A new social network is just AI chatbots, all the time. A vision of heaven, or hell? CC-licensed photo by James Royal-Lawson 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: all hail the algorithm.
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.
A new era in sabotage: turning ordinary devices into grenades, on a mass scale • The New York Times
David Sanger:
»
the presumed Israeli sabotage of hundreds or thousands of pagers, walkie-talkies and other wireless devices used by Hezbollah has taken the murky art of electronic sabotage to new and frightening heights. This time the targeted devices were kept in trouser pockets, on belts, in the kitchen. Ordinary communication devices were turned into miniature grenades.
And while the target was Hezbollah fighters, the victims were anyone standing around, including children. Lebanese authorities say 11 people died and more than 2,700 were injured in Tuesday’s attack. On Wednesday, at least 20 more people were killed and 450 injured in a second round of attacks with exploding walkie-talkies.
There is reason to fear where this attack on Hezbollah fighters might go next. The history of such sabotage is that once a new threshold is crossed, it becomes available to everyone.
Of course, there is nothing new about sabotaging phones or planting bombs: terrorists and spy agencies have done that for decades. What made this different was the mass scale, the implantation of explosives on so many devices at once. Such subterfuge is difficult to pull off, because it requires getting deep into the supply chain. And that, in a way, is the best reason for people not to be afraid of their internet-connected refrigerators and computers.
But our sense of vulnerability about how everyday implements connected to the internet can become deadly weapons may be just beginning.
«
There’s a suggestion that the pagers and walkie-talkies were routed through Iran – after all, it’s hard to get a licence to export to Hezbolla – which will increase the group’s uncertainty about what can be trusted.
At The Atlantic, Eliot Cohen points out:
»
Hezbollah members will now be unlikely to trust any form of electronics: car keys, cellphones, computers, television sets. Myth and legend, no doubt reinforced by an information-warfare campaign, will magnify Israel’s success in getting inside black boxes no matter how big or how small. An army skittish about any kind of electronics is one that is paralyzed—an individual leader, like Hamas’s Yahya Sinwar, can communicate without a phone, but an entire organization cannot.
The Iranians, already reeling from the assassination of the political head of Hamas in a Revolutionary Guard Corps guesthouse on the day of the inauguration of the new president, now have much to wonder about as well. How, they must ask themselves, did the Israelis penetrate the supply chain? How did they get access to the pagers? How did they know that this batch was going to Hezbollah? How did they manage to foil whatever security precautions had been taken?
«
The ground war is inevitable, Cohen says:
»
Both Iran and Hezbollah have to know that Israel now believes itself to be fighting an existential fight, with a different set of rules.
«
AI-generated code is causing outages and security issues in businesses • Tech Republic
Fiona Jackson:
»
Bilkent University researchers found that the latest versions of ChatGPT, GitHub Copilot, and Amazon CodeWhisperer generated correct code just 65.2%, 46.3%, and 31.1% of the time, respectively.
Part of the problem is that AI is notoriously bad at maths because it struggles to understand logic. Plus, programmers are not known for being great at writing prompts because “AI doesn’t do things consistently or work like code,” according to Wharton AI professor Ethan Mollick.
In late 2023, more than half of organisations said they encountered security issues with poor AI-generated code “sometimes” or “frequently,” as per a survey by Snyk. But the issue could worsen, as 90% of enterprise software engineers will use AI code assistants by 2028, according to Gartner.
Tariq Shaukat, CEO of Sonar and a former president at Bumble and Google Cloud, is “hearing more and more about it” already. He told TechRepublic in an interview, “Companies are deploying AI-code generation tools more frequently, and the generated code is being put into production, causing outages and/or security issues.
“In general, this is due to insufficient reviews, either because the company has not implemented robust code quality and code-review practices, or because developers are scrutinising AI-written code less than they would scrutinise their own code.
“When asked about buggy AI, a common refrain is ‘it is not my code,’ meaning they feel less accountable because they didn’t write it.”
«
No actual evidence is provided of these outages – we’re told the CEO of a company is “hearing more and more” about them – and I’d also like to know how good your average Joe or Jane programmer is at generating “correct” code. (The linked paper doesn’t specify that – big miss! – though it did introduce me to the concept of a “code smell” which I’d somehow never come across before.)
unique link to this extract
Google employees’ attempts to hide messages from investigators might backfire • The Verge
Lauren Feiner on how Google’s peculiar approach to retaining documents on legal request might not help it in its adtech trial, because the judge could “draw an adverse inference” about why that happened:
»
Brad Bender, another Google ad tech executive who testified earlier in the week, described conversations with colleagues over chat as more akin to “bumping into the hall and saying ‘hey we should chat.’” The DOJ also questioned former Google executive Rahul Srinivasan about emails he marked privileged and confidential, asking what legal advice he was seeking in those emails. He said he didn’t remember.
Google employees were well aware of how their written words could be used against the company, the DOJ argued, pointing to the company’s “Communicate with Care” legal training for employees. In one 2019 email, Srinivasan copied a lawyer on an email to colleagues about an ad tech feature and reminded the group to be careful with their language. “We should be particularly careful when framing something as a ‘circumvention,’” he wrote. “We should assume that every document (and email) we generate will likely be seen by regulators.” The email was labeled “PRIVILEGED and CONFIDENTIAL.”
While the many documents shown by the DOJ demonstrate that Google often discussed business decisions in writing, at other times, they seemed to intentionally leave the documentation sparse. “Keeping the notes limited due to sensitivity of the subject,” a 2021 Google document says. “Separate privileged emails will be sent to folks to follow up on explicit [action items].”
“We take seriously our obligations to preserve and produce relevant documents,” Google spokesperson Peter Schottenfels said in a statement.
«
The judge in a previous trial didn’t think Google took those obligations seriously enough. Remains to be seen what this one will think.
unique link to this extract
Norway: electric cars outnumber petrol for first time in ‘historic milestone’ • AFP via The Guardian
»
Electric cars now outnumber petrol cars in Norway for the first time, an industry organisation has said, a world first that puts the country on track towards taking fossil fuel vehicles off the road.
Of the 2.8m private cars registered in the Nordic country, 754,303 are all-electric, against 753,905 that run on petrol, the Norwegian road federation (OFV) said in a statement.
Diesel models remain the most numerous at just under 1m, but their sales are falling rapidly.
“This is historic. A milestone few saw coming 10 years ago,” said OFV director Øyvind Solberg Thorsen.
“The electrification of the fleet of passenger cars is going quickly, and Norway is thereby rapidly moving towards becoming the first country in the world with a passenger car fleet dominated by electric cars.”
Norway, paradoxically a major oil and gas producer, has set a target for all new cars being sold to be zero emission vehicles – mostly EVs since the share of hydrogen cars is so small – by 2025, 10 years ahead of the EU’s goal.
In August, all-electric vehicles made up a record 94.3% of new car registrations in Norway, boosted by sales of the Tesla Model Y.
«
There’s a nice read from back in March about how Norway got there, apparently involving Morten from A-ha before he was Morten from A-ha.
unique link to this extract
“Dead Internet theory” comes to life with new AI-powered social media app • Ars Technica
Benj Edwards:
»
For the past few years, a conspiracy theory called “Dead Internet theory” has picked up speed as large language models (LLMs) like ChatGPT increasingly generate text and even social media interactions found online. The theory says that most social Internet activity today is artificial and designed to manipulate humans for engagement.
On Monday, software developer Michael Sayman launched a new AI-populated social network app called SocialAI that feels like it’s bringing that conspiracy theory to life, allowing users to interact solely with AI chatbots instead of other humans. It’s available on the iPhone app store, but so far, it’s picking up pointed criticism.
After its creator announced SocialAI as “a private social network where you receive millions of AI-generated comments offering feedback, advice & reflections on each post you make,” computer security specialist Ian Coldwater quipped on X, “This sounds like actual hell.” Software developer and frequent AI pundit Colin Fraser expressed a similar sentiment: “I don’t mean this like in a mean way or as a dunk or whatever but this actually sounds like Hell. Like capital H Hell.”
SocialAI’s 28-year-old creator, Michael Sayman, previously served as a product lead at Google, and he also bounced between Facebook, Roblox, and Twitter over the years. In an announcement post on X, Sayman wrote about how he had dreamed of creating the service for years, but the tech was not yet ready. He sees it as a tool that can help lonely or rejected people.
“SocialAI is designed to help people feel heard, and to give them a space for reflection, support, and feedback that acts like a close-knit community,” wrote Sayman. “It’s a response to all those times I’ve felt isolated, or like I needed a sounding board but didn’t have one. I know this app won’t solve all of life’s problems, but I hope it can be a small tool for others to reflect, to grow, and to feel seen.”
«
No. Nonononononono. This is Black Mirror stuff. You could almost, if you looked sidewise enough, think it might be something that those at end of life or with dementia might use, but even so it feels like denying humans the thing they most need: human companionship.
unique link to this extract
Solar power continues to surge in 2024 • Ember
Euan Graham and Nicolas Fulghum:
»
Ember’s analysis of the latest data on monthly capacity installations shows that the world is on track to reach 593 GW of solar installations by the end of this year. This would once again surpass most industry forecasts, and comes after 2023 showed record growth in solar installations of 86% compared to 2022. Countries need to plan ahead to make the most of the high levels of solar capacity being built today and ensure the continued build-out of capacity in the coming years.
Ember estimates that at the current rate of additions, the world will install 593 GW of solar panels this year. That’s 29% more than was installed last year, maintaining strong growth even after an estimated 87% surge in 2023. In 2024, an estimated 292 GW of solar capacity was installed by the end of July.
«
Half of that will be in China, but countries such as Pakistan and Saudi Arabia are also jumping on this very fast-moving bandwagon. And most countries are accelerating their installs. (Pakistan because its grid company is heavily indebted to.. China.)
Now we just need the battery capacity to gather pace.
unique link to this extract
Elon Musk’s X finds way around Brazil ban and goes live again for many users • The New York Times
Jack Nicas:
»
In his continuing fight with the Brazilian authorities, score one for Elon Musk — at least briefly.
On Wednesday, his social network, X, suddenly went live again for many across Brazil after three weeks of being blocked under orders from Brazil’s Supreme Court. The reason? X made a technical change to how it routes its internet traffic, enabling the site to evade the digital roadblocks set up in recent weeks by Brazilian internet providers.
But by Wednesday night, the president of Brazil’s telecommunications regulator, Anatel, said his agency believed it would soon be able to restore the block.
The new twist showed how Mr. Musk appears far from backing down in Brazil, making the dispute a significant test of strength between national sovereignty and the borderless power of internet companies. Brazil’s Supreme Court blocked X because the company defied orders to remove certain accounts and then closed its offices in the country to avoid consequences.
…“You can’t just block Cloudflare because you would block half of the internet,” said Basílio Perez, president of Abrint, the trade group for Brazilian internet providers. He said Cloudflare supported more than 24 million websites, including those of the Brazilian government and banks.
But hours later, Anatel’s president, Carlos Baigorri, said in an interview that Cloudflare had agreed to isolate internet traffic from X, enabling Brazilian internet providers to easily target and block that traffic.
“Cloudflare has been extremely cooperative,” he said. “It really shows the diverse reactions from two companies, Cloudflare and X.”
«
Unclear if this really was Musk trying to find a way around the Brazilian block or – as claimed in this BBC version of the story – just an “inadvertent mistake”.
unique link to this extract
The wordfreq dataset will not be updated any more · GitHub
Robyn Speer:
»
The wordfreq data is a snapshot of language that could be found in various online sources up through 2021. There are several reasons why it will not be updated anymore.
1: Generative AI has polluted the data
I don’t think anyone has reliable information about post-2021 language usage by humans.
The open Web (via OSCAR) was one of wordfreq’s data sources. Now the Web at large is full of slop generated by large language models, written by no one to communicate nothing. Including this slop in the data skews the word frequencies.
…As one example, Philip Shapira reports that ChatGPT (OpenAI’s popular brand of generative language model circa 2024) is obsessed with the word “delve” in a way that people never have been, and caused its overall frequency to increase by an order of magnitude.
2: Information that used to be free became expensive
wordfreq is not just concerned with formal printed words. It collected more conversational language usage from two sources in particular: Twitter and Reddit.
The Twitter data was always built on sand. Even when Twitter allowed free access to a portion of their “firehose”, the terms of use did not allow me to distribute that data outside of the company where I collected it (Luminoso). wordfreq has the frequencies that were built with that data as input, but the collected data didn’t belong to me and I don’t have it anymore.
«
(Thanks wendyg for the link.)
unique link to this extract
Marree Man: The enduring mystery of a giant outback figure (2018) • BBC News
Frances Mao:
»
This week [end of June 2018] marks 20 years since a helicopter pilot flying over central Australia spotted the outline of a giant man drawn into the earth.
The 4.2km (2.5 miles) tall figure, on a remote plateau in South Australia, is often thought to depict an Aboriginal hunter. Dubbed Marree Man after a nearby town, it is one of the world’s largest designs to be etched into the ground.
But mystery surrounds who created it – and why.
Earlier this week, Australian entrepreneur Dick Smith offered a A$5,000 (£2,800; $3,700) reward for any information about the artwork’s origins. “How has it been kept secret for 20 years?” he said on the Australian Broadcasting Corporation (ABC) on Monday, external.
Marree Man has been a subject of fascination since its discovery in the desert about 700km north of Adelaide. It has gained popularity on tourism flights because it is too large to be viewed from the ground.
With an outline measuring a total of 28km, Marree Man had an initial depth of about 35cm (14 inches), according to local media reports. Locals believe it portrays an Aboriginal man carrying a woomera – a throwing stick – in his left hand.
Marree publican Phil Turner says he is convinced that its creator, or creators, were “professionals” who possibly used GPS technology.
«
I came across this because someone on Twitter did a thread about how they spotted it out of an airplane window the other day and were very proud of tracking it down via Google Earth and old Landsat photos. (Overkill, really.)
Plenty of ideas about who might have made it; clearly they used earthmoving equipment, and satellite photos show it was created between 27 May and 12 June 1998. It’s 28km of tracks: to do that accurately in two weeks you need big plant and accurate GPS – which in those days wasn’t guaranteed to civilian users.
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 