In Utah, the Great Salt Lake is drying up and now America wonders if market forces can refill it. CC-licensed photo by Scott Elliott Smithson 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 10 links for you. Dessicated. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.
Google just patented the end of your website • Forbes
Joe Toscano:
»
A patent granted to Google on January 27, 2026 titled “AI-generated content page tailored to a specific user” describes a system that evaluates your company’s landing page in real time and, if it decides the page won’t perform well enough for a specific user, replaces it with an AI-generated version assembled on the fly. The user never sees what your team built, they see what Google’s machine learning model thinks they should see instead.
This isn’t a feature announcement, it’s a patent, meaning Google has legally protected the ability to do this. Whether and when they deploy it is a separate question, but the direction is unmistakable – your website may soon be optional.
The system described in the patent is more sophisticated than a simple redirect. When a user submits a query, Google generates a standard search result page. But simultaneously, the system scores the most relevant landing page using signals like conversion rate, bounce rate, click-through rate, and design quality. If that score falls below a threshold – or if the page simply lacks the desired content – search results maybe be updated to include a navigation link to an AI-generated alternative.
That alternative page isn’t a cached copy of your site. It’s a dynamically assembled page built from the user’s current query, their search history, their account context, and whatever Google can extract from your original page. The patent describes possible elements including personalized headlines, suggested product filters, a product feed, sitelinks to product detail pages, and even an embedded AI chatbot. In other words, a complete brand experience built by Google. Not you.
There’s one particularly striking detail buried in the patent’s claims: “In some instances, the navigation link can be included in a sponsored content item.” What the patent does not say is how that sponsored unit would be billed, who sets it up, or whether it requires advertiser consent.
«
Google has decided that at some point people might be just no good at writing websites. Or maybe that AI is no good at writing websites, which is happening more and more. Between this and the rewriting of headlines on articles in results, Google really seems to think the rest of the web isn’t trying hard enough. (Thanks Wendy G for the link.)
unique link to this extract
Elon Musk pauses plan to improve X • Business Insider
Peter Kafka:
»
On Tuesday, [X’s head of product since last summer, Nikita] Bier announced a change to X’s revenue-sharing system: It would encourage X users who make money by posting on the platform to post about “content that resonates with people in your country, in neighboring countries and people who speak your language.”
That is: Bier’s plan was meant to discourage X users based in places like India, Thailand, and Eastern Europe from riling up big audiences in the US with incendiary posts about US politics.
“While we appreciate everyone’s opinion on American politics, we hope this will disincentivize gaming the attention of US or Japanese accounts and instead, drive diverse conversations on the platform,” Bier wrote.
That seems like a good idea. It’s also one that builds on work Bier had done previously on the platform, when he started disclosing where X users were actually based, which made it clear that lots of accounts that purported to be very interested in American politics and culture wars were run by people outside America.
Bier’s plan wouldn’t prevent those posters from weighing in on US elections or anything else. But it would make it harder for those posters to make money doing that.
And now it’s not happening, per Musk. “We will pause moving forward with this until further consideration,” he told an X user who was complaining about the new policy early Wednesday morning.
This may have been news to Bier, who was on the platform at the same time, defending the policy and telling some users they would make more money with the new system.
So what happened? I’ve asked X and Musk for comment. [Which hasn’t been forthcoming.]
It’s hard to imagine Bier making a decision that would affect power users on X without consulting with his boss, who is also X’s biggest power user. But who knows? Maybe he did, and Musk changed his mind — something he does all the time.
Or maybe he did and Musk forgot, because he’s got a lot on his mind. For instance, he is getting ready to launch what could be the biggest IPO in history, when he files to take SpaceX public — something that could happen in the next few days, per The Information.
That also means he would be taking Twitter/X public, since he merged that company with his xAI business last year, and then merged those companies into SpaceX earlier this year.
«
Bier seemed to be taking cautious steps towards making X a more trustworthy place; the revelation of where people are posting from made a huge difference in helping root out bad actors. The whole monetisation scheme has created a morass of perverse incentives which need to be unwound.
unique link to this extract
Meta and YouTube designed addictive products that harmed young people, jury finds • The Guardian
Dara Kerr:
»
Meta and YouTube have been found liable for deliberately designing addictive products that hooked a young user and led to her being harmed, a jury ruled on Wednesday. Jurors found the tech companies to be both negligent and having failed to provide adequate warnings about the potential dangers of their products.
The jury awarded the plaintiff in the case damages of $6m, with Meta to pay 70% and YouTube the remainder. It took nearly nine days of deliberations for the Los Angeles jury to reach its verdict. This lawsuit, over social media’s alleged harm to young people, was the first of its kind to go to trial.
Over the course of the six-week trial, which took place in Los Angeles superior court, jurors heard from top executives at Meta and YouTube, whistleblowers, expert witnesses on social media and addiction, and a 20-year-old woman at the center of the lawsuit, who has gone by the initials KGM for court proceedings.
KGM testified that she became addicted to YouTube at age six and Instagram at nine, which she said had deleterious effects on her wellbeing. By age 10, she said, she had become depressed and was engaging in self-harm as a result. Her social media use allegedly caused her to have strained relationships with her family and in school. When she was 13, KGM’s therapist diagnosed her with body dysmorphic disorder and social phobia, which KGM attributes to her use of Instagram and YouTube.
“How do you make a child never put down the phone? That’s called the engineering of addiction. They engineered it, they put these features on the phones,” Mark Lanier, KGM’s lawyer said during closing arguments last week. “These are Trojan horses: they look wonderful and great … but you invite them in and they take over.”
KGM’s lawyers say her experience is emblematic of what tens of thousands of young people have faced on social media and in their offline lives.
«
Juries really don’t like Meta or YouTube, do they. The question is, will this change the behaviour or design of any of their products? If the trial doesn’t come with some sort of government sanction attached, it’s hard to see anything changing.
unique link to this extract
Jury finds Meta liable in case over child sexual exploitation on its platforms • CNN Business
Clare Duffy:
»
A jury on Tuesday found Meta violated New Mexico law in a case accusing it of failing to warn users about the dangers of its platforms and protect children from sexual predators.
The jury found Meta liable on all counts, including for willfully engaging in “unfair and deceptive” and “unconscionable” trade practices, and ordered the company to pay $375m in damages.
Meta for years has faced concerns about risks to kids and teens on its platforms from parents, whistleblowers, advocates and lawmakers. Tuesday’s decision marks the first time the company has been held accountable in a jury trial for those issues.
A Meta spokesperson said the company “respectfully” disagrees and plans to appeal the decision.
New Mexico Attorney General Raúl Torrez sued Meta in 2023 for allegedly creating a “breeding ground” for child predators on Facebook and Instagram, claims that the company denies. The jury’s award was smaller than the billions in damages New Mexico had sought, but a later portion of the case to be presented directly to the judge could also force Meta to make changes to its platforms and pay additional penalties.
The case is part of a wave of legal pressure Meta and other social media platforms are facing over the safety of young users. As jurors in New Mexico state court delivered a verdict, jurors in Los Angeles are considering a separate case against Meta and YouTube accusing them of intentionally creating addictive features that harmed a young woman’s mental health. Social media giants are also facing hundreds of other cases from individuals, school districts and state attorneys general — some of which are set to go to trial later this year.
Closing arguments on Monday followed a six-week trial that included testimony from Meta executives and former employees-turned-whistleblowers. Details from the attorney general’s undercover investigation into child sexual exploitation on Meta’s platforms, which led to three arrests, were also discussed in the courtroom.
«
The fine amounts to only a few hours’ revenue for Meta as a whole, and perhaps a few days’ profit, but it’s the cumulative effect which might start to be relevant here. There is a certain echo of the tobacco trials about it all, and Don Draper is nowhere to be found.
unique link to this extract
‘Dead end’: radical 20-year study reveals genetic cloning hits a limit • ScienceAlert
Carly Cassella:
»
Scientists have pushed the limits of mammal cloning until the whole house of cards has come tumbling down.
After two decades of continuous work, researchers in Japan have discovered a genetic ‘dead end’ to mammal cloning.
The study began in 2005, when researchers, led by scientists at the University of Yamanashi in Japan, cloned a single female mouse. They then re-cloned that clone by transferring its nuclear DNA into an egg ’emptied’ of nuclear DNA, and so on and so forth, for 57 more generations, producing more than 1,200 mice from that single original donor.
Two decades later, the team was on their 58th generation, and the re-cloned mice had accumulated so many genetic mutations that they died the day after they were born.
The study is the first peer-reviewed research to ‘serially’ clone a mammal to this end.
“It has long been unclear whether mammals, unlike plants and some lower animals, could sustain their species through clonal reproduction alone,” write the research team, led by geneticist Sayaka Wakayama. “[O]ur results align closely with Muller’s ratchet theory,” they add. “This model predicts that in asexual lineages, deleterious mutations inevitably accumulate, ultimately producing mutational meltdown and extinction.”
«
This is why sexual reproduction exists: it’s an efficient way of removing deleterious mutations, since they don’t survive to reproducing age. Then again, 57 generations is a lot; mice reach reproductive maturity at about eight weeks old. If you did this in humans, you’d get to about 900 years before the end of the experiment (assuming no disastrous mutations).
unique link to this extract
Fibre optic cables reveal a big problem at the heart of farming • Grist
Matt Simon:
»
When farmers first started putting down roots, they’d plant and tend their crops by hand. With the power of oxen, they could drag plows across their fields before sowing, which boosted soil fertility and eliminated weeds. Today, that job has been made even easier by giant machines that rake the landscape.
Millennia of tilling, though, has come at a cost. While plowing releases nutrients in the short term, it degrades soil fertility in the long term, requiring farmers to load their fields with synthetic fertilizers. (The burst of microbial activity after roiling the ground also chews through accumulated carbon, returning it to the atmosphere as planet-warming greenhouse gas.) In addition, all this cultivation destroys the natural subterranean structures that hold onto water, meaning less is delivered to crops.
Fibre optic cables, of all things, have now exposed just how badly tilling messes with a farm’s ability to retain moisture. Using a technology known as distributed acoustic sensing, or DAS, scientists analyzed how seismic waves disturbed the cable as they rippled through harrowed fields compared to adjacent undisturbed plots. This created subtly distinct signals, showing that plowing obliterates the “capillaries” that carry water like tiny interconnected reservoirs.
The findings point to a serious problem with modern agriculture, to be sure, but also to solutions. “Regenerative farming practices based on principles of no-till — combined with cover crops and a diversity of crops — can basically lead to less agrochemical reliance, better soil organic matter contents, comparable yields, [and] lower diesel use,” said David Montgomery, a geomorphologist at the University of Washington and coauthor of a new paper describing the research.
DAS exploits the extreme sensitivity of fibre optic cables, which transmit information as pulses of light. If there’s a disturbance along the path — an earthquake or even someone walking overhead — a tiny bit of light bounces back to the source. With a device called an interrogator, researchers can send pulses along a length of cable and analyze what returns. Because they know the speed of light, they can differentiate a disturbance a mile down the line from one just a few hundred feet away, as the former will take just a tiny bit longer to return. Whereas a traditional seismometer takes readings at a single point, a DAS system turns miles upon miles of fibre optic cable into one continuous sensor.
«
I tried to prove I’m not AI. My aunt wasn’t convinced • BBC Future
Thomas Germain:
»
As we worked through my interview questions I stopped and asked [Hany] Farid [a digital forensics professor at the University of California, Berkeley and co-founder of GetReal Security, which works to mitigate the threat of AI deepfakes] if there was anything I could do, right now, to prove to him that I wasn’t an AI.
His answer was simple: No.
“There are things I could do to probe the system and make it less likely,” Farid says. “If you were a full-blown agentic AI, I wouldn’t hear you typing. And I can see a shadow in the background that’s pretty physically consistent as you move, and a reflection in your glasses.” There were other signs, too like the way I kept looking down as I took notes, something a deepfake wouldn’t bother with. “But at the end of the day, you’re in New York. I’m in Berkeley, California,” he says. “We’re on a video call. The reality is that you could be faking this.”
Without taking additional steps before or after our call, Farid says there’s nothing I could do to make him 100% certain I was the real Tom Germain. “No,” he says. “It’s over.”
[Chair of disinformation studies at the University of Pittsburgh in the US, Samuel] Woolley was just as hard to convince. “I could call the BBC and ask someone to double check that that you called me”, but that would take too long to figure out while we’re on the phone, he says. “For the average person, and even for people who are savvy to technological manipulation, it is very difficult to verify that someone is real,” says Woolley. For all he knows, I’m just another robot.
The solution the world’s leading experts have landed on is one your grandparents could have come up with: codewords. You, your family, business partners and anyone else you communicate with about important subjects need to come up with a secret phrase that no-one else knows you can use in an emergency to verify each other’s identities. Think of it like a convoluted form of the multi-factor authentication we all use to login online.
«
Of course any disinformation/deepfake expert isn’t going to say “oh yeah, you’re definitely human” because the downside of being wrong is too big. But they were probably 99.9% sure. However it does seem that codewords are the way forward – just as they have been for a long time in the world of anonymous scams.
unique link to this extract
Orbital data centers, part 1: There’s no way this is economically viable, right? • Ars Technica
Eric Berger:
»
Instead of being stored in 19-inch racks, the individual server elements would instead be built around—and attached to—a “satellite bus.” This is a spacecraft with large solar arrays to gather energy, thermal systems to manage heat (in a vacuum, heat must be radiated away), propulsion for orbit-keeping and maneuvering, and high-bandwidth communications gear. And it is not a theoretical idea. A company called Starcloud recently modified and launched an Nvidia H100 GPU to a small satellite bus where it is running Gemini in space.
There’s a catch, though. Replicating the output of even a single large terrestrial data center would require, at a minimum, hundreds of these satellites.
Historically, building things in space has been enormously expensive. The International Space Station, which has about the same amount of habitable space as the average American home, cost more than $150bn to construct in space. That’s on the order of 1 million times more than the cost of building a single-family home. Until recently, it cost $10,000 to put a single kilogram of payload into orbit, but costs can now be as low as one-third of that.
So yes, it does sound strange to try to build something in space that is fairly commonplace on Earth. But there is a logic here.
The biggest and most obvious advantage of putting data centers in space is the abundant energy provided by the Sun, which matters because data centers are notoriously avaricious consumers of electricity. The gathering power of a solar panel in space is five to seven times greater than a panel on Earth, depending on cloud cover, the latitude of the solar panel, and other factors that limit surface-based solar power.
Another significant advantage comes on the regulatory side. People on Earth don’t like living near data centers, considering them noisy neighbors that affect local water supplies and electricity prices. A tide of NIMBY opposition is already building.
«
With all that said, the hostility of space, the cost of assembly, the difficulty of maintenance – this all sounds like tech bros wanting to replicate what they’ve seen in SF series.
unique link to this extract
Matt Brittin: Ex-Google boss confirmed as new BBC director general • BBC
Paul Glynn and Helen Bushby:
»
Former Google executive Matt Brittin has been confirmed as the BBC’s new director general.
The ex-president of Google’s Europe, Middle East and Africa operations will replace Tim Davie, who said he was resigning in November following a storm over how Panorama edited a speech by US President Donald Trump.
BBC chairman Samir Shah said Brittin, who left Google in 2025 after 18 years, “brings to the BBC deep experience of leading a high-profile and highly-complex organisation through transformation”.
Brittin, 57, said he “can’t wait to start this work”, describing it as “a moment of real risk, yet also real opportunity”.He said the UK needs “a thriving BBC that works for everyone in a complex, uncertain and fast changing world”.
Setting out his agenda, he added: “The BBC needs the pace and energy to be both where stories are, and where audiences are. To build on the reach, trust and creative strengths today, confront challenges with courage, and thrive as a public service fit for the future.”
Shah called Brittin “an outstanding leader” with “the skills needed to navigate the organisation through the many changes taking place in the media market and in audience behaviours”.
«
One has to wonder how Brittin will cope working in an organisation which doesn’t have money to burn. Though he did work briefly for Trinity Mirror (the news group) and joined the board of the Guardian Media Group in 2025. The challenge is going to be figuring out where digital media is going, and getting there quickly.
unique link to this extract
The Great Salt Lake is running out of water — and time • Washington Post
Shawn Regan:
»
The Great Salt Lake is in trouble. After the lake’s water level reached a historic low in 2022, Utah is experiencing its worst snowpack on record. The lake is plunging toward a new low, risking the increased exposure of hundreds of square miles of lake bed contaminated with arsenic and heavy metals. When the wind blows, toxic dust blankets Salt Lake City and other towns along the Wasatch Front, one of the fastest-growing regions in one of the fastest-growing states in the country.
…The stakes are high. More than 2.5 million people live downwind of the drying lake bed. Evaporation from the lake is responsible for nearly half of the region’s precipitation. The lake also plays an underappreciated role in the national food supply; it produces minerals that provide fertilizer for organic crops and nearly half of the world’s brine shrimp supply, supporting 10 million metric tons of seafood.
To restore the lake, Utah must get more water flowing into it, and fast. Agriculture takes up roughly two-thirds of diverted water consumption in the Great Salt Lake basin. Rather than forcing farmers to give up water, Utah is paying them to lease it voluntarily and sending the savings downstream to the lake. After the lake reached a record low in 2022, Utah passed several reforms that make such transactions legally viable by allowing water rights holders to lease their water for lake restoration. The state created a $40m trust to buy or lease water for the lake and committed more than $1bn to water conservation measures. Utah is, in effect, attempting to create the freest market for water conservation in the American West.
«
The bad outcome – drying up – looks uncomfortably likely without legal restriction, which Utah is reluctant to impose.
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 