Start Up No.2743: Russia recruits UK vandals online, AI Odyssey film loses the plot, Samsung hurries to unfold new foldable, and more


Given multiple albums’ track listings, could you build a chart-topping ensemble? A new online game lets you try. CC-licensed photo by Tom Collins on Flickr.

You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.


A selection of 10 links for you. Spinning around. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.


Russian spies “recruiting UK teenagers” to carry out acts of violence and sabotage • The i paper

Lizzie Dearden:

»

Russian spies are targeting violence-obsessed UK teenagers and offering them money, guns and training on making explosives and poisons, a new investigation suggests.

It marks a worrying new escalation in Russian-linked use of “disposable” agents in the West to carry out acts of violence, sabotage and intelligence-gathering.

Analysis of open source material has found online networks have been used to attempt to recruit young people in Britain willing to carry out acts of violence. At least two incidents have already taken place in Britain in connection with one of the Russia-linked groups, The i Paper can reveal.

…The online groups where the Russian-linked actors are targeting youngsters are so-called “Com networks”. Short for “community” networks or “The Com”, these online groups, whose members are predominantly teenagers and young adults, collaborate, compete, and gain status by inflicting or recording extreme harm.

Law enforcement agencies across the UK, US and Europe have already raised the alarm about these chat groups where violence-obsessed young people incite each other to commit ever-more extreme acts, ranging from vandalism to animal torture, sexual abuse and murder.

The analysis, shared exclusively with The i Paper, was carried out by the Institute for Strategic Dialogue (ISD), an organisation that monitors extremism, with a branch based in London.

The research has revealed “multiple indicators suggesting possible Russian state-linked support” for a section of the “Com network” movement, using tactics that “mirror established Russian methods for recruiting ‘disposable agents’ for violence abroad”.

At least two incidents have taken place in the UK in connection with one of the Russia-linked groups, which The i Paper is not naming to avoid signposting its online presence. Both were acts of vandalism filmed by a person who declared their allegiance by holding up a piece of paper displaying the name of the organisation, its leaders and allies.

In one video shared on the group’s official channel on the Telegram app, a person wearing black gloves used a brick to smash a car’s front and rear windscreens before jumping onto the bonnet and stamping on the vehicle. The i Paper has established that the incident took place in Bedford on 14 April, and targeted a car used by carers working with elderly people to visit their clients. They were not aware the video had been recorded.

«

unique link to this extract


The 2.5-hour AI-generated Odyssey movie is 2.5 hours too long • The Verge

Andrew Webster:

»

Christopher Nolan’s engrossing take on The Odyssey dominated at the box office and spurred a newfound interest in classic literature among filmgoers. But a new retelling of the story made entirely with AI is so bad that it might just make viewers hate the original tale altogether.

The new film, called Odysseus: The Fall, comes via AI firm Fountain 0 and is written and directed by company cofounder Ash Koosha, who also lent his likeness to the title character and created all of the music. (The film’s credits are very short.) The studio’s only previous work is an AI film called Dreams of Violets that made it into Tribeca earlier this year. The Fall is, on its face, a much more ambitious project, billed as “the first completely AI generated film produced at the level of a Hollywood blockbuster.” In reality, it’s nothing close to “the level of a Hollywood blockbuster.” That may not necessarily be surprising, but after watching the entire 2.5-hour production I nonetheless found myself constantly surprised by just how much worse things could get. The Fall is the opposite of a glimpse at the future of AI filmmaking — it’ll make you wonder if there’s really any future in it at all.

…The Fall is plagued by two main issues. The first is consistency. The world of the film is constantly morphing, to the point that it’s incredibly distracting. The height of the cyclops that traps Odysseus and his men changes from shot to shot, waves move in the opposite direction of the wind, mouths don’t match the words spoken (and sometimes they move without producing a sound at all), and oars bend and twist as sailors row. In one scene a huge fleet of boats is racing through the ocean and one of them is inexplicably moving sideways. The AI-produced voice-overs can be incomprehensible, and at various points the names of both Odysseus and Zeus are pronounced wrong.

If it were just the odd hiccup it might be understandable, but there are so many faults in The Fall that it can be hard to focus on anything else. I watched the movie with my 13-year-old kid and it basically became a game of finding the weirdest mistakes.

«

Headline of the year. One could say that this is the worst that AI filmmaking will ever be. But on the other hand, if the capability to actually direct and correct it doesn’t improve, then it will never rise above that.
unique link to this extract


How to play Debutle: build a chart-topping album

»

Spin real albums. Pick one song from each. Build a 12-track debut and see where it charts.

Debutle is a free music draft game in the browser. No install, no login. You build a Franken-album from a curated pool of real records, then the game projects a chart debut – anywhere from missing the chart to a #1 week. Play free whenever you want, or take on the Daily Debut, one shared puzzle a day.

«

It’s a fun game which goes back a fair distance (the 1990s) for its albums. How was it built?

»

Every song carries a rating from 55 to 99. Those numbers are computed, not assigned by feel. The inputs are verified Spotify playcounts, YouTube view counts, and hand-curated chart peaks and RIAA certifications, combined on a calibrated ladder and recomputed from scratch whenever the catalog changes. Six automated audits have to pass before a rating ships, covering things like duplicate track IDs, interludes inheriting a hit’s streaming numbers, and ratings drifting out of step with a fresh recompute.

The interesting decisions are the debatable ones. Streams are discounted by how recently a song was released, so a 2024 release is not measured against a 1996 release on raw plays.

«

The creator(s) don’t give a name.
unique link to this extract


Iran war has the US military on the brink of a “breakdown” • The Intercept

Nick Turse:

»

The US military is straining under the pressures of the Iran war, according to two US officials familiar with operations in the Middle East. One cautioned that this stress could lead to a “breakdown” in the military’s ability to effectively continue a war that President Donald Trump said would end in early April.

As the conflict has engulfed the region, the military has grappled with a growing list of challenges: Iranian attacks on US bases and warships in the Middle East, depleted stockpiles of defensive munitions, increasing casualties, maintenance issues on ships, lengthy troop deployments, complicated logistics, flagging morale and a host of other issues that have taken a heavy toll on troops, damaged military readiness, and a strained Pentagon budget — contradicting months of rosy pronouncements by the Trump administration.

Both officials said that Iran is facing extreme economic pressure, according to intelligence reports. The first official characterized recent Iranian attacks on warships with ballistic missiles as a “go for broke” strategy to increase economic pressures on America ahead of the US midterms. That official said it was unclear which country would “blink first.”

Trump suggested on Sunday that the US could “stay and keep the oil” in Iran, like the neo-colonial resource grab the administration is carrying out in Venezuela. The second official said that a US occupation of Iran would be “almost impossible” and that the idea was “nonsensical obviously.”

Iran’s ability to overwhelm US air defenses in the Middle East using attack drones and advanced ballistic missiles, as previously reported by The Intercept, has left the US military in a precarious state. The Trump administration failed to appreciate Iran’s military prowess, said the first official. The other pointed to a “complete lack of planning” for the war.

«

Nothing Trump says about Iran should be believed (or else the war has been won more than 100 times). It’s very difficult to get any picture of what is happening inside Iran; communications from within seem to have shut down despite the internet still working.
unique link to this extract


Trump administration orders Kalshi to scrap AI price tracker over national security concerns • Semafor

Reed Albergotti and Liz Hoffman:

»

The US Commerce Department last month ordered Kalshi to take down one of its products tracking the price of AI compute, the crucial power from data centers that’s driving the artificial intelligence boom.

Commerce officials cited national security concerns when they told Kalshi to unpublish its AI-compute future curve, said people familiar with the matter. The product pulls together data from several markets that allow users to bet on the cost to rent Nvidia chips to create an overall picture of where AI compute costs are heading. Kalshi quietly complied, though many underlying markets remain open for trading.

Separately, Commerce also pushed the Commodity Futures Trading Commission, which oversees prediction and futures markets, to effectively freeze approval of new compute contracts for 60 days, some of the people said. It was a rare intervention that surprised an industry used to White House enthusiasm for both AI and financial-market innovation.

Kalshi declined to comment, while a Commerce spokesperson said the department “has never once asked Kalshi to take down this market or any other markets.” The CFTC didn’t respond to a request for comment. “This story is false,” a Commerce spokesman said.

It’s unclear why Commerce is worried about the nascent market, which aims to do for AI computing what oil futures do for crude — let buyers and sellers of compute lock in prices, and give traders a way to bet on where those prices go.

«

Besides polluting more, the Trump administration is removing more and more data from the public realm, no matter who it’s published by. At least the (plentiful) food warnings aren’t being removed. For now.
unique link to this extract


How a fringe AI movement convinced Washington the end is near • The Washington Post

Ian Duncan and Nitasha Tiku:

»

When Sen. Bernie Sanders traveled to the Bay Area this year to question the impact of artificial intelligence, he sought out different perspectives. The result was a “handful” of meetings, including with tech executives, the senator’s longtime political strategist Faiz Shakir said. “If they wanted to be on camera, he said, ‘Great, let’s do this publicly!’” Shakir said.

But when Sanders posted a video from his trip in March, it showed him talking with only one controversial faction: people convinced that advanced AI is likely to lead to human extinction.

“Once they don’t need the humans, the humans are discarded,” Eliezer Yudkowsky, who has spent two decades warning about AI eradicating people, said in the video. Sanders, sitting with his chief of staff, appeared shocked. “That’s scary,” he said.

The meeting between Sanders and advocates of the idea that powerful AI will threaten humanity’s future occurred amid a years-long drive by activists to forge connections in Congress and elevate their concerns. In recent days, that work appeared to have paid off.

After Anthropic researcher Jacob Coxon resigned last week and said the people building AI feared it could “kill us all by the end of the decade,” existential debate gripped Washington. Lawmakers from both major parties began to echo warnings that AI could snuff out humans.

Connor Leahy, head of the U.S. arm of ControlAI, an organization seeking to ban artificial superintelligence, said the shift has been years in the making. His team spoke with 200 congressional offices in the past 18 months, he said, working to spread what they see as a crucial warning.

“The way a lot of these kind of politics tends to work is that something is completely impossible until it’s very suddenly not,” Leahy said. “Things can change very quickly. And we’ve been seeing a lot of very-suddenly-not’s lately.”

«

Much easier to get people to listen if you say this new technology is scary and dangerous than if you say that nah, it’s nothing to worry about. And Bernie Sanders might be particularly predisposed to listening and repeating it.
unique link to this extract


A coffee shop owner used AI for a menu. Cue the angry DMs • Business Insider

Katie Notopoulos:

»

Penny’s Coffee Shop would switch seasonal drinks every few months, and Endeladze would photograph the new items herself, and it would end up being a full-day event. “You have to time it so the foam doesn’t deflate,” Endeladze explained to Business Insider.

Then, one of the more artistically gifted baristas would create a chalkboard sign for the new drink menus or promotional specials. Over the years, there had been several employees skilled with lettering and design, and the duty had always rotated among them, always done in-house.

But this fall, the aesthetically inclined barista who most recently chalked the signs had moved out of town, Endelaydsze said, and there was no one on staff at the moment who had the skill set to do a chalkboard. So Endelaydsze decided to get a poster printed up instead. She considered doing what she has done in the past: taking her own photos and editing them together with Canva. But with a newborn baby, she figured she could save herself some time and do it with AI.

“I felt like a genius, and made posters using AI and frame them,” she told me. ‘Not knowing how much damage it would cause my coffee shop.”

She used ChatGPT to make an image of the menu, and got the poster professionally printed up for around $150. She also bought a plastic poster holder that cost around $200.

Then, she posted a photo of the new menu sign to Instagram. Very soon, the negative comments started rolling in:

“AI Slop.”

“Really a shame honestly.”

«

It got worse – which wasn’t much fun for the mother of a newborn trying to keep up with the barrage of hate. We find new things to be outraged about all the time, but the internet keeps making outrage easy.
unique link to this extract


Samsung reportedly rushing new trifold phone in response to iPhone Duo • 9to5 Mac

Ben Lovejoy:

»

It was already clear that Samsung was worried about the impact of the iPhone Duo on its foldable phone range, and there’s now further evidence of the Korean company’s concern.

A blogger who aggregates reports from Chinese social media says there are signs that the company is bringing forward the launch of its Galaxy Z Trifold 2, despite the seeming failure of its predecessor.

Given Samsung’s poor track record on the robustness and reliability of its foldable phones, it’s not surprising that it felt threatened by the build quality of the iPhone Duo. Even before the new model was officially unveiled, the company was criticizing the reported weight of the device.

Once it was announced, it was clear that Apple had addressed a key weakness of Samsung’s foldables: vulnerability to damage caused by the ingress of dust and dirt. While Samsung’s latest folding model is only rated to IP48, leaving it vulnerable to particles as large as a grain of sand, the iPhone Duo is rated to IP68, meaning fully dustproof.

Samsung now appears to be hoping that if one fold isn’t enough, perhaps two will do the trick.

«

The first Samsung TriFold cost $2,899 – below cost price, apparently – and wasn’t a roaring success by all accounts. The idea that Apple just makes something and then everyone goes gaga for it is only half true; Apple has a tendency to get the tiny important details right (such as beating dust in the hinge, assuming it has). Though even that couldn’t save the Vision Pro. Some concepts are beyond saving.
unique link to this extract


Two Chinese firms used Binance to launder Iranian oil money, US says • The New York Times

David Yaffe-Bellany and Michael Forsythe:

»

Two Chinese companies used accounts on the cryptocurrency exchange Binance to launder illegal proceeds from the sale of Iranian oil, the Justice Department said in court papers.

In a civil forfeiture claim filed in federal court in Manhattan on Monday, prosecutors said that the two companies had moved funds intended to finance the Iranian military, including the Islamic Revolutionary Guards Corps, which the United States has classified as a terrorist organization. Binance was not charged with any wrongdoing.

A group of crypto wallets linked to Iran received more than $1.5bn of the illegal proceeds, the complaint said, though it wasn’t clear exactly how much of that money had moved through Binance, the world’s largest crypto trading platform. The Justice Department plans to seize $61m, currently held in a set of accounts overseen by Tether, another crypto firm, the complaint said.

The flow of illicit funds through Binance was flagged internally last year by a group of the exchange’s compliance officials, according to reports in The New York Times and other outlets. Binance fired or suspended the officials after they pointed out the Iran transactions; the company said the discipline had nothing to do with their warnings.

The reports in The Times raised concerns in Congress, where a top Senate Democrat opened an investigation into Binance. For years, the company has struggled to eliminate financial crime on its platform. It paid $4.3bn in penalties in 2023 after pleading guilty to violating US financial laws.

«

Handy to have $4.3bn lying around so you can pay a fine. And what a remarkable correlation between the comparatively small use of crypto and the relatively large number of times its practitioners fall foul of anti-money laundering laws and similar.
unique link to this extract


Nine-year-old runs up $118,000 bill on dad’s corporate credit card advertising his Roblox YouTube channel • The Register

Brandon Vigliarolo:

»

There’s kids running up a few hundred bucks on in-app purchases, and then there’s the case of nine-year old YouTuber Mighty Mike Plays, who just cost his dad $118,000 – and possibly his job – after running a bunch of Google ad campaigns for his channel on papa’s company credit card.

The young Roblox and Minecraft YouTuber posted a video Tuesday – or perhaps it’s more apt to say his father Dave posted the video to explain why Might Mike Plays might be going dark for a while. In it, nine-year-old Mike explains “I think this might be the end,” before handing the mic to his dad.

“My son has been running some high-budget campaigns on YouTube that I was pretty much unaware of,” Dave explained after taking the mic.

Dave didn’t figure this out after spotting the charges himself, however. He learned about it after being called into a meeting with his manager and the finance department at his employer, an unnamed “large” company where he works in marketing. Dave was confronted with a spreadsheet with a full breakdown of three weeks of advertising spending from his corporate card that totaled $118K and asked to explain it.

“It dawned on me that, oh crap, this is probably something my son did,” Dave said in the video. “The number sitting on the page in front of me had about five more digits than it should have.” 

«

Sure have moved on from kids hitting credit limits on parents’ cards with in-app purchases in games. Though maybe this is just the same, but the app is different.
unique link to this extract


• Why do social networks drive us a little mad?
• Why does angry content seem to dominate what we see?
• How much of a role do algorithms play in affecting what we see and do online?
• What can we do about it?
• Did Facebook have any inkling of what was coming in Myanmar in 2016?

Read Social Warming, my latest book, and find answers – and more.


Errata, corrigenda and ai no corrida: none notified

Leave a comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.