The subtleties of flavouring crisps, and choosing how to name the same flavours in different countries, are all part of the snack business. CC-licensed photo by Leonard J Matthews 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.

It’s Friday, so there’s another post due at the Social Warming Substack at about 0845 UK time.

A selection of 9 links for you. Tasty? I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. Observations and links welcome.

The race to 5G is over — now it’s time to pay the bill • The Verge

Allison Johnson:

»

At CES in 2021, 5G was just about everywhere you looked. It was the future of mobile communications that would propel autonomous vehicles, remote surgery, and AR into reality. The low latency! The capacity! It’ll change everything, we were told. Verizon and AT&T wrote massive checks for new spectrum licenses, and T-Mobile swallowed another network whole because it was very important to make the 5G future happen as quickly as possible and win the race.

CES 2024 is just around the corner, and while telecom executives were eager to shout about 5G to the rafters just a few years ago, you’ll probably be lucky to hear so much as a whisper about it this time around. While it’s true that 5G has actually arrived, the fantastic use cases we heard about years ago haven’t materialized. Instead, we have happy Swifties streaming concert footage and a new way to get internet to your home router. These aren’t bad things! But deploying 5G at the breakneck speeds required to win an imaginary race resulted in one fewer major wireless carrier to choose from and lots of debt to repay. Now, network operators are looking high and low for every bit of profit they can drum up — including our wallets.

If there’s a poster child for the whole 5G situation in the US, it’s Verizon: the loudest and biggest spender in the room. The company committed $45.5 billion to new spectrum in 2021’s FCC license auction — almost twice as much as AT&T. And we don’t have to guess whether investors are asking questions about when they’ll see a return — they asked point blank in the company’s most recent earnings call. CEO Hans Vestberg fielded the question, balancing the phrases “having the right offers for our customers” and “generating the bottom line for ourselves,” while nodding to “price adjustments” that also “included new value” for customers. It was a show of verbal gymnastics that meant precisely nothing. 

«

This is very reminiscent of the 3G bidding wars in the UK in 2000, which raised £22.5bn for five licences – and then saw huge writedowns by the licence holders a few years later. But they then did recoup it once the phones arrived that could make use of 3G. But that took until about 2010 for substantial penetration and adoption.

On that basis, it might be some time in the 2030s before we see 5G really making a mark.
unique link to this extract

USING AI to search for case law and make submissions: it makes cases up – it really does • Civil Litigation Brief

Gordon Exall:

»

If ever there was a judgment where the clue is in the name, it is Harber v Commissioners for His Majesty’s Revenue and Customs (INCOME TAX – penalties for failure to notify liability to CGT – appellant relied on case law which could not be found on any legal website – whether cases generated by artificial intelligence such as ChatGPT) [2023] UKFTT 1007.

This is a case that exemplifies the danger of relying on “Artificial Intelligence” to make legal submissions. In this case the appellant cited cases that do not exist. “Having considered all the points set out above, we find as a fact that the cases in the Response are not genuine FTT judgments but have been generated by an AI system such as ChatGPT.”

The appellant appealed to the First Tier Tax Tribunal in relation to a penalty arising from capital gains tax.  The procedure involved her filing a Response. That Response set out a number of previous decisions that appeared to assist the appellant. However there was no citation and, upon close examination, it was clear that the cases did not in fact exist. The Tribunal concluded that this was because the Response had been generated by an AI system.

«

The linked judgment did, I confess, make me laugh out loud. This is the first paragraph:

»

Mrs Harber disposed of a property and failed to notify her liability to capital gains tax (“CGT”). HMRC issued her with a “failure to notify” penalty of £3,265.11. Mrs Harber appealed the penalty on the basis that she had a reasonable excuse, because of her mental health condition and/or because it was reasonable for her to be ignorant of the law.

«

“It was reasonable for her to be ignorant of the law”?? A core principle of the law is that “ignorance is no excuse.” Secondly, this is an appeal, which means some costs have already been racked up. And all over a demand for three thousand pounds from a property sale? This surely has cost Mrs Harber a lot more than that. Plus an embarrassing place in British legal history as the first known attempt to win a case via AI-generated case law.
unique link to this extract

Early impressions of Google’s Gemini aren’t great • TechCrunch

Kyle Wiggers:

»

A “lite” version of Gemini, Gemini Pro, began rolling out to Bard yesterday, and it didn’t take long before users began voicing their frustrations with it on X (formerly Twitter).

The model fails to get basic facts right, like 2023 Oscar winners. Note that Gemini Pro claims incorrectly that Brendan Gleeson won Best Actor last year, not Brendan Fraser — the actual winner.

I tried asking the model the same question and, bizarrely, it gave a different wrong answer.

“Navalny,” not “All the Beauty and the Bloodshed,” won Best Documentary Feature last year; “All Quiet on the Western Front” won Best International Film; “Women Talking” won Best Adapted Screenplay; and “Pinocchio” won Best Animated Feature Film. That’s a lot of mistakes. [It also offers a link to “the official Oscars website” which is not the official Oscars website oscars.org.]

Science fiction author Charlie Stross found many more examples of confabulation in a recent blog post. (Among other mistruths, Gemini Pro said that Stross contributed to the Linux kernel; he never has.)

Translation doesn’t appear to be Gemini Pro’s strong suit, either. It struggles to give a six-letter word in French [it suggested “amour” to one Twitter user]. When I ran the same prompt through Bard (“Can you give me a 6-letters word in French?”), Gemini Pro responded with a seven-letter word instead of a five-letter one — which gives some credence to the reports about Gemini’s poor multilingual performance.

«

This is about as unimpressive as you can get. As with anything, fast and unreliable is not preferable over slow and reliable. Plus: Google’s video showing it off was heavily edited (which isn’t that surprising, to be honest).
unique link to this extract

‘How do you reduce a national dish to a powder?’: the weird, secretive world of crisp flavours • The Guardian

Amelia Tait:

»

Reuben and Peggy’s jobs are not top secret in the way top secret jobs usually are. They don’t have guns, for example – and the grey conference table they sit at is much the same as you’d find in any office in the UK. They even have LinkedIn profiles that tell you their job titles. But this is where things get odd: search the name of the company they work for – a name I have agreed not to print – and you’ll find little information about the work Reuben and Peggy do. You could click through every page on their company’s website and leave with no idea that it creates the most beloved crisp flavours in the world.

Reuben and Peggy are not their real names. Reuben is a snacks development manager and Peggy is a marketer, and they work for a “seasoning house”, a company that manufactures flavourings for crisps.

I meet the pair on Zoom, hoping they can answer a question that has consumed me for years. In January 2019, I was visiting Thailand when I came across a pink packet of Walkers with layered pasta, tomato sauce and cheese pictured on the front. Lasagne flavour, the pack said. You can’t get lasagne Walkers – or Lay’s, as they are known in most of the world – in Italy. Relatively speaking, Italians have a small selection of Lay’s – paprika, bacon, barbecue, salted and Ricetta Campagnola, a “country recipe” flavour featuring tomato, paprika, parsley and onion. I’ve sampled Hawaii-style Poké Bowl crisps in Hungary and chocolate-coated potato snacks in Finland; I have turned away from Sweet Mayo Cheese Pringles in South Korea. So why can you get lasagne flavour Lay’s in Thailand but not in Italy, home of the dish? Who figures out which country gets which crisps?

«

This is wonderful. It was referred to by Stuart Maconie and Mark Radcliffe, who do the RadMac show on BBC 6 Music on weekend mornings, because they do the amazing “Crisps on the radio” segment in which a listener sends in a packet of crisps – the weirder and more obscure the flavour the better – and they have to try to work out what it is from a live taste test. It’s as strange and wonderful as it sounds.
unique link to this extract

Podcasters took up her sister’s murder investigation. Then they turned on her • The New York Times

Sarah Viren:

»

[Liz] Flatt was at a crossroads in what she had taken to calling her journey, a path embarked on after a prayer-born decision five years earlier to try and find who killed her sister, Deborah Sue Williamson, or Debbie, in 1975. It was now 2021. Flatt was middle-aged and coming out of one of the darkest moments of her life. Her mother had died, quite suddenly, two years earlier, and the grief from her death almost destroyed Flatt. Her father was gone, too — dead from a heart attack after years of fighting for the police to reinvestigate Debbie’s killing — and her older brother, Ricky, who was once a suspect in the murder, took his own life five years before that.

She had come to Austin [in Texas] for a conference, CrimeCon, which formed around the same time that Flatt began her quest, at a moment now seen as an inflection point in the long history of true crime, a genre as old as storytelling but one that adapts quickly to new technologies, from the printing press to social media. The gathering was smaller in 2021 because of the pandemic, but Nancy Grace, queen of true crime’s TV era, still showed up, as did Dr. Phil. On “Podcast Row,” Flatt wandered among booths for “Cults, Crimes & Cabernet” and “Murderish,” for “True Crime Garage” and “Die-alogue,” less a fan of the genre, which she never liked that much, than a scout on a search.

She ran into a podcaster who covered Debbie’s story a couple of years before, a man who goes by the name Vincent Strange, and she commiserated with a woman whose mother’s murder also remained unsolved. Then, at another booth, Flatt met a woman who would later put her in touch with two investigators who presented at the conference that year: George Jared and Jennifer Bucholtz. They were podcasters, but Jared was also a journalist and Bucholtz an adjunct professor of forensics and criminal justice at the for-profit American Military University. Their presentation was on another cold case, the murder of Rebekah Gould in 2004, whose killer they claimed to have helped find using a technique that has quickly become a signature of the changing landscape of true crime: crowdsourcing.

«

Crowdsourcing, however, means Facebook, and a group of people trying to “solve” a crime on Facebook means you have a tiger by the tail.
unique link to this extract

COP28 so far: a cheat sheet • Heatmap News

Jessica Hullinger:

»

• The loss and damage fund: On day one of the conference, world leaders reached a landmark deal to help vulnerable nations deal with the costly effects of climate change. The early accomplishment set an optimistic tone for the summit — although The Guardian notes that wealthy countries have so far pledged $700m to the fund, “far short of what is needed.” In total, countries have announced $57bn of various funding pledges at the conference.

• Methane cuts: About 50 oil and gas companies pledged to slash their methane leaks by 2030. Critics cry greenwashing, but as Heatmap’s Emily Pontecorvo points out, recent technological advances in methane monitoring – including satellites, drones, and handheld detectors – could help in the international effort to hold these companies accountable. A planned $40m infusion from billionaire philanthropist Michael Bloomberg will bolster the cause, too.

• A renewables pledge: At least 120 countries backed a pledge to triple global renewable energy capacity by 2030. That goal made it into an early draft of the global stocktake report, the summit’s final deliverable, but that’s no guarantee it will be formally adopted.

• A nuclear energy declaration: More than 20 countries including the US, Canada, the UK, and the United Arab Emirates, pledged to triple global nuclear energy capacity by 2050.

• Growing support for a fossil phase-out: The number of countries pledging to voluntarily end oil and gas extraction and exploration grew to 24 when Spain, Kenya, and Samoa joined the Beyond Oil & Gas Alliance

• A global cooling pledge: More than 60 countries pledged to reduce their cooling-related emissions by at least 68% by 2050.

«

That last one puzzled me, so I looked it up: it’s about emissions caused by air conditioning and similar. Heatmap News is an interesting new media site covering climate and related issues. Worth a look.
unique link to this extract

iPad Air 12.9-inch and MacBook Air with M3 expected in March 2024 • Apple Insider

Mike Wuerthele:

»

The end of the winter may herald hardware refreshes for Apple, with a new report from the industry’s most prolific leaker predicting the long-rumoured larger iPad Air and refreshed MacBook Air models will hit store shelves by the end of March 2024.

To combat sales doldrums for Mac and iPad, Apple is rumored to be prepping many new releases before the first calendar quarter of 2024 ends. On tap are allegedly a larger iPad Air, new iPad Pros with OLED screens, and a New MacBook Air model, presumably with M3 processor.

«

The OLED iPad Pros should do well – deeper blacks and wider colours is attractive. Apparently there are new keyboards for the iPads coming too.

I linked to this rather than Mark Gurman’s original report at Bloomberg because 1) this version avoids the strangulated “people familiar with the situation who asked not to be identified” formulation for “my sources in the supply chain” 2) it also avoids the struggling construction that Apple’s doing this “to combat [a] sales slump”. Nope, it’s just doing this because it refreshes products. Sales go up and down, and – iPhones apart – it’s unusual to make a big difference.
unique link to this extract

23andMe is updating its TOS to force binding arbitration with a limited opt-out window • Stackdiary

Alex Ivanovs:

»

23andMe, the personal genomics and biotechnology company, has been trying to contain a security breach that was first disclosed on October 6th. On October 19th, 23andMe disclosed another security breach by the same hacker who had initially claimed responsibility. The hacker said he had access to more than 4 million genetic profile records this time. And on December 4th, 23andMe confirmed that the total scope of the breach was 6.9 million users in total.

The fallout of this disclosure, which started in October, was swift. By October 14th, several individuals had already filed lawsuits against 23andMe for negligence, as Stack Diary reported. Likewise, the general consensus of 23andMe users has been that the company handled the situation very poorly.

To add insult to injury, Stack Diary can reveal that 23andMe is now rolling out an update to its Terms of Service. This change will force its users into binding arbitration, which is a means to resolve disputes (such as a cybersecurity breach leaking your DNA data) outside of court.

In this process, both parties in a disagreement present their cases to an arbitrator, who is a neutral third party. The arbitrator listens to both sides, reviews the evidence, and decides. The key aspect of binding arbitration is that the arbitrator’s decision is final and legally enforceable, meaning both parties must accept it and cannot appeal to a regular court.

«

Users get 30 days to opt out of these terms which Stack Diary says “significantly reduce their rights”, adding

»

The email doesn’t mention that you must email the “arbitrationoptout@23andme.com” address to opt out of forced arbitration, as outlined in the updated Terms of Service

«

I’d call that fundamentally sneaky, but the news is going to get around pretty quickly.
unique link to this extract

‘Signs of life’: Sycamore Gap tree will live on, experts say • NPR

Bill Chappell:

»

The tree occupied a magical spot in the landscape of Northumberland, England, and in the hearts of people who visited it. So the news that efforts to propagate the ancient tree will likely succeed is being welcomed now, after the tree was felled in September.

“[We] are encouraged by positive signs of life, and are hopeful that over 30% of the mature seeds and half of the cuttings (scions) will be viable,” said Andy Jasper, the National Trust’s director of gardens and parklands, in a statement sent to NPR.

“Over the next year, we’ll be doing all we can to nurture the seeds and cuttings, in the hope that some will grow into strong, sturdy saplings,” Jasper said, “providing a new future for this much-loved tree.”

The sycamore’s trunk might also regrow, Jasper said, but it could be several years before it’s known whether that will bear out.

…A 16-year-old boy was arrested shortly after the tree was cut down, in what police said was an act of deliberate vandalism. But Northumbria Police recently said the teen “will now face no further action by police.” Instead, their focus is on three men — two in their 30s and one in his 60s — who were arrested in the weeks since the incident.

Police haven’t divulged many details about the three remaining suspects, but media reports have suggested at least one of them is a former lumberjack who was in possession of a large chainsaw.

«

Pah, just circumstantial. Did you take the tree’s DNA, copper? Did ya? Oh… really, you did?
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

Golf balls will be altered so they don’t fly as far, under rules being introduced from 2028, as pros drive them further and further. CC-licensed photo by cretinbob 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. Clubbing together. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. Observations and links welcome.

Google just launched Gemini, its long-awaited answer to ChatGPT • WIRED

Will Knight:

»

Gemini, a new type of AI model that can work with text, images, and video, could be the most important algorithm in Google’s history after PageRank, which vaulted the search engine into the public psyche and created a corporate giant.

An initial version of Gemini starts to roll out from Wednesday inside Google’s chatbot Bard for the English language setting. It will be available in more than 170 countries and territories. Google says Gemini will be made available to developers through Google Cloud’s API from December 13. A more compact version of the model will from today power suggested messaging replies from the keyboard of Pixel 8 smartphones. Gemini will be introduced into other Google products including generative search, ads, and Chrome in “coming months,” the company says. The most powerful Gemini version of all will debut in 2024, pending “extensive trust and safety checks,” Google says.

“It’s a big moment for us,” Demis Hassabis, CEO of Google DeepMind, told WIRED ahead of today’s announcement. “We’re really excited by its performance, and we’re also excited to see what people are going to do building on top of that.”

Gemini is described by Google as “natively multimodal,” because it was trained on images, video, and audio rather than just text, as the large language models at the heart of the recent generative AI boom are. “It’s our largest and most capable model; it’s also our most general,” Eli Collins, vice president of product for Google DeepMind, said at a press briefing announcing Gemini.

Google says there are three versions of Gemini: Ultra, the largest and most capable; Nano, which is significantly smaller and more efficient; and Pro, of medium size and middling capabilities.

From today, Google’s Bard, a chatbot similar to ChatGPT, will be powered by Gemini Pro, a change the company says will make it capable of more advanced reasoning and planning.

«

You can use it in the Bard Chatbot right now if you want. The little video that went with it is… entertaining, but I still want something that will organise calendars, reply sensibly to emails for me, point to odd things happening which shouldn’t be in my electronic life. (But would you trust a chatbot with all of your life? There’s a Black Mirror episode waiting to be written.)
unique link to this extract

Animate Anyone • Institute for Intelligent Computing, Alibaba Group

Li Hu and others:

»

In this paper, we leverage the power of diffusion models and propose a novel framework tailored for character animation. To preserve consistency of intricate appearance features from reference image, we design ReferenceNet to merge detail features via spatial attention. To ensure controllability and continuity, we introduce an efficient pose guider to direct character’s movements and employ an effective temporal modeling approach to ensure smooth inter-frame transitions between video frames.

By expanding the training data, our approach can animate arbitrary characters, yielding superior results in character animation compared to other image-to-video methods. Furthermore, we evaluate our method on benchmarks for fashion video and human dance synthesis, achieving state-of-the-art results.

«

You have to see the video clips, really, but this is the sort of thing that actors are concerned about: their images being used to create moving pictures, made to be doing anything.
unique link to this extract

Governments spying on Apple, Google users through push notifications, US senator reveals • Reuters

Raphael Satter:

»

Unidentified governments are surveilling smartphone users via their apps’ push notifications, a US senator warned on Wednesday.

In a letter to the Department of Justice, Senator Ron Wyden said foreign officials were demanding the data from Alphabet’s Google and Apple. Although details were sparse, the letter lays out yet another path by which governments can track smartphones.

Apps of all kinds rely on push notifications to alert smartphone users to incoming messages, breaking news, and other updates. These are the audible “dings” or visual indicators users get when they receive an email or their sports team wins a game. What users often do not realize is that almost all such notifications travel over Google and Apple’s servers.

That gives the two companies unique insight into the traffic flowing from those apps to their users, and in turn puts them “in a unique position to facilitate government surveillance of how users are using particular apps,” Wyden said. He asked the Department of Justice to “repeal or modify any policies” that hindered public discussions of push notification spying.

In a statement, Apple said that Wyden’s letter gave them the opening they needed to share more details with the public about how governments monitored push notifications. “In this case, the federal government prohibited us from sharing any information,” the company said in a statement. “Now that this method has become public we are updating our transparency reporting to detail these kinds of requests.”

Google said that it shared Wyden’s “commitment to keeping users informed about these requests.”

«

unique link to this extract

Running Signal will soon cost $50 million a year • WIRED

Andy Greenberg:

»

Signal was originally founded with money from the US government-funded Open Technology Fund, but the service has since turned to donations to keep afloat. When the Signal Foundation was created in 2018 and WhatsApp cofounder Brian Acton left Facebook to become its president, he donated $50m. But with Signal’s growing user base and staff, that donation wouldn’t cover much more than a year’s current budget for the company. Other major donors continue to cover the foundation’s costs, Whittaker says—Twitter cofounder Jack Dorsey, for instance, has pledged $1m a year, and others Whittaker declines to name have given similarly large contributions.

But Signal hopes to increasingly rely on donations of as little as $3 that can be made through the app itself. Monthly donations of $5 or more are rewarded with a badge for the user’s account. Those small donations, Signal says, now account for 25% of its operating costs, up from 18% last year, the first full year after Signal enabled in-app contributions. But for Signal to continue to exist and grow without depending on a few wealthy individuals, Whittaker says small user donations will need to ramp up significantly.

With a nearly $50m annual budget, can Signal actually survive on those donations? “We have to,” says Whittaker. “Signal needs to find a way to survive in perpetuity because it is the tool that we have to ensure meaningfully private communications.”

Whittaker says that charging users has never been an option—Signal would never have grown its network to a degree that could compete with iMessage or WhatsApp if it hadn’t been free all along. Nor can Signal adopt a venture capital-funded business model that would leave the service vulnerable to investors or shareholders demanding a profitable exit. Exhibit one: Elon Musk’s acquisition of Twitter and his decisions that triggered an exodus of its users.

«

What, another article about Signal and its funding? Yes, because it’s been pointed out to me (thanks, Paul C) that yesterday’s article was written by an author who could be thought of as unreliable in claiming that the CIA has suddenly cut funding. As this shows, that happened quite a while back. Anyway, if you use it, donate.
unique link to this extract

iMessage will reportedly dodge EU regulations, won’t have to open up • Ars Technica

Ron Amadeo:

»

The EU is deciding what should and shouldn’t be under the new rules set out by the “Digital Markets Act.” The idea is that Big Tech “gatekeepers” will be subject to certain interoperability, fairness, and privacy rules. So far the wide-ranging rules have targeted 22 different services, including app stores on iOS and Android, browsers like Chrome and Safari, the Android, iOS, and Windows OSes, ad platforms from Google, Amazon, and Meta, video sites YouTube and TikTok, and instant messaging apps like WhatsApp and Facebook Messenger.

Google recently rolled out a campaign to implore the EU to qualify iMessage for regulation, as Android’s iMessage incompatibility is a big deal in the US. iMessage hasn’t made the list, though, and that’s despite meeting the popularity metrics of 45 million monthly active EU users. In the EU and most other parts of the world, the dominant messaging platform is WhatsApp, and with the Digital Market Act’s focus on business usage, not general consumers, iMessage will just squeak by. Right now the EU is “investigating” a handful of borderline additions to the Digital Markets Act, with a deadline in February 2024.

Qualifying for the law would have forced iMessage to allow interoperability with other services, so theoretically, you’d be allowed to log in to iMessage from WhatsApp, Facebook Messenger, and whatever else.

«

unique link to this extract

The inside story of Microsoft’s partnership with OpenAI • The New Yorker

Charles Duhigg was embedded in OpenAI when Everything Happened:

»

Some members of the OpenAI board had found Altman an unnervingly slippery operator. For example, earlier this fall he’d confronted one member, Helen Toner, a director at the Center for Security and Emerging Technology, at Georgetown University, for co-writing a paper that seemingly criticized OpenAI for “stoking the flames of AI hype.”

Toner had defended herself (though she later apologized to the board for not anticipating how the paper might be perceived). Altman began approaching other board members, individually, about replacing her. When these members compared notes about the conversations, some felt that Altman had misrepresented them as supporting Toner’s removal. “He’d play them off against each other by lying about what other people thought,” the person familiar with the board’s discussions told me. “Things like that had been happening for years.” (A person familiar with Altman’s perspective [it’s Altman – Overspill Ed] said that he acknowledges having been “ham-fisted in the way he tried to get a board member removed,” but that he hadn’t attempted to manipulate the board.)

… when four members of the board—Toner, D’Angelo, Sutskever, and Tasha McCauley—began discussing his removal, they were determined to guarantee that he would be caught by surprise. “It was clear that, as soon as Sam knew, he’d do anything he could to undermine the board,” the person familiar with those discussions said.

The unhappy board members felt that OpenAI’s mission required them to be vigilant about AI becoming too dangerous, and they believed that they couldn’t carry out this duty with Altman in place. “The mission is multifaceted, to make sure AI benefits all of humanity, but no one can do that if they can’t hold the C.E.O. accountable,” another person aware of the board’s thinking said. Altman saw things differently. The person familiar with his perspective said that he and the board had engaged in “very normal and healthy boardroom debate,” but that some board members were unversed in business norms and daunted by their responsibilities. This person noted, “Every step we get closer to AGI [artificial general intelligence], everybody takes on, like, ten insanity points.”

«

So that seems to be the story: basically the board didn’t like Altman or his attitude about AGI. Nothing much more complicated than that.
unique link to this extract

Is this the end of ‘Intel Inside’? • WSJ

Christopher Mims:

»

The threats to Intel are so numerous that it’s worth summing them up: The Mac and Google’s Chromebooks are already eating the market share of Windows-based, Intel-powered devices. As for Windows-based devices, all signs point to their increasingly being based on non-Intel processors. Finally, Windows is likely to run on the cloud in the future, where it will also run on non-Intel chips.

Apple has moved almost entirely away from Intel’s chips, which it used for over a decade for all of its desktop and notebook computers. At the same time, its overall market share for desktops and notebooks has climbed from around 12% of devices in the US in 2013 to nearly one in three today, according to Statcounter.

These days, it’s not just Apple moving away from Intel’s chips. Microsoft is accelerating its yearslong effort to make Windows run on ARM-based processors, so that the entire PC ecosystem isn’t doomed by Intel’s failure to keep up with Apple and TSMC. Google’s Chrome OS, which works with either Intel or ARM-based chips, is also an emerging threat to Microsoft.

This means the threat to Intel comes from a whole ecosystem of companies with deep pockets and sizable profit margins, each trying to take their piece of the company’s market share. In many ways, it really is Intel versus the world—and “the world” includes nearly every tech giant you can name. 

It wasn’t always this way. For decades, Intel enjoyed PC market dominance with its ride-or-die partner, Microsoft, through their “Wintel” duopoly.

It’s ironic, then, that Microsoft is one of the companies leading the charge away from Intel’s chips.

«

That Statcounter figure seems quite optimistic; the caveat is that it’s only the US, and it’s measured via browsers (so that won’t include PCs used just on intranets without external connections). Intel, though, is in all sorts of trouble.
unique link to this extract

Maybe we already have runaway machines • The New Yorker

Gideon Lewis-Kraust:

»

One of the things that make the machine of the capitalist state work is that some of its powers have been devolved upon other artificial agents—corporations. Where [Cambridge professor David] Runciman compares the state to a general AI, one that exists to serve a variety of functions, corporations have been granted a limited range of autonomy in the form of what might be compared to a narrow AI, one that exists to fulfill particular purposes that remain beyond the remit or the interests of the sovereign body.

Corporations can thus be set up in free pursuit of a variety of idiosyncratic human enterprises, but they, too, are robotic insofar as they transcend the constraints and the priorities of their human members. The failure mode of governments is to become “exploitative and corrupt,” Runciman notes. The failure mode of corporations, as extensions of an independent civil society, is that “their independence undoes social stability by allowing those making the money to make their own rules.”

There is only a “narrow corridor”—a term Runciman borrows from the economists Daron Acemoglu and James A. Robinson—in which the artificial agents balance each other out, and citizens get to enjoy the sense of control that emerges from an atmosphere of freedom and security. The ideal scenario is, in other words, a kludgy equilibrium.

«

This is a review of Runciman’s book, in which he points out that states and corporations have in effect been uncontrolled AIs of a sort for quite some time already; and so the concerns about the new machine-based AIs have already been rehearsed, just in a different context. (And how well, exactly, have we managed them?)
unique link to this extract

New golf ball rules: R&A and USGA opt to limit distance ball will travel in air • BBC Sport

Iain Carter:

»

Modern premium golf balls (which cost around £6 each) when struck with the latest large-headed drivers have never flown as far as they do today.

The PGA Tour’s biggest hitter, Rory McIlroy, is among several players whose drives average more than 320 yards, with 98 pros beating the circuit’s average of 299.9 yards last season.

In 2002 only one player, John Daly (306 yards), beat the 300 yard barrier. This year the Masters was forced to lengthen Augusta’s famous par-five 13th hole from 510 to 545 yards to make sure it remains an appropriate challenge.

The new measures come into force in January 2028 for the elite game, with a phased introduction for recreational golfers in 2030.

Golf balls must conform to the rules and pass strict testing protocols which determine their ‘Overall Distance Standard’. The playing characteristics of a ball can be altered through its composition and/or dimple patterns which in turn can affect spin rates that could limit the distance it flies.

Under current regulations, a ball struck by a robotic club swung in laboratory conditions at 120mph (193kph) is only allowed to travel 317 yards (289.9m) (with three yards/1m tolerance). The new rules will maintain the same distance outcome, but for a club swung at the increased rate of 125mph, which is the top end of the speed generated by pros.

“We feel very strongly that we need to act and update the rules for the modern game,” Slumbers said. “It is 20 years since we last updated the golf ball and a lot has changed in sport, and in golf, in that time.”

The St Andrews-based boss added: “We feel that [a reduction of] 15 yards for the longest hitters is fair and will have a meaningful impact.

“But it is very important to understand that for the average recreational golfer we will see an impact of less than five yards.”

«

And what’s going to happen? Top golf pros will figure out how to swing their club faster – perhaps 140mph. Though 160 mph could be the top speed humanly possible. (But if you make the club longer…) I do like these tales of equipment being reined in to try to keep sports within their stadia.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

How old was the world’s oldest dog? And how hard has Guinness World Records tried to confirm it? CC-licensed photo by Daniel Spils 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. Following a lead. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. Observations and links welcome.

Signal facing collapse after CIA cuts funding • Kit’s Newsletter

Kit Klarenberg:

»

On November 16th, Meredith Whittaker, President of Signal, published a detailed breakdown of the popular encrypted messaging app’s running costs for the very first time. The unprecedented disclosure’s motivation was simple – the platform is rapidly running out of money, and in dire need of donations to stay afloat. Unmentioned by Whittaker, this budget shortfall results in large part due to the US intelligence community, which lavishly financed Signal’s creation and maintenance over several years, severing its support for the app.

Never acknowledged in any serious way by the mainstream media, Signal’s origins as a US government asset are a matter of extensive public record, even if the scope and scale of the funding provided has until now been secret. The app, brainchild of shadowy tech guru ‘Moxie Marlinspike’ (real name Matthew Rosenfeld), was launched in 2013 by his now-defunct Open Whisper Systems (OWS). The company never published financial statements or disclosed the identities of its funders at any point during its operation.

Sums involved in developing, launching and running a messaging app used by countless people globally were nonetheless surely significant. The newly-published financial records indicate Signal’s operating costs for 2023 alone are $40m, and projected to rise to $50m by 2025. Rosenfeld boasted in 2018 that OWS “never [took] VC funding or sought investment” at any point, although mysteriously failed to mention millions were provided by Open Technology Fund (OTF).

«

This is not good news. But it’s also puzzling: what has changed that has led the US to stop funding Signal? Klarenberg doesn’t know. And nor do we.
unique link to this extract

Linda Yaccarino, Elon Musk and X: behind the mess, ad exodus • The Hollywood Reporter

Kim Masters:

»

By now, Yaccarino has become one of the best-known CEOs in America, if not for the most desirable reasons. Amid all the noise and controversy, prominent voting-rights attorney Marc Elias posted: “I had never heard of Linda Yaccarino before her joining X, but was she this ridiculous in her last job?”

According to many former associates at NBCU, the answer is a qualified no. Though several describe her as a difficult and volatile boss or colleague, they say she was an extremely hardworking and capable ad-sales executive. Advertisers — who she was, of course, always courting — also praise her. In mid-November, after Forbes reported that marketing leaders were urging Yaccarino to resign, Axios quoted Lou Paskalis, founder and CEO of marketing consultancy AJL Advisory, saying that “the advertising community is now working to save the reputation of a beloved member of our industry who does not share Elon Musk’s views.” 

In fact, it’s unclear what Yaccarino thinks of Musk’s views; after his Nov. 15 tweet endorsing an antisemitic trope as “the actual truth,” she touted the site’s “efforts to combat antisemitism and discrimination.” (Her views on Donald Trump are much clearer. Associates say she was an enthusiastic supporter. She was appointed to the President’s Council on Sport, Fitness and Nutrition during his administration.)

Based on conversations with multiple sources who worked with or for Yaccarino at NBCU, the word “beloved” is not one that many would use to describe the way she was seen internally. “She was good at ad sales but wrecked the culture,” says a former insider. “She was not collegial. She was a scorched-earth manager.”

«

This is not, it is safe to say, a hagiography.
unique link to this extract

Seven new features to express yourself on Google Messages • Google Blog

Sanaz Ahari, VP and GM of Android and Business Communications at Google:

»

Today marks a new milestone that we are incredibly proud of: there are now more than one billion monthly active users with RCS enabled in Google Messages. We are grateful to our partners and our users that have advocated for RCS over the years — it’s been a lot of work to get here, and we want to thank you.

Beyond Google Messages, there are other messaging clients that use RCS and we are pleased that Apple also took their first step two weeks ago in announcing that they’re embracing RCS.

«

One billion, but monthly users – so someone who uses that once in a month, just a single time, counts. And it’s not a huge proportion of all the Android users. Getting RCS onto iOS will certainly expand that number significantly… in the US. I suspect that in the rest of the world, people use WhatsApp and Signal and so on, and don’t get hung up about blue and green bubbles. (Or they just use WhatsApp or Signal from the off.)
unique link to this extract

A new kind of solar cell is coming: is it the future of green energy? • Nature

Mark Peplow:

»

A few niche perovskite-based PV products are already on the market, but announcements this year signal that many more are set to join them. Case says that end users should get their hands on solar panels made from Oxford PV’s cells around the middle of next year, for example. In May, a large silicon PV manufacturer, Hanwha Qcells, headquartered in Seoul, said it plans to invest US$100m in a pilot production line that could be operational by the end of 2024.

Silicon is the workhorse material inside 95% of solar panels. Rather than replace it, Oxford PV, Qcells and others are piggybacking on it — layering perovskite on silicon to create so-called tandem cells. Because each material absorbs energy from different wavelengths of sunlight, tandems could potentially deliver at least 20% more power than a silicon cell alone; some scientists project much greater gains.

Perovskite supporters say that this extra electricity could more than offset the additional costs of tandem cells, particularly in crowded urban areas or industrial sites where space is at a premium. “Our biggest initial demand is from utilities, because they simply don’t have enough accessible land,” says Case.

«

Back in August 2020 I linked to a Guardian article about perovskites which said they’d be in production by 2021. Oh well, a few years here and there.. though the problem is degradation, which might delay things by another couple of years. Again.
unique link to this extract

Your 2023 WebMD Wrapped • McSweeney’s Internet Tendency

Like Spotify Wrapped, but for your hypochondria. McSweeney’s is always a great read.
unique link to this extract

Streaming apps are trying to bundle their way out of customer disenchantment • Ars Technica

Scharon Harding:

»

A Netflix-Max bundle through Verizon and a potential bundle with Apple TV+ and Paramount+ follow a trend that sees streaming apps partnering with other apps (including rivals) and other types of companies with subscription-based revenue to ultimately offer TV streaming at a lower monthly price.

Similarly to Verizon, T-Mobile offers bundles for its mobile services with Netflix and Apple TV+. You can get Disney+ with Hulu and ESPN, and Disney, which will soon own all of Hulu, is launching a unified Disney+ and Hulu app. HBO’s Max and Discovery+ merged into Max. Paramount+ offers Showtime content, and Showtime’s Anytime app (for people subscribed to Showtime via a TV provider) is shutting down on December 14. Other streaming-related bundle deals currently being pushed include Paramount+ with Walmart+ and Peacock with Instacart+ or Xfinity.

Striking a deal between multiple conglomerates is complex, though. Companies see less revenue per user when adding customers through promotions and bundles compared to direct sales, WSJ reported in October 2022. Involved companies need to agree on how to divide monthly subscription fees, customer data, and advertising sales. As a result, “such talks in the industry have progressed slowly,” The Washington Post reported at the time.

But over a year after WSJ’s report, streaming bundles are happening “faster than we thought” and are “here to stay,” Erin McPherson, senior vice president and chief content officer at Verizon Communications, said, according to WSJ’s report Friday. Verizon’s CEO, Hans Vestberg, added at a UBS conference today that creating new types of bundles is a company priority, as per The Hollywood Reporter.

«

So basically, there were terrestrial channels, then cable bundled lots together, then streaming unbundled them, and now streaming is bundling them back together again. Place your bets on when the cycle turns around, and what prompts the next unbundling.
unique link to this extract

Thieves rob DC Uber Eats driver, reject Android phone for not being iPhone • KATV Washington

Carl Willis:

»

After working into the early morning hours, a woman who asked not to be identified said her husband insisted he meet her outside of their apartment in Northwest [Washington] D.C. and go park the car.

“As soon as he parked the car two masked gentlemen came up to him, armed,” she said. “They robbed him, took everything he had in his pockets, took the keys to my truck and got in and pulled off.”

She said one of them approached on foot in the 2400 block of 14th Street, NW. The other was in a black BMW, both of them armed with guns. She said the robbers were bold taking her husband’s phone, but then giving it back because it wasn’t to their liking. “They basically looked at that phone and was like ‘Oh, that’s an Android? We don’t want this. I thought it was an iPhone,'” she said.

The bizarre encounter lasted only seconds, but she said the impact turned her life upside-down.

“That [truck] was my income,” she said. “That was the way I made money. I did Uber Eats and Instacart so, that was our livelihood.”

«

The opening paragraph is a bit confusing. I interpret what happened as the following: the husband arrives home, and parks his car. The two thieves confront him: they have had their eyes on the truck, so they tell him to call his wife (because he doesn’t have its key) and get her to come outside; she does. Et voila.

And the Android phone? Insufficient resale value.
unique link to this extract

Misinformation researcher Joan Donovan accuses Harvard of bowing to Facebook • The Washington Post

Joseph Menn:

»

A prominent disinformation scholar has accused Harvard University of dismissing her to curry favor with Facebook and its current and former executives in violation of her right to free speech.

Joan Donovan claimed in a filing with the Education Department and the Massachusetts attorney general that her superiors soured on her as Harvard was getting a record $500m pledge from Meta founder Mark Zuckerberg’s charitable arm.

As research director of Harvard Kennedy School projects delving into mis- and disinformation on social media platforms, Donovan had raised millions in grants, testified before Congress and been a frequent commentator on television, often faulting internet companies for profiting from the spread of divisive falsehoods.
Last year, the school’s dean told her that he was winding down her main project and that she should stop fundraising for it. This year, the school eliminated her position. The surprise dismissal alarmed fellow researchers elsewhere, who saw Donovan as a pioneer in an increasingly critical area of great sensitivity to the powerful and well-connected tech giants.

Donovan has remained silent about what happened until now, filing a 248-page legal statement obtained by The Washington Post that traces her problems to her acquisition of a trove of explosive documents known as the Facebook Papers and championing their importance before an audience of Harvard donors that included Facebook’s former top communications executive.

Harvard disputes Donovan’s core claims, telling The Post that she was a staff employee and that it had not been able to find a faculty sponsor to oversee her work, as university policy requires. It also denies that she was fired, saying she “was offered the chance to continue as a part-time adjunct lecturer, and she chose not to do so.”

«

The timing does look suspicious, but at the same time it’s all coincidental; the money from the Zuckerberg foundation is (as the filing says) for “a university-wide centre on artificial intelligence”. One suspects that any nudges and winks about Donovan’s position, if they occurred, happened verbally behind closed doors.
unique link to this extract

Was Bobi the world’s oldest dog—or a fraud? • WIRED

Matt Reynolds:

»

On October 21, 2023, Bobi the dog died. As with most celebrity deaths, the press coverage was wall-to-wall, but Bobi’s demise wasn’t unexpected. At 31 years and 163 days (or 217 in “human” years), he was old. So old, in fact, that in February 2023 Bobi had been crowned the “oldest dog ever” by Guinness World Records, which is the authority when it comes to these kinds of things.

Or is it? Shortly after Bobi’s death, experts started raising questions about the Portuguese mastiff’s advanced years. “Not a single one of my veterinary colleagues believe Bobi was actually 31 years old,” veterinarian Danny Chambers told The Guardian. “For the Guinness Book of Records to maintain their credibility and authority in the eyes of the veterinary profession, they really need to publish some irrefutable evidence.”

The reputation of the world’s foremost Irish dry stout turned recordkeeper was on the line here. Someone needed to establish the truth about the oldest dog to ever have lived. That someone—it turned out—was me.

A quick email to Guinness World Records would clear this up, I thought. This is the organization that verified the fastest time to eat a banana with no hands (17.82 seconds) and the longest human tunnel traveled through by a skateboarding dog (30 pairs of legs). For more than 60 years, Guinness World Records has cataloged the stinkiest flowers, widest mouths, and largest chicken nuggets. It had the receipts for the world’s oldest horses, cats, flags, trees, headstanders, llamas (in captivity), customer complaints, working post offices, and road surfaces. Dating the world’s oldest dog would be child’s play.

«

Of course it wasn’t, and the trail includes questions about a conspiracy by Big Dog Food. (That’s not food for big dogs, it’s.. anyway.) An entertaining read. Note in passing: the average confirmed dog age at death is a bit over 11 years.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

A new theory that aims to unite Einstein’s equations and quantum theory suggests we should look at weight as a key to unification. CC-licensed photo by Janet Ramsden 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 10 links for you. Does it scale? I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. Observations and links welcome.

Sellafield nuclear site hacked by groups linked to Russia and China • The Guardian

Anna Isaac and Alex Lawson:

»

The UK’s most hazardous nuclear site, Sellafield, has been hacked into by cyber groups closely linked to Russia and China, the Guardian can reveal.

The astonishing disclosure and its potential effects have been consistently covered up by senior staff at the vast nuclear waste and decommissioning site, the investigation has found.

The Guardian has discovered that the authorities do not know exactly when the IT systems were first compromised. But sources said breaches were first detected as far back as 2015, when experts realised sleeper malware – software that can lurk and be used to spy or attack systems – had been embedded in Sellafield’s computer networks.

It is still not known if the malware has been eradicated. It may mean some of Sellafield’s most sensitive activities, such as moving radioactive waste, monitoring for leaks of dangerous material and checking for fires, have been compromised.

Sources suggest it is likely foreign hackers have accessed the highest echelons of confidential material at the site, which sprawls across 6 sq km (2 sq miles) on the Cumbrian coast and is one of the most hazardous in the world.

The full extent of any data loss and any ongoing risks to systems was made harder to quantify by Sellafield’s failure to alert nuclear regulators for several years, sources said. The revelations have emerged in Nuclear Leaks, a year-long Guardian investigation into cyber hacking, radioactive contamination and toxic workplace culture at Sellafield.

The site has the largest store of plutonium on the planet and is a sprawling rubbish dump for nuclear waste from weapons programmes and decades of atomic power generation.

Guarded by armed police, it also holds emergency planning documents to be used should the UK come under foreign attack or face disaster. Built more than 70 years ago and formerly known as Windscale, it made plutonium for nuclear weapons during the cold war and has taken in radioactive waste from other countries, including Italy and Sweden.

The Guardian can also disclose that Sellafield, which has more than 11,000 staff, was last year placed into a form of “special measures” for consistent failings on cybersecurity, according to sources at the Office for Nuclear Regulation (ONR) and the security services.

«

Well that would be fun if the hackers gained any control of the systems. At a guess: the poor security is the result of ancient computer systems which are almost impossible to lock down, and there’s too little money allocated to computer security. (Storage is, of course, the potential downside of nuclear power.)
unique link to this extract

Federal government investigating multiple hacks of US water utilities • POLITICO

Maggie Miller and John Sakellariadis:

»

The [US] federal government is investigating multiple hacks suspected to have been launched by an Iranian government-linked cyber group against US water facilities that were using Israeli-made technology, according to two individuals familiar with the probes.

One of the breaches made headlines Saturday after the Tehran-linked Cyber Av3ngers group claimed responsibility for hitting a water authority in Pennsylvania. In total, the government is aware of and examining a “single digit” number of facilities that have been affected across the country, according to the two people who were granted anonymity to discuss details that had not yet been made public.

None of the hacks caused significant disruption, according to the individuals, while cyber experts familiar with the Pennsylvania incident say the activity appears designed to stoke fears about using Israeli devices.
Washington has been bracing for increased cyber breaches from Iran since the latest conflict broke out between Israel and the militant group Hamas, which Tehran has long supported. It also comes amid a spate of recent drone and rocket attacks on American troops in the Middle East, conducted by Iranian proxy groups.

Water facilities in general are a particularly vulnerable part of U.S. infrastructure, often due to a lack of funding and personnel for the issue at smaller utilities. The Biden administration has sought to address this problem, including through expanding partnerships with private organizations involved in the water sector.

«

This happened near the end of last week. (On Friday the US put out a formal alert about it: use better passwords!) Lucky that the attacks were of the script kiddie defacement level, rather than actively manipulating controls at water plants.
unique link to this extract

New theory claims to unite Einstein’s gravity with quantum mechanics • Phys.org

»

A radical theory that consistently unifies gravity and quantum mechanics while preserving Einstein’s classical concept of spacetime has been announced in two papers published simultaneously by UCL (University College London) physicists.

Modern physics is founded upon two pillars: quantum theory on the one hand, which governs the smallest particles in the universe, and Einstein’s theory of general relativity on the other, which explains gravity through the bending of spacetime. But these two theories are in contradiction with each other and a reconciliation has remained elusive for over a century.

The prevailing assumption has been that Einstein’s theory of gravity must be modified, or “quantized,” in order to fit within quantum theory. This is the approach of two leading candidates for a quantum theory of gravity, string theory and loop quantum gravity.

But a new theory, developed by Professor Jonathan Oppenheim (UCL Physics & Astronomy) and laid out in a paper in Physical Review X, challenges that consensus and takes an alternative approach by suggesting that spacetime may be classical—that is, not governed by quantum theory at all.

Instead of modifying spacetime, the theory—dubbed a “postquantum theory of classical gravity”—modifies quantum theory and predicts an intrinsic breakdown in predictability that is mediated by spacetime itself. This results in random and violent fluctuations in spacetime that are larger than envisaged under quantum theory, rendering the apparent weight of objects unpredictable if measured precisely enough.

A second paper, published simultaneously in Nature Communications and led by Professor Oppenheim’s former Ph.D. students, looks at some of the consequences of the theory, and proposes an experiment to test it: to measure a mass very precisely to see if its weight appears to fluctuate over time.

«

This comes under the heading of “big if true”. (Also: hard to really understand.)
unique link to this extract

Spotify to lay off 17% of workforce, its third round of job cuts this year • WSJ

Anne Steele:

»

Despite efforts to reduce costs, [founder and CEO Daniel] Ek said Spotify is still spending too much money. The audio streaming company has been squeezed by slower economic growth as well as interest-rate increases that have made it more expensive to borrow, he said.

“The Spotify of tomorrow must be defined by being relentlessly resourceful in the ways we operate, innovate, and tackle problems,” he said in a 1,000-word letter to staff. “Being lean is not just an option but a necessity.”

…Spotify, like other technology companies, grew in size and scope during the pandemic, with its head count nearly doubling over the past three years to more than 8,000 workers, as a result of hiring and acquisitions. As investors have become more focused on profitability than growth, many streaming-focused companies have aggressively cut costs. 

At Spotify that meant scaling back a $1bn bet on podcasting, including through layoffs earlier this year. It continues to back top podcasters Joe Rogan, Alex Cooper and Emma Chamberlain, and stopped making a number of other shows such as Meghan Markle’s “Archetypes.” 

Spotify, which reported a €462m loss in the first nine months of the year, is trying to balance investments in emerging areas such as its growing ad business with the need to become consistently profitable. The company also is focused on its audiobooks offering, which rolled out to subscribers in the US last month. 

Last year, during its first investor day since going public, Ek said he wants Spotify to be the world’s largest audio company and announced ambitious growth targets, such as generating $100bn in revenue by 2030. He said the company plans to reach profitability by 2024.

«

The memo is classic management blurb, but its key points are simply: capital used to be cheap, so we took on lots of people, now it isn’t, goodbye then.
unique link to this extract

Gmail’s AI-powered spam detection is its biggest security upgrade in years • Ars Technica

Ron Amadeo:

»

The latest post on the Google Security blog details a new upgrade to Gmail’s spam filters that Google is calling “one of the largest defense upgrades in recent years.” The upgrade comes in the form of a new text classification system called RETVec (Resilient & Efficient Text Vectorizer). Google says this can help understand “adversarial text manipulations”—these are emails full of special characters, emojis, typos, and other junk characters that previously were legible by humans but not easily understandable by machines. Previously, spam emails full of special characters made it through Gmail’s defenses easily.

…Emails like this have been so difficult to classify becuase, while any spam filter could probably swat down an email that says, “Congratulations! A balance of $1,000 is available for your jackpot account,” that’s not what this email actually says. A big portion of the letters here are “homoglyphs”—by diving into the endless depths of the Unicode standard, you can find obscure characters that look like they’re part of the normal Latin alphabet but actually aren’t.

For instance, the subject “𝐂𝐡𝐞𝐜𝐤_𝐘𝐨𝐮𝐫_𝐀𝐜𝐜𝐨𝐮𝐧𝐭” is weirdly bolded not because it has bolded styling but because it uses Unicode glyphs like the “Mathematical Bold Capital C.” It’s a math symbol that happens to look like the letter “C” to people, but the robot doing spam filtering accurately views it as a math symbol and doesn’t understand the intended English meaning. The closer you look at an email like this, the worse it gets: “C0NGRATULATIONS” has a zero replacing one of the “O” characters, the underlined letters in “Jᴀ̲ᴄ̲ᴋ̲pot” are so strange they don’t even come up in Unicode searches, and a lot of spaces are swapped out for periods or underscores. The result is that a spam filter looks at this hot mess of an email and basically gives up. (I don’t understand why illegible emails default to “inbox” instead of “spam,” but I’m not in charge.)

Google says RETVec is here to save the day: “RETVec is trained to be resilient against character-level manipulations including insertion, deletion, typos, homoglyphs, LEET substitution, and more. The RETVec model is trained on top of a novel character encoder which can encode all UTF-8 characters and words efficiently. Thus, RETVec works out-of-the-box on over 100 languages w

«

unique link to this extract

Federal judge vows to investigate Google for intentionally destroying chats • The Verge

Sean Hollister:

»

Judge James Donato is overseeing Epic v. Google, a case that could determine the future of the Android app store — but testimony in this case may have more repercussions for Google too.

On Friday, Judge Donato vowed to investigate Google for intentionally and systematically suppressing evidence, calling the company’s conduct “a frontal assault on the fair administration of justice.” We were there in the courtroom for his explanation.

“I am going to get to the bottom of who is responsible,” he said, adding he would pursue these issues “on my own, outside of this trial.”

Testimony in the Epic v. Google trial — and in a parallel DOJ antitrust suit against Google in Washington, DC — revealed that Google automatically deleted chat messages between employees, and that employees all the way up to CEO Sundar Pichai intentionally used that to make certain conversations disappear. Pichai, and many other employees, also testified they did not change the auto-delete setting even after they were made aware of their legal obligation to preserve evidence.

And Pichai, among other employees, admitted that they marked documents as legally privileged just to keep them out of other people’s hands.

On November 14th, Pichai told the court that he relied on his legal and compliance teams to instruct him properly, particularly Alphabet chief legal officer Kent Walker — and so Judge Donato hauled Walker into court two days later.

But the judge was not satisfied with Walker’s testimony, either, accusing him of “tap-dancing around.”

«

Google has so many court cases going on that you’d think everything would automatically get filed as potentially liable for discovery.
unique link to this extract

Don’t be fooled: “carbon capture and storage” is no solution to oil and gas emissions • Climate Change News

Laurence Tubiana and Emmanuel Guérin:

»

At the Cop28 climate conference taking place in Dubai, oil and gas producers are counting on carbon capture and storage (CCS) for a social license to keep drilling as usual. Don’t fall for it.

While it can be helpful at the margins, CCS cannot possibly deliver reductions in greenhouse gas emissions on the scale needed to avert climate disaster. This can only happen if the main sources of emissions – fossil fuels – are phased out.

CCS is expected to deliver less than a tenth of the cumulative carbon dioxide emission reductions, over the 2023-2050 period, needed to hold global warming to 1.5C.

In the International Energy Agency net zero emission (NZE) scenario, CCS captures approximately 1.5 billion tons (GT) of CO2 in 2030, and 6 GT by 2050. But very little of that is applied to emissions from fossil fuel production and combustion. It is primarily used to capture CO2 from sectors where emissions are harder and more expensive to reduce, such as cement production or chemicals.

Is the IEA NZE scenario the only way to achieve net-zero emission and limit the temperature increase to 1.5ºC? Certainly not. …scenarios coming out of models are not to be confused with reality. The fossil fuel industry claims it can achieve the same objectives as in the IEA NZE scenario, while producing more oil and gas, by relying more heavily on CCS. Is this true?

…Another IEA scenario, the stated policies scenario, gives the answer. Reaching net-zero carbon emissions in this way would require the capture of 32 GT of CO2 emissions by 2050, including 23 GT through direct air capture (DAC).

At this scale, DAC alone would require 26,000 TWh of electricity to operate, which is more than the total global electricity demand today.

«

A while back I considered writing a book about carbon capture – the climate saviours! Then I looked into it and realised it was all Not Going To Happen.
unique link to this extract

Meet the first Spanish AI model earning up to €10,000 per month • Euronews

Laura Llach:

»

Last summer, Rubén Cruz, her designer and founder of the agency The Clueless, was going through a rough patch because he didn’t have many clients.

“We started analysing how we were working and realised that many projects were being put on hold or cancelled due to problems beyond our control. Often it was the fault of the influencer or model and not due to design issues,” Cruz told Euronews.

So they decided to create their own influencer to use as a model for the brands that approached them.

They created Aitana, an exuberant 25-year-old pink-haired woman from Barcelona whose physical appearance is close to perfection. The virtual model can earn up to €10,000 a month, according to her creator, but the average is around €3,000.

“We did it so that we could make a better living and not be dependent on other people who have egos, who have manias, or who just want to make a lot of money by posing,” said Cruz.

Aitana’s income is quite scattered. She earns just over €1,000 per advert, and has recently become the face of Big, a sports supplement company, and as if that weren’t enough, she uploads photos of herself in lingerie to Fanvue, a platform similar to OnlyFans.

In just a few months, she has managed to gain more than 121,000 followers on Instagram and her photos get thousands of views and reactions. She even receives private messages from celebrities who are unaware that she is not an actual person.

“One day, a well-known Latin American actor texted to ask her out. This actor has about five million followers and some of our team watched his TV series when they were kids,” said Cruz. “He had no idea Aitana didn’t exist.”

«

Plus also they don’t get stroppy. Which turns out to be a big plus.
unique link to this extract

Elon Musk’s xAI will launch ‘Grok’ chatbot this week: what to expect • VentureBeat

Shubham Sharma:

»

The company has already opened signups for the program, but the caveat is that it will be only available to those who have taken the most expensive paid plan of the social networking platform. Those on other plans or using X for free will not get access just yet.

…While many details remain under wraps, the X posts shared by Musk and his team at xAI indicate that Grok will be a ChatGPT competitor, which will be able to engage in back-and-forth conversations when prompted. It is expected to handle all sorts of queries from users, right from mathematical problems to code challenges. 

However, unlike other players in the AI race, Grok will differentiate with exclusive access to X and its realtime, user-generated posts and information. The model behind the assistant has been trained on billions of posts (formerly called tweets) on X and will have access to the most recent data posted on X, enabling it to provide up-to-date information when asked about a current issue. 

«

It’s trained on tweets? This thing is going to make previous racist chatbots look like amateurs.
unique link to this extract

Vendor lock-in is a good thing? HP’s CFO thinks so • The Register

Paul Kunert:

»

Tech vendors – software, hardware, and cloud services – generally avoid terms that suggest they’re perhaps in some way pinning down customers in a strategic sales hold.

But as Marie Myers, chief financial officer at HP, was this week talking to the UBS Global Technology conference, in front of investors, the thrust of the message was geared toward the audience.

“We absolutely see when you move a customer from that pure transactional model … whether it’s Instant Ink, plus adding on that paper, we sort of see a 20% uplift on the value of that customer because you’re locking that person, committing to a longer-term relationship.”

Instant Ink is a subscription in which ink or toner cartridges are dispatched when needed, with customers paying for plans that start at $0.99 and run to $25.99 per month. As of May last year, HP had more than 11 million subscribers to the service. Since then it has banked double-digit percentage figures on the revenues front.

By pre-pandemic 2019, HP had grown weary of third-party cartridge makers stealing its supplies business. It pledged to charge more upfront for certain printer hardware (“rebalance the system profitability, capturing more profit upfront”).

HP also set in motion new subscriptions, and launched Smart Tank hardware filled with a pre-defined amount of ink/toner. These now account for 60% of total shipments.

Myers told the UBS Conference she was “really proud” that HP could “raise the range on our print margins” based on “bold moves and shifting models.”

«

As El Reg points out, it’s not often that CxOs say the quiet part out loud, but they will in front of an investor conference.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

A company in the US is trying to bring back the dodo in a Jurassic Park-style revival. CC-licensed photo by allispossible.org.uk 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. They won’t fly away, though. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. Observations and links welcome.

OpenAI agreed to buy $51m of AI chips from a startup backed by CEO Sam Altman • WIRED

Paresh Dave:

»

Sam Altman was reinstated soon after being fired as OpenAI CEO last month, but still stood to gain had the company continued to develop ChatGPT without him. During Altman’s tenure as CEO, OpenAI signed a letter of intent to spend $51m on AI chips from a startup called Rain AI into which he has also invested personally.

Rain is based less than a mile from OpenAI’s headquarters in San Francisco and is working on a chip it calls a neuromorphic processing unit, or NPU, designed to replicate features of the human brain. OpenAI in 2019 signed a nonbinding agreement to spend $51m on the chips when they became available, according to a copy of the deal and Rain disclosures to investors this year seen by WIRED. Rain told investors Altman had personally invested more than $1m into the company. The letter of intent has not been previously reported.

The investor documents said that Rain could get its first hardware to customers as early as October next year. OpenAI and Rain declined to comment.

OpenAI’s letter of intent with Rain shows how Altman’s web of personal investments can entangle with his duties as OpenAI CEO. His prior position leading startup incubator Y Combinator helped Altman become one of Silicon Valley’s most prominent dealmakers, investing in dozens of startups and acting as a broker between entrepreneurs and the world’s biggest companies. But the distraction and intermingling of his myriad pursuits played some role in his recent firing by OpenAI’s board for uncandid communications, according to people involved in the situation but not authorized to discuss it.

«

Welllll. There’s a lot more detail in this story (Saudi Arabia forced to sell stake in company by US government! Attempt to corner market for AI chips!). But maybe this is the smoking gun that explains what the previous OpenAI board meant when it said, you’ll recall, that Altman “was not consistently candid in his communications with the board, hindering its ability to exercise its responsibilities.”

The story doesn’t say that. But there’s an undercurrent in this that the board didn’t like it.
unique link to this extract

Not so dead as a dodo: “de-extinction” plan to reintroduce bird to Mauritius • CNN

Tom Page:

»

US-based biotechnology and genetic engineering company Colossal Biosciences, which is pursuing the “de-extinction” of multiple species, including the woolly mammoth, has entered a partnership with the Mauritian Wildlife Foundation to find a suitable location for the large flightless birds.

The dodo has been extinct since 1681; a combination of predation by humans and animals introduced by humans led to its downfall, turning it into a textbook case for extinction. But according to the partners, its return to Mauritius could benefit the dodo’s immediate environment and other species.

Colossal first announced its intention to resurrect the dodo in January 2023. Exactly when it will be able to do so remains unclear, but fresh details regarding how it plans to recreate the species have been revealed.

The full genome of the dodo has been sequenced by Beth Shapiro, lead paleogeneticist at Colossal. In addition, the company says it has now sequenced the genome of the solitaire, an extinct relative of the dodo from Rodrigues Island, close to Mauritius, and the Nicobar pigeon, the dodo’s closest living relative, which resides on islands in Southeast Asia spanning the Indian and Pacific Oceans.

Geneticists at Colossal have found cells that act as a precursor for ovaries or testes in the Nicobar pigeon can grow successfully in a chicken embryo. They are now researching to see if these cells (called primordial germ cells, or PGCs) can turn into sperm and eggs.

«

They start with dodos, then pretty soon they’re saying “maybe just a small dinosaur? Ooh, how about this one for Mauritius too?” Also, we’ll need a new aphorism. Dead as a..?
unique link to this extract

Cop28 president says there is ‘no science’ behind demands for phase-out of fossil fuels • The Guardian

Damian Carrington and Ben Stockton:

»

The president of Cop28, Sultan Al Jaber, has claimed there is “no science” indicating that a phase-out of fossil fuels is needed to restrict global heating to 1.5C, the Guardian and the Centre for Climate Reporting can reveal.

Al Jaber also said a phase-out of fossil fuels would not allow sustainable development “unless you want to take the world back into caves”.

The comments were “incredibly concerning” and “verging on climate denial”, scientists said, and they were at odds with the position of the UN secretary general, António Guterres.

Al Jaber made the comments in ill-tempered responses to questions from Mary Robinson, the chair of the Elders group and a former UN special envoy for climate change, during a live online event on 21 November. As well as running Cop28 in Dubai, Al Jaber is also the chief executive of the United Arab Emirates’ state oil company, Adnoc, which many observers see as a serious conflict of interest.

More than 100 countries already support a phase-out of fossil fuels and whether the final Cop28 agreement calls for this or uses weaker language such as “phase-down” is one of the most fiercely fought issues at the summit and may be the key determinant of its success. Deep and rapid cuts are needed to bring fossil fuel emissions to zero and limit fast-worsening climate impacts.

Al Jaber spoke with Robinson at a She Changes Climate event. Robinson said: “We’re in an absolute crisis that is hurting women and children more than anyone … and it’s because we have not yet committed to phasing out fossil fuel. That is the one decision that Cop28 can take and in many ways, because you’re head of Adnoc, you could actually take it with more credibility.”

Al Jaber said: “I accepted to come to this meeting to have a sober and mature conversation. I’m not in any way signing up to any discussion that is alarmist. There is no science out there, or no scenario out there, that says that the phase-out of fossil fuel is what’s going to achieve 1.5ºC.”

«

Perhaps he’s been reading the wrong scenarios, because I’ve certainly seen one which suggests how to stay within 1.5ºC of warming. But it essentially requires stopping use of fossil fuels almost immediately. Again, as Upton Sinclair said: can’t get someone to understand something when their salary depends on their not understanding it. And once again Cop is revealed as a fossil fuel talking shop.
unique link to this extract

1960s chatbot ELIZA beat OpenAI’s GPT-3.5 in a recent Turing test study • Ars Technica

Benj Edwards:

»

In the recent study, listed on arXiv at the end of October, UC San Diego researchers Cameron Jones (a PhD student in Cognitive Science) and Benjamin Bergen (a professor in the university’s Department of Cognitive Science) set up a website called turingtest.live, where they hosted a two-player implementation of the Turing test over the Internet with the goal of seeing how well GPT-4, when prompted different ways, could convince people it was human.

Through the site, human interrogators interacted with various “AI witnesses” representing either other humans or AI models that included the aforementioned GPT-4, GPT-3.5, and ELIZA, a rules-based conversational program from the 1960s. “The two participants in human matches were randomly assigned to the interrogator and witness roles,” write the researchers. “Witnesses were instructed to convince the interrogator that they were human. Players matched with AI models were always interrogators.”

The experiment involved 652 participants who completed a total of 1,810 sessions, of which 1,405 games were analyzed after excluding certain scenarios like repeated AI games (leading to the expectation of AI model interactions when other humans weren’t online) or personal acquaintance between participants and witnesses, who were sometimes sitting in the same room.

Surprisingly, ELIZA, developed in the mid-1960s by computer scientist Joseph Weizenbaum at MIT, scored relatively well during the study, achieving a success rate of 27%. GPT-3.5, depending on the prompt, scored a 14% success rate, below ELIZA. GPT-4 achieved a success rate of 41%, second only to actual humans.

«

Good old Eliza, still going strong all these years on.
unique link to this extract

It’s official: Evernote will restrict free users to 50 notes • TechCrunch

Ivan Mehta:

»

Days after Evernote started testing with limited users a free plan with access to only one notebook and 50 notes, the company has now made this its new default free plan. The notetaking app said that this change will be applicable for all new and existing free users starting December 4.

In a post on its blog, Evernote specified that users can delete content from their notebooks to add other content within the limit.

“From December 4, the Evernote Free experience has changed. Going forward, new and existing Free users will have a maximum of fifty notes and one notebook per account. These limits refer to the number of notes and notebooks a user can have in their account at one time: you can always delete unwanted content to remain below the threshold,” the company, owned by Milan-based Bending Spoons, said.

Users with more than 50 notes in their existing free accounts will be able to export additional notes and notebooks. Evernote mentioned on its blog that these restrictions will reflect on its compare plans page on December 4, but didn’t specify if limits or pricing of other plans are also changing.

Earlier this week, Evernote confirmed to TechCrunch on its website that the new limited-free plan was part of a test with “less than 1% of its free users.” The test was trying to get people to pay the higher limit plans, which are priced at $14.99 and $17.99 per month.

The company said that most free users fall below the newly set limit. However, Evernote acknowledged that this change might push customers towards “reconsidering” their “relationship with Evernote.”

«

Bought a year ago, laid off 129 people in February, “unprofitable for years”. Started in February 2008 but doubtful it ever made money. Another ZIRP casualty. Bending Spoons also canned the entire staff of filmmaking app Filmic on Friday.
unique link to this extract

Thames Water told by auditors it could run out of money by April • The Guardian

Miles Brignall:

»

The parent company of Thames Water has been warned by its auditors that it could run out of money by April if shareholders do not inject more cash into the debt-laden firm.

In accounts signed off in July and published on the Companies House website last week, PricewaterhouseCoopers said there was “material uncertainty” about whether the main company behind the water supplier can continue as a going concern.

The disclosure was made in the 2022-23 accounts of Kemble Water Holdings, the company at the top of Thames Water’s byzantine ownership structure.

PwC made its assertion after noting that there were no firm arrangements in place to refinance a £190m loan at one of its subsidiary companies.

Thames Water is expected to face further scrutiny over its debt levels when it issues its results on Tuesday, and a possible investigation into whether it misled MPs earlier this year.

In June, it emerged that contingency plans for the collapse of Thames Water were being drawn up by the UK government amid fears that Britain’s biggest water company would not survive because of its huge debt pile.

Sir Robert Goodwill, chair of the environment, food and rural affairs select committee, said it was considering a fresh investigation after the Financial Times reported that Thames Water had originally presented a loan from its shareholders to its parent as new equity funding.

«

This would be quite an event. Water companies in the UK were privatised in 1989, and none has collapsed into bankruptcy or similar problems. Until now. Its debts in June were about 80% of its value (about £17.5bn). The current government won’t like having to take that onto its books, so the question is: who will be left with the hot (wet) potato?
unique link to this extract

Switzerland put vertical solar panels on a roadside retaining wall • Electrek

Michelle Lewis:

»

The canton of Appenzell Ausserhoden in northeastern Switzerland is aiming to generate at least 40% of its electricity from renewables by 2035. So, it exercised a little creativity and covered a roadside retaining wall with 756 glass-glass solar panels.

The panels have an output of 325 kW and an energy yield of around 230,000 kWh annually. This is equivalent to the consumption of about 52 Swiss households. The energy will be fed into the grid of energy supplier St. Gallisch-Appenzellische Kraftwerke, and the canton will get a feed-in tariff in return.

…K2 Systems says that “especially in the winter months (when consumption and dependence on foreign electricity imports are at their highest), the vertically aligned modules will achieve a very good electricity yield.”

«

The calculation (230,000 kWh/yr / 245.7 kW / 365 day/yr) works out to 2h33m average per day; ironically, more during the winter because the sun will be lower.
unique link to this extract

Amazon deal for iRobot may restrict competition, European Commission says • WSJ

Ben Glickman:

»

The EC said that the deal may restrict competition in the making of robot vacuum cleaners and could allow Amazon to fortify its position as an online marketplace services provider.

The commission said Amazon may have the “ability and the incentive” to foreclose iRobot’s rivals by preventing them from selling on Amazon’s platform.

“We continue to work through the process with the European Commission and are focused on addressing its questions and any identified concerns at this stage,” an Amazon spokesperson said in response to the release.

The spokesperson said iRobot [which makes the Roomba] faces “intense competition” in the market for vacuum cleaner products and that the company believes it can invest in iRobot while lowering prices for consumers.

The acquisition by Amazon, announced in August 2022, was cleared by U.K. regulators in June, but faces an ongoing investigation by the U.S. Federal Trade Commission.

The EC has until February 14 to make a final decision on the deal.

«

Amazon signed the deal back in August 2022 for $1.7bn, all cash. And it’s still bumping back and forth into regulatory barriers.
unique link to this extract

iPhone glitch photo explained. It is NOT photoshop. • Threads

So there was a big kerfuffle over the weekend about a photo of a British comedian/actress in a wedding dress which showed her with her arms in three different poses.. in the same photo. It became this year’s blue/silver dress meme. And here a guy called Faruk explains it in a short video. You need the video really.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

An AI system at Google DeepMind has predicted the structure of nearly half a million novel crystals for future materials technologies. CC-licensed photo by Francisco Anzola 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.

It’s Friday, so there’s another post due at the Social Warming Substack at about 0845 UK time.

A selection of 9 links for you. Use them wisely. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. Observations and links welcome.

Biden administration stops warning some social platforms of foreign meddling • The Washington Post

Naomi Nix and Cat Zakrzewski:

»

Meta no longer receives notifications of global influence campaigns from the Biden administration, halting a longtime partnership between the federal government and the world’s largest social media company, senior security officials said Wednesday. Federal agencies have also stopped communicating about political disinformation with Pinterest, according to the company.

The developments underscore the far-reaching impact of a conservative legal campaign against initiatives established to avoid a repeat of the 2016 election, when Russia manipulated social media in an attempt to sow chaos and swing the vote for Donald Trump. Republican lawmakers even have proposed cutting funding for combatting foreign disinformation and subpoenaed government agencies, including the State Department’s Global Engagement Center, which counters foreign propaganda.

For months, researchers in government and academia have warned that a barrage of lawsuits, congressional demands and online attacks are having a chilling effect on programs intended to combat health and election misinformation. But the shift in communications about foreign meddling signals how ongoing litigation and Republican probes in Congress are unwinding efforts once viewed as critical to protecting U.S. national security interests.

Ben Nimmo, chief of global threat intelligence for Meta, said government officials stopped communicating foreign election interference threats to the company in July.

«

This could be seen as Definitely Bad. But: X/Twitter is now an absolute morass of misinformation, with a falling user base, so Russia isn’t going to have much joy there. (Nor anyone else.) Facebook, well, it’s where the boomers hang out – nobody much else – and they’re always susceptible to any old crap, if they’re susceptible at all. Instagram, folk don’t get taken in by that. TikTok, well, that’s China, and doesn’t really listen to the US. What else is there? (Thanks G for the link.)
unique link to this extract

Meta sues FTC, hoping to block ban on monetizing kids’ Facebook data • Ars Technica

Jon Brodkin:

»

Meta sued the Federal Trade Commission on Wednesday in a lawsuit that challenges the FTC’s authority to impose new privacy obligations on the social media firm.

The complaint stems from the FTC’s May 2023 allegation that Meta-owned Facebook violated a 2020 privacy settlement and the Children’s Online Privacy Protection Act. The FTC proposed changes to the 2020 privacy order that would, among other things, prohibit Facebook from monetizing data it collects from users under 18.

Meta’s lawsuit against the FTC challenges what it calls “the structurally unconstitutional authority exercised by the FTC through its Commissioners in an administrative reopening proceeding against Meta.” It was filed against the FTC, Chair Lina Khan, and other commissioners in US District Court for the District of Columbia. Meta is seeking a preliminary injunction to stop the FTC proceeding, pending resolution of the lawsuit.

Meta argues that in the FTC’s administrative proceedings, “the Commission has a dual role as prosecutor and judge in violation of the Due Process Clause.” Meta asked the court to “declare that certain fundamental aspects of the Commission’s structure violate the US Constitution, and that these violations render unlawful the FTC Proceeding against Meta.”

Meta says it should have a right to a trial by jury and that “Congress unconstitutionally has delegated to the FTC the power to assign disputes to administrative adjudication rather than litigating them before an Article III court.” The FTC should not be allowed to “unilaterally modify the terms” of the 2020 settlement, Meta said.

The FTC action “would dictate how and when Meta can design its products,” the lawsuit said.

«

Well, yes, that’s sort of the point of the FTC order, Meta. The core of Meta’s suit is that the FTC is itself unconstitutional because its commissioners can act as prosecutor, judge and jury in a case. (Explained here; others have tried the same.)
unique link to this extract

Nearly 100,000 Gaza buildings may be damaged, satellite images show • BBC News

Dominic Bailey, Erwan Rivault and Daniele Palumbo:

»

New satellite images commissioned by the BBC reveal the extent of destruction across northern Gaza, before the start of the temporary ceasefire between Israel and Hamas.

The satellite images were taken last Thursday, just before the suspension of hostilities came into force, following weeks of Israeli air strikes and on-the-ground fighting. Separate satellite data analysis also provides a snapshot of the destruction across the whole of Gaza.

Drone footage and verified video also show buildings and entire neighbourhoods reduced to rubble. While northern Gaza has been the focus of the Israeli ground offensive and has borne the brunt of the destruction, widespread damage extends across the entire strip.

Israel says northern Gaza, which includes the major urban centre of Gaza City, was a “centre of gravity of Hamas”, the group behind the deadly 7 October attacks on Israel. Israel says its bombing campaign has successfully targeted Hamas commanders and fighters and accuses the group of embedding itself in civilian areas.

Satellite data analysis suggests that almost 98,000 buildings across the whole Gaza Strip may have suffered damage, with most of it concentrated in the north – as shown in the above map.

«

On the technology front, what’s amazing here is that a media company can commission satellite images and then get them analysed in less than a week to derive information like this.

On the human front, it’s utterly shocking. Yes, Hamas’s murder of more than a thousand Israelis on October 7 was evil and deserved retribution. But even at the last election, in 2006, it only got 44% of the vote, and much of Gaza’s population on October 7 had never voted at all. The idea that there were Hamas commanders and fighters in every single one of those levelled properties is for the birds.
unique link to this extract

‘A mass assassination factory’: Inside Israel’s calculated bombing of Gaza • 972

Yuval Abraham:

»

The Israeli army’s expanded authorization for bombing non-military targets, the loosening of constraints regarding expected civilian casualties, and the use of an artificial intelligence system to generate more potential targets than ever before, appear to have contributed to the destructive nature of the initial stages of Israel’s current war on the Gaza Strip, an investigation by +972 Magazine and Local Call reveals. These factors, as described by current and former Israeli intelligence members, have likely played a role in producing what has been one of the deadliest military campaigns against Palestinians since the Nakba of 1948.

The investigation by +972 and Local Call is based on conversations with seven current and former members of Israel’s intelligence community — including military intelligence and air force personnel who were involved in Israeli operations in the besieged Strip — in addition to Palestinian testimonies, data, and documentation from the Gaza Strip, and official statements by the IDF Spokesperson and other Israeli state institutions.

…According to the investigation, another reason for the large number of targets, and the extensive harm to civilian life in Gaza, is the widespread use of a system called “Habsora” (“The Gospel”), which is largely built on artificial intelligence and can “generate” targets almost automatically at a rate that far exceeds what was previously possible. This AI system, as described by a former intelligence officer, essentially facilitates a “mass assassination factory.”

According to the sources, the increasing use of AI-based systems like Habsora allows the army to carry out strikes on residential homes where a single Hamas member lives on a massive scale, even those who are junior Hamas operatives. Yet testimonies of Palestinians in Gaza suggest that since October 7, the army has also attacked many private residences where there was no known or apparent member of Hamas or any other militant group residing. Such strikes, sources confirmed to +972 and Local Call, can knowingly kill entire families in the process.

«

The Israel Defence Force (IDF) has been very cagey about talking about its AI system, and what it does, and how it’s used. This is the first report I’ve seen which describes what its role could be.

About 972 Magazine: “+972 Magazine is an independent, online, nonprofit magazine run by a group of Palestinian and Israeli journalists.” The name comes from the dialling code, which reaches numbers in Israel and Palestine.
unique link to this extract

What a bloody San Francisco street brawl tells us about the age of citizen surveillance • WIRED

Lauren Smiley:

»

One day in July, I spotted a camera on the front of a tidy house on Magnolia—the place whose driveway had been the site of the 2021 attack where a man was sprayed in the face for five seconds. I walked up to the open garage and asked the man inside if I could talk to him about the Ring cam footage that had emerged of the incident. He said he’d rather not, that he was thinking he would be subpoenaed. [Assault-accused homeless man Garret] Doty’s trial was coming up—it’s now set for mid-November—and the man said he had to get an attorney himself. He hadn’t authorized the release of his footage to the media, he explained, and he wasn’t happy that it wound up on national TV.

As I became yet another Marina [Bay area in San Francisco] surveillant [by recording some video of a homeless man], I thought of what one regular chronicler of the homeless in the neighborhood had told me—that he doubted all the careful surveillance had added up to much other than “On my phone I have a bunch of stupid photos now.” San Francisco’s police chief has called the city’s bonanza of surveillance footage a “golden” tool for solving crimes, but it can just as easily set people free: The public defender’s office is one of the main requesters of footage from the Tenderloin’s camera network. One defense attorney in the city, Elizabeth Hilton, told me that in many of her cases the trove of San Francisco video evidence ends up helping the accused, contradicting victims’ and witnesses’ accounts of what went down.

…As I kept my eye on the Marina, I couldn’t stop thinking about the guy I’d met in his garage that day in July, bewildered that his Ring footage had ended up on the national news, that this little piece of hardware had unleashed something bigger than he’d ever intended. His reaction struck me as genuine and understandable—what most people would feel in his position. Yet it also seemed quaintly naive, a reminder that those engaged in citizen surveillance in 2023 still don’t totally get what it means to have a camera watching the street.

«

An absorbing long read about the incredible tensions between residents and homeless on San Francisco’s streets.
unique link to this extract

Message received • The Rebooting

Brian Morrissey:

»

as publishers know well, the ad business is often illogical and possibly more mercurial than [Elon] Musk. It can gyrate wildly. And publishers have little control of their distribution, putting themselves in a double bind. I don’t think it took very long for Musk to figure this out. The parts of the ad business that the engineering mind like are in the direct marketing part that’s driven by math and quantifiable results. Ironically, he bought the one scaled tech platform that is more reliant on the other part of advertising, the part that’s more about branding, since Twitter doesn’t have the kind of intent signals other platforms have. 

Big brands care about flaccid concepts like brand safety and alignment. And sure, they’re hypocrites – aren’t we all? – but you only get a pass if you perform. Twitter has never been a must-buy on the basis of performance. Musk made sure to specifically call out Bob Iger in the audience during his GFY, since Disney is among the advertisers who have paused spending due to Musk’s latest controversial post that strayed into the territory of Protocols of the Elders of Zion. The truth is Disney’s business is not going to be hurt by not advertising on X. 

I believe he knows this. The subscription model is the only path for X, at least as the core of its business if it is truly going to maximize free speech. I believe that’s why he shrunk the size of the organization. The ad business will be niche, more digital [junk catalogue] SkyMall than vying for the TV ad budgets [CEO] Linda Yaccarino has been adept at grabbing. I’m unclear the point of her role at the company with that focus.

«

Yaccarino is now being referred to online as the “CNO”, or CENO – CEO In Name Only. Certainly can’t have been her most comfortable moment sitting in the front row as Musk torpedoed her role. Shall we start a resignation clock?
unique link to this extract

Millions of new materials discovered with deep learning • Google DeepMind

Amil Merchant and Ekin Dogus Cubuk:

»

Modern technologies from computer chips and batteries to solar panels rely on inorganic crystals. To enable new technologies, crystals must be stable otherwise they can decompose, and behind each new, stable crystal can be months of painstaking experimentation.

Today, in a paper published in Nature, we share the discovery of 2.2 million new crystals – equivalent to nearly 800 years’ worth of knowledge. We introduce Graph Networks for Materials Exploration (GNoME), our new deep learning tool that dramatically increases the speed and efficiency of discovery by predicting the stability of new materials.

With GNoME, we’ve multiplied the number of technologically viable materials known to humanity. Of its 2.2 million predictions, 380,000 are the most stable, making them promising candidates for experimental synthesis. Among these candidates are materials that have the potential to develop future transformative technologies ranging from superconductors, powering supercomputers, and next-generation batteries to boost the efficiency of electric vehicles.

GNoME shows the potential of using AI to discover and develop new materials at scale. External researchers in labs around the world have independently created 736 of these new structures experimentally in concurrent work. In partnership with Google DeepMind, a team of researchers at the Lawrence Berkeley National Laboratory has also published a second paper in Nature that shows how our AI predictions can be leveraged for autonomous material synthesis.

«

AI for good. (Though I’m sure someone will come along presently and suggest it can be used by terrorists to do.. something bad.)

In passing, you can tell this didn’t go through a subeditor. The second sentence of the first paragraph should read more like “To enable new technologies, crystals must be stable, or else they can decompose. Finding a new, stable crystal can take months of painstaking experimentation.”
unique link to this extract

China’s wind and solar surge threatens Australian coal exports • Australian Financial Review

Ben Potter:

»

Two new reports, from global energy consultancy Wood Mackenzie and Sydney-based Climate Energy Finance, show that China is building wind and solar at twice the rate of the US and Europe combined, and also leading the way with huge energy storage installations.

The startling acceleration comes as world leaders and officials prepare to descend on Dubai for the United Nations climate change conference (COP28) and wrangle over phasing out fossil fuels and supporting developing countries’ clean energy projects.

“China’s demand for Australian exports of thermal and coking coal is set to decline structurally over the longer term due to the greening of China’s power sector and economy,” said Climate Energy Finance director Tim Buckley. “The report recommends that to minimise economic risk, Australia urgently comprehends and responds at speed to align with China’s massive investment pivot.”

China’s accelerating clean energy shift challenges Australia, which has planned for continued exports of coal and gas while cautiously backing US-led efforts to wean the West off China’s clean energy goods and commodities and build up alternative sources of supply.

Wood Mackenzie expects China to continue to command 80% of the global supply chain for solar energy until at least 2026. It says in a new report, “How China became the global renewables leader”, that the giant economy is on track to build 230 gigawatts of wind and solar power this year at a cost of $US140bn ($210bn), compared with 75GW for Europe and 40GW for the US. A gigawatt is the size of a small coal-fired power station.

«

So much implied judgement in this piece. “Startling acceleration” – not really; China loses out from runaway climate change too. And as for China getting ahead on making renewables: Australia could have grasped that opportunity decades ago (it has lots of empty space for wind and solar). It chose not to. Cry me a river.
unique link to this extract

Inside Foxconn’s India iPhone factory expansion • Rest of World

Viola Zhou and Nilesh Christopher:

»

Foxconn began manufacturing iPhones at Sunguvarchatram in 2019, starting with the iPhone XR. At that point, the model was more than a year old. When Li [Hai, from Foxconn China] arrived at Foxconn Sunguvarchatram in early 2023, the factory was making iPhone 14s, for which production in India had begun two months after its launch. This year, the goal was to have a shipment of made-in-India iPhone 15s ready to go as soon as the model was announced.

The iPhone plant is part of a sprawling 60-hectare campus where Foxconn also makes phones for other brands. About 35,000 employees go to work inside half a dozen white, three-story factory buildings. Li may as well have been walking back into the Chinese plant he was familiar with at home: the same advanced equipment, the same rows of tables with workers repeating tasks thousands of times a day, the same final product. But there was one obvious difference. Unlike in China, the assembly line was staffed almost exclusively by young women.

When electronics manufacturing took off in China in the 1980s, rural women who had just begun moving to the cities made up the majority of the factory workforce. They didn’t have many other options. Managers at companies like Foxconn preferred to hire women because they believed them to be more obedient, Jenny Chan, a sociologist at Hong Kong Polytechnic University who studies labor issues at Foxconn, told Rest of World.

Over the past 30 years, that’s changed. Today, most of China’s iPhone workers are men; women have moved into less arduous service sector jobs. But in India, Foxconn and other electronics manufacturers are once again recruiting from a female workforce beginning to migrate for better jobs.

…Foxconn also had to find a workaround for employing married women. The company typically requires workers to pass through metal detectors when entering and exiting its factories in order to prevent leaks about upcoming products, according to reports. But in India, married women wear a mangalsutra, a metal pendant; and a metti, a metal toe ring. These workers are searched manually and have their jewelry logged in a notebook.

«

unique link to this extract

Errata, corrigenda and ai no corrida: none notified