tech bros and nerds everywhere so antennas sticking out of a computer is not going to scare people,” Schloss said. “But everyone kept coming up to us – not to ask us if we were supposed to be there, but to ask us if we were going to fix the Wi-Fi.”
After wandering the building, Schloss and his team came to the marketing department where one of the trophies, which he estimates was worth at least $250,000 (or, perhaps, priceless as there are only three), was sitting in a case. Knowing that his job was to test overall security and not just network security, he opened up the case and proceeded to remove the trophy.
Someone from the marketing department saw Schloss pulling the trophy out of the case and talked to him while he was doing it. Their question: “Are you here to fix the Wi-Fi?” When he answered “yes,” the marketing people ignored him as he slipped the trophy into his backpack.
«
It’s increasingly recognised that the best disguise is a hard hat, fluoro tabard and a clipboard. This takes it to a new level. (Red teaming is where hackers you trust are paid to try to break into your stuff and report back on how hard or easy it is.)
unique link to this extract
“Cool in 90 seconds: the fake portable air conditioners sweeping the internet • BBC News
Laura Cress:
»
As parts of the UK brace for another hot weekend, online adverts have been appearing for portable air conditioners claiming to be “designed by former Nasa engineers” and able to “cool a room in 90 seconds”.
The adverts have emerged on platforms including Facebook and YouTube, but the Advertising Standards Authority (ASA) has now warned the products are often “too good to be true”.
YouTuber Stuart Matthews, who bought several devices to test on his Proper DIY channel, told the BBC that despite paying £70 for one machine, it turned out to be “a small, simple fan worth only a few pounds”.
The BBC has approached Meta and YouTube for comment.
The ASA told the BBC that some of the adverts it had seen online in recent weeks made exaggerated claims, including that a small device could cool an entire home within minutes or used very little electricity.
It also said the adverts frequently featured fake customer reviews describing dramatic temperature drops or exceptional performance.
The adverts direct shoppers to websites selling the devices, typically for between £70 and £120.
Many of the adverts also appeared to be AI-generated, using visuals such as copper coils and metallic boxes to make the products seem more sophisticated.
«
The problem is that the claims are fake, the reviews are probably fake, so how do you figure out what’s real?
unique link to this extract
I spent a week using the Trump phone; it sucks • The Verge
Dominic Preston:
»
The T1 Phone is a curved slab of cheap gold plastic, the smartphone equivalent of a pair of knockoff wraparound Oakleys. The gold finish — more yellow in certain light, though it certainly does shine and shimmer — is tacky in every sense, with a sticky friction that makes it feel distinctly unpleasant to the touch. My phone arrived with a tiny scratch in the top-right corner.
The phone is fairly thin, and light, but its excessively curved waterfall display feels immediately dated. It also loses one of the chief advantages of that design — better in-hand feel — thanks to the oddly angular frame, which juts into my palm as I hold it.
Almost every detail speaks to bad design. There’s the American flag logo, missing a stripe. The fact that “Trump Mobile” appears on the back twice, in two different orientations and two different fonts. Or the camera module, where the three lenses are spaced at irregular intervals.
There are things to like. The 3.5mm headphone jack will have its fans, as will the microSD card slot inside the phone, or the fact that the phone ships with a case, charger, and braided USB cable. These are things that a certain type of Android fan has lamented the absence of for years.
I, for one, am more excited to be reviewing a phone with a notification light again, a true treat that I thought we’d lost forever. It’s a glimpse of a better world, one I didn’t expect from Trump Mobile of all companies. But like the curved screen, even these welcome touches betray that this is a dated, old-fashioned phone, one based on an old HTC design that already felt like a throwback two years ago.
«
The bad design and lack of attention to detail should be no surprise from a bunch of people who either used AI or incredibly lazy designers to make the standard American eagle logo for an airport they have rebranded with Trump’s name.
unique link to this extract
Apparently, Netflix is shocked that no one is watching their crap • The World Outside Our Window
Max Bolen:
»
Netflix has spent the past 15 years blitzkrieging its way to market dominance by sacrificing everything that actually makes an audience care about television. Focused entirely on hyper-growth, the streamer abandoned traditional 13-plus-episode seasons, abolished predictable yearly release schedules, and rarely bothered hiring more than a skeleton crew of writers for a temporary “mini-room.” Outside of their massive, star-driven flagship titles, they hardly market new seasons, leaning instead on an opaque algorithm to feed potential viewers whatever happens to be fresh that Tuesday. This cheap, stripped-down mode of production reliably delivered mountains of content each week, keeping Wall Street happy and cementing the myth of Netflix’s inevitable victory in a war of their own creation.
But the streaming wars are over, and Netflix has a real problem: they won.
Legacy studios have retreated to traditional theatrical windows and weekly episode drops, having realized that burning billions to replicate Netflix’s model was a corporate suicide pact. Yet, after a decade of chasing mindless growth for growth’s sake, Netflix must now reorient into something it has never had to be: a stable, trustworthy home for culture. Left to their own devices, they are being forced to contend with a simple, fatal fact: they broke TV.
The streamer’s hallmark innovation was, of course, the binge model. Guided by the data-heavy metrics of Silicon Valley, Netflix crafted shows around a single variable: immediate streams. This instant-gratification delivery system paid enormous dividends for a rapidly growing tech venture. Releasing a season all at once engineered an artificial scarcity of social relevance—watch all 10 hours in 48 hours or be locked out of the discourse. Once the platform had you hooked, its expansive library of licensed Hollywood history was there to keep you trapped. Like a casino obscuring the exit doors, all Netflix had to do was let the psychology of digital addiction take care of the rest.
But the binge model never provided a foundation for sustainability. A binged show leaves the human brain almost as quickly as it enters it. When a tech platform drops a second season two or three years later—assuming the consumer even hears about it through noise of modern society—the odds are high that whatever memory or interest they possessed has completely evaporated.
«
The point of this does seem worth reinforcing: Netflix winning is not good for TV. For instance, it doesn’t have soap operas – not in the sense that a normal terrestrial channel would. Yet those provide a subtle shared social bond through storylines that develop over months. Its nod to the destruction it has wreaked is to try to buy live sports and get more short video. That’s not TV. It’s a simulacrum of it.
unique link to this extract
The power of collaboration: how we can reduce traffic congestion • Google Research Blog
Neha Arora and Aboudy Kreidieh:
»
We ran an experiment in 10 major US cities to demonstrate the effectiveness of targeted low-cost routing interventions in improving overall traffic conditions. For this study, the Google Maps algorithm was modified to prefer alternative routes with similar travel times and segment types, effectively guiding trips away from the pre-selected congested segments.
Over a six month period, we adopted a city-wide switchback (also known as crossover) experimental design, alternating between this treatment and the control (unaltered) routing algorithm over consecutive days to appropriately measure the effect of this intervention. Rather than randomly selecting individual trips, the intervention was applied systematically across the entire city. During “treatment” days, the modified routing guided all trips that encountered the pre-selected congested segments toward alternative routes with similar travel times. Under 2% of observed trips received altered routing recommendations as a result of this experiment.
To set up the experiment, cities were chosen based on the congestion levels and ground truth availability. For each city, we selected roughly 100 road segments based on historical congestion patterns, characterised by recurring bottlenecks or high traffic density during peak demand.
«
The idea, basically, is that you send all the people using your routing apps on slightly different routes, and that turns out to reduce congestion. (Raise your hand if you thought apps like Waze – owned by Google – were already doing this.)
The full paper is published at Nature Cities: “Urban congestion relief experiments through routing-app interventions“.
unique link to this extract
Apple sues OpenAI, accuses ex-employees of stealing trade secrets • 9to5Mac
Chance Miller:
»
The complaint, filed in the U.S. District Court for the Northern District of California, alleges that [one of the accused, Tang] Tan used insider knowledge of Apple’s confidential projects to grill job candidates in interviews and learn more confidential information. Additionally, Tan directed job candidates still working at Apple to bring actual Apple hardware components and samples for “show and tell” sessions.
»
When interviewing Apple employees for jobs at OpenAI, Mr. Tan uses Apple’s confidential information to gain access to even more insider knowledge. He has used an Apple internal project codename to ask, “What’s the plan[?]” for an unannounced Apple product.
He has directed job candidates still working for Apple to bring “Actual parts” from Apple to their interviews for “show and tell” sessions in which he and his team at OpenAI can elicit still more Apple confidential information. These directions to bring Apple’s parts to OpenAI job interviews surprised at least one of the candidates, who commented that he “didn’t even know we could take those from the office.”
OpenAI has been instructing Apple employees to bring “CAD/design artifacts” and “prototypes” to their interviews and to divulge details about their work such as “subsystem and component selection,” the “tools or methodologies you use for system integration, such as CAD software, simulation tools,” and “Vendor selection and communication/collaboration with vendors.”
«
Furthermore, Apple says a candidate began “screenshotting and downloading files relating to a highly confidential Apple project” hours before interviewing with Tan, who then “solicited more information about that same Apple project” once the interview started. This became an “established pattern,” Apple says.
Tan also allegedly possessed and distributed an internal Apple “Need to Know” document to new OpenAI hires before they gave their notice to Apple. The document included Apple’s departure security protocols. As part of its investigation, Apple found a “pattern by employees who depart for OpenAI of taking steps to evade the security processes intended to protect Apple’s confidential information.”
Meanwhile, Apple also claims former engineer [Chang] Liu exploited a security bug to download confidential engineering files after leaving the company. Rather than report the exploit, Liu allegedly joked about it in messages (“LOL,” “so funny”). Liu also failed to return an Apple-issued laptop after his departure.
«
OpenAI is, one suspects, going to be very sorry that it hired these people if this gets proven in court. The complaint makes them sound like absolute loose cannons. Anyway, sayonara ChatGPT integration with Siri!
unique link to this extract
What a Monopoly vendor learned when making things in America • NPR
Scott Horsley:
»
The board game Monopoly has always taught some important economic lessons: The benefits of owning real estate. The profit potential of railroad mergers. The value of a get-out-of-jail-free card.
Now a special edition of the board game is teaching a new lesson—about how hard it is to make things in the USA.
The game is being marketed by the WS Game Company, which produces most of its high-end board games in China, just like almost every other toy maker.
After getting hit with a seven-figure tariff bill last year, CEO Jonathan Silva decided to see if it was possible to produce a profitable board game in the United States. He opted for a custom version of Monopoly, pegged to the country’s 250th birthday. But the experiment almost didn’t pass go. One big problem: No dice.
“We turned over every single leaf trying to find someone who would make 10,000 dice for us in the US,” Silva says. “It requires special machinery. It requires investment. And that type of stuff just can’t happen on a random Tuesday and be ready in a couple of months.” Silva ultimately had to settle for imported dice.
He was able to find the rest of what he needed domestically, but it wasn’t easy. A former Hasbro factory in Massachusetts prints the Monopoly board. A company called Pioneer Packaging makes the tray that holds the Monopoly money. And a small business in Indiana cranked out custom metal game tokens, in all-American shapes like a cowboy hat, a covered wagon and an apple pie. [So a Donner Party theme? – Overspill Ed.]
Just assembling all those different players took more than a year, so Silva missed the first half of the 250th birthday selling season. And the cost to manufacture the games — which retail for $80 — was at least double what it would have been in China.
«
Ten thousand dice doesn’t sound like a lot, and it doesn’t sound like the highest tech challenge ever.
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 