The chances of AI designing a bioweapon are between zero and none because it takes real work by humans in a laboratory. CC-licensed photo by DisconnecTomas 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. Not weaponised. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. Observations and links welcome.
Israel detonates Hezbollah walkie-talkies a day after pager attack • Axios
Barak Ravid:
»
Israel on Wednesday blew up thousands of two-way personal radios used by Hezbollah members in Lebanon in a second wave of an intelligence operation that started on Tuesday with the explosions of pager devices, two sources with knowledge of the operation told Axios. More than a dozen people were killed and hundreds of others were wounded.
The second wave of clandestine attacks is another serious security breach in Hezbollah’s ranks and increases pressure on the militant Lebanese group.
Lebanon’s health ministry said 14 people were killed and 450 wounded in the attacks on Wednesday. The walkie-talkies were booby-trapped in advance by Israeli intelligence services and then delivered to Hezbollah as part of the militia’s emergency communications system, which was supposed to be used during a war with Israel, the sources said.
The attack further damages Hezbollah’s military command and control system.
«
Lots of detail has come out in the intervening 24 hours. A Taiwanese company licensed the manufacture of the pagers in Europe to a Hungarian company, but in fact a Bulgarian company made them, and they sat on some docks for three months.
The walkie-talkies had ICOM branding, but the company says it hasn’t made that model for 20 years. So they’re knockoffs, or fakes, and contained more explosive than pagers.
There’s no denying this is terrorism – civilians and children were injured, some with life-altering injuries. The only question is whether it’s justified. Will Hezbollah roll over? Or will the rocket attacks intensify? Making everyday technology into a weapon changes everything.
unique link to this extract
Massive China-state IoT botnet went undetected for four years—until now • Ars Technica
Dan Goodin:
»
The FBI has dismantled a massive network of compromised devices that Chinese state-sponsored hackers have used for four years to mount attacks on government agencies, telecoms, defence contractors, and other targets in the US and Taiwan.
The botnet was made up primarily of small office and home office routers, surveillance cameras, network-attached storage, and other internet-connected devices located all over the world. Over the past four years, US officials said, 260,000 such devices have cycled through the sophisticated network, which is organized in three tiers that allow the botnet to operate with efficiency and precision. At its peak in June 2023 Raptor Train, as the botnet is named, consisted of more than 60,000 commandeered devices, according to researchers from Black Lotus Labs, making it the largest China state botnet discovered to date.
Raptor Train is the second China state-operated botnet US authorities have taken down this year. In January, law enforcement officials covertly issued commands to disinfect Internet of Things devices that hackers backed by the Chinese government had taken over without the device owners’ knowledge. The Chinese hackers, part of a group tracked as Volt Typhoon, used the botnet for more than a year as a platform to deliver exploits that burrowed deep into the networks of targets of interest.
«
Half of it was in the US; a quarter in Europe. (Only 3% in the UK. Hurrah.)
unique link to this extract
Exclusive: Google offered to sell part of ad tech business, not enough for EU publishers, sources say • Reuters via MSN
Foo Yun Chee and Jody Godoy:
»
Alphabet’s Google took a major step this year to end an EU antitrust investigation with an offer to sell its advertising marketplace AdX but European publishers rejected the proposal as insufficient, two people with direct knowledge of the matter said on Wednesday.
Google’s lucrative ad tech business attracted EU regulatory scrutiny last year following a complaint from the European Publishers Council.
The European Commission subsequently charged Google with favouring its own advertising services, opening its fourth case against the world’s most-popular search engine.
Google has never before offered to sell an asset in an antitrust case, according to three lawyers involved in antitrust cases who did not have permission to speak publicly.
The company is on trial in the US, fighting claims by antitrust authorities who seek to make Google sell its Ad Manager product, which contains AdX and Google’s publisher ad server, known as DFP.
Publishers rejected Google’s proposal because they want it to divest more than just AdX to address conflicts of interest due to its presence in almost all levels of the ad tech supply chain, the people said. They said the EU antitrust enforcer was aware of the offer.
“As we have said before, the European Commission’s case about our third-party display advertising products rests on flawed interpretations of the ad-tech sector, which is fiercely competitive and rapidly evolving. We remain committed to this business,” a Google spokesperson said.
«
Doesn’t sound like you were committed to the business if you offered to sell it, to be honest. Google is fighting this case in multiple locations: EU, UK, US. Not going well in any of them.
unique link to this extract
Google gets win from European court as €1.5bn fine overturned • FT via Ars Technica
Javier Espinoza:
»
Google has won an appeal against a €1.5bn competition fine from the European Commission in a victory for the Big Tech group as it comes under growing scrutiny from Brussels regulators.
The EU’s General Court said on Wednesday that while it accepted “most of the commission’s assessments” that the company had used its dominant position to block rival online advertisers, it annulled the hefty fine levied against Google in the case.
When launching the action against Google in 2019, Margrethe Vestager, the bloc’s competition chief, said that the search giant had imposed anti-competitive restrictions on third-party websites for a decade between 2006 and 2016. She justified the €1.5bn fine by arguing that it reflected the “serious and sustained nature” of the infringement.
However, the Luxembourg-based General Court found that the commission, the EU’s executive arm, had failed “to take into account all the relevant circumstances in its assessment of the duration of the contractual clauses that it had found to be unfair.”
The commission, which is likely to appeal, said it took “note” of the judgment and “will carefully study the judgment and reflect on possible next steps.”
«
Pretty hard to keep track of who’s paying and who’s repaying with all these nine-digit fines.
unique link to this extract
iPhone 16 Pro Camera Review: Kenya — Travel Photographer • Austin Mann
The photographer has been testing the new camera (which has a phone attached):
»
Last week at the Apple keynote event, the iPhone camera features that stood out the most to me were the new Camera Control button, upgraded 48-megapixel Ultra Wide sensor, improved audio recording features (wind reduction and Audio Mix), and Photographic Styles.
When I finally landed in Nairobi (after more than 15 hours of flight delays), these were the features I was most excited to put to the test.
Over the past week we’ve traveled over a thousand kilometers across Kenya, capturing more than 10,000 photos and logging over 3TB of ProRes footage with the new iPhone 16 Pro and iPhone 16 Pro Max cameras. Along the way, we’ve gained valuable insights into these camera systems and their features.
As iPhone cameras have improved over the years, finding their boundaries has become more challenging, but Kenya’s vast and diverse landscape has provided the ideal setting to really push these devices to their limits.
My question today is the same as it’s always been: how will this new tech make our pictures and videos better?
«
This set of pictures definitely gave me an absolutely irresistible urge to.. go on safari in Kenya.
unique link to this extract
Against prophecy among the machines
Anselm Levskaya:
»
The backers of California’s SB1047 routinely cite AI-enabled bioweapons as a threat justifying the radical regulatory regime that places a locus of liability on general computational models, rather than on particular dangerous applications or criminal acts.
For six years I’ve worked on scaling generative AI models at a leading industrial lab – I’ve worked on some of the largest language models ever trained. But before that I was an experimental lab-scientist for two decades working in synthetic biology, optogenetics, and immuno-oncology.
As a biologist I feel compelled to comment on this risk-scenario, speaking solely on my own behalf as a resident of California, and as someone who generally admires my state senator Scott Wiener, the primary sponsor of the bill. The claims being made about biological risks do not reflect scientific reality.
Language models won’t create bioterrorists any more readily than celebrity cookbooks have created Michelin star chefs.
Those of us who have made a living in the practical, experimental arts of the world know: it is not easy to explain to someone how a line cook tests the elasticity of a steak to measure doneness, how a machinist listens to the sound of a cutting bit for signs of a good feed rate, or how a biologist reads the fluorescent smears in a gel to see that a polymerase chain reaction’s annealing temperature was low. These tacit skills are acquired by hand, via repetition, under apprenticeship.
The preparedness and safety studies written by various AI labs on the subject of biological risks all try to assay the ability of users to elicit dangerous experimental recipes from language models. None publish the protocols or results in the sufficient detail necessary for an independent review, but it is clear that none of them actually assay the ability of any of these study participants to translate their dangerous chats into actual bench practice.
Having trained many people in experimental biology, I can tell you that it is a long and cruel apprenticeship requiring the disabuse of enthusiasts’ ideas that manipulating the machines of life is straightforward. One can watch their sanity fray as they slowly learn to stitch DNA constructs together through the arcane and interminable mixing of clear and precious droplets of fluid. One can watch their bright eyes dim as they deal with failed DNA assemblies, failed cell transfections, fritzing instruments, and their own faltering punctiliousness. One can only try to cheer their fatalism as they learn that even the smallest of ambitions will demand the sacrifice of their youth to the lonely nights of a laboratory.
«
AI started as a dream to save humanity. Then Big Tech took over • Bloomberg
Parmy Olson, with an extract from her new book “Supremacy”:
»
Generative AI promises to make people more productive and bring more useful information to our fingertips through tools like ChatGPT. But every innovation has a price to pay. Businesses and governments are adjusting to a new reality where the distinction between real and “AI-generated” is a crapshoot. Companies are throwing money at AI software to help displace their employees and boost profit margins. And devices that can conduct new levels of personal surveillance are cropping up.
We got here after the visions of two innovators [Demis Hassabis and, separately, Sam Altman] who tried to build AI for good were eventually ground down by the forces of monopoly. Their story is one of idealism but also one of naivety and ego — and of how it can be virtually impossible to keep an ethical code in the bubbles of Big Tech and Silicon Valley. Altman and Hassabis tied themselves into knots over the stewardship of AI, knowing that the world needed to manage the technology responsibly if we were to stop it from causing irreversible harm. But they couldn’t forge AI with godlike power without the resources of the world’s largest tech firms. With the goal of enhancing human life, they would end up empowering those companies, leaving humanity’s welfare and future caught in a battle for corporate supremacy.
After selling DeepMind to Google in 2014, Hassabis and his co-founders tried for years to spin out and restructure themselves as a nonprofit-style organization. They wanted to protect their increasingly powerful AI systems from being under the sole control of a tech monolith, and they worked on creating a board of independent luminaries that included former heads of state like Barack Obama to oversee its use. They even designed a new legal charter that would prioritize human well-being and the environment. Google appeared to go along with the plan at first and promised its entity billions of dollars, but its executives were stringing the founders along.
«
Lionsgate, studio behind ‘John Wick,’ signs deal with AI startup Runway • WSJ
Jessica Toonkel:
»
The entertainment company behind “The Hunger Games” and “Twilight” plans to start using generative artificial intelligence in the creation of its new movies and TV shows, a sign of the emerging technology’s advance in Hollywood.
Lions Gate Entertainment has agreed to give Runway, one of several fast-evolving AI startups, access to its content library in exchange for a new, custom AI model that the studio can use in the editing and production process.
The deal—the first of its kind for Runway and one that could become a blueprint in the entertainment industry—comes as creatives, actors and studio executives debate whether to use the new technology and how to protect their copyright material. Advocates say generative AI can enhance creators’ work and help a cash-strapped industry save time and money.
Michael Burns, vice chairman of Lionsgate Studio, expects the company to be able to save “millions and millions of dollars” from using the new model. The studio behind the “John Wick” franchise and “Megalopolis” plans to initially use the new AI tool for internal purposes like storyboarding—laying out a series of graphics to show how a story unfolds—and eventually creating backgrounds and special effects, like explosions, for the big screen.
“We do a lot of action movies, so we blow a lot of things up and that is one of the things Runway does,” Burns said.
«
Runway, it should be pointed out, is one of the AI startups against which a copyright case is proceeding. And this is also not going to be a popular move at all.
unique link to this extract
ITV launches ITV Kerching • Advanced Television
»
ITV has launched its new consumer-facing affiliate marketing brand, ITV Kerching – a browser extension which aims to simplify the process of finding discount codes across hundreds of online retailers.
Available from the ITVX website and mobile app, ITV Kerching uses tech company Kindred’s technology to search for discount codes from retailers to help consumers search for the best price available for their online purchases.
Following activation, when a user lands on a retailer’s website, coupons are searched for, applied and redeemed with minimum effort.
ITV research found that 79% of ITVX viewers already use discount codes and one-third said their usage has increased in the last year. Among those that don’t use discount codes, 48% say they don’t know where to look, a problem that ITV Kerching solves by doing the hard work for you.
The launch of ITV Kerching is being supported by marketing promotion across ITV’s linear channels, ITVX, social media and email marketing.
In addition to the launch, ITV is also investing up to £8.5m of advertising inventory across ITV’s channels and ITVX in return for a minority equity stake in Kindred, the company powering ITV Kerching.
«
Google long since passed ITV for UK advertising revenue; so the plan now is to find any additional revenue you can, even if it’s down-the-sofa stuff like this.
unique link to this extract
Character Limit by Kate Conger and Ryan Mac review – Musk’s Twitter takeover • The Guardian
James Ball:
»
Even people who never used Twitter more or less know its story over the last few years: Musk bought it, gave it a juvenile new name, X, and the whole thing seems to have been a complete mess that has made everyone miserable, including Musk himself.
That makes the job of New York Times reporters Kate Conger and Ryan Mac in their chronicle of the takeover and its resulting fallout, a difficult one: almost anyone who might actually read this book is pretty familiar with how things have played out. How can you make a story compelling when each step along the way has already been so heavily covered?
Conger and Mac’s answer to that is their astonishing ability to take the reader into almost every room that mattered during the contentious $44bn acquisition. The book opens with a Twitter data scientist getting ready to meet Musk, ostensibly in a bid to keep his job. The employee, however, has already decided he’s quitting and is instead using the opportunity to level with the new boss. The encounter goes predictably badly, leading him to accuse Musk of being one of the most gullible men on the planet. The book records Musk’s response as being two words long: “Fuck you.”
At other moments, the narration seems to know Musk’s exact movements when he was at home with his then-girlfriend Claire Elise Boucher (better known as the musician Grimes), or the conversations that take place on his plane. Such is the apparent omniscience that impressive accounts of goings-on in boardrooms and executive suites during the takeover seem par for the course.
Musk himself did not grant the authors an interview. Some of their insights come from court documents and other reporting, but there is no doubt that Conger and Mac enjoyed unmatched access to a range of characters from all sides. You couldn’t hope for a better ringside seat on the unfolding drama.
«
Twitter (now X) has given rise to a multitude of books, charting its rise and plateau and now fall. This certainly looks worth the time.
unique link to this extract
| • Why do social networks drive us a little mad? • Why does angry content seem to dominate what we see? • How much of a role do algorithms play in affecting what we see and do online? • What can we do about it? • Did Facebook have any inkling of what was coming in Myanmar in 2016? Read Social Warming, my latest book, and find answers – and more. |
Errata, corrigenda and ai no corrida: none notified
The Overspill: when there's more that I want to say 