The drum sound that defined a lot of 1980s music came about through a studio accident, like all the best discoveries. CC-licensed photo by badgreeb RECORDS 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. Badum-tisch. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.
Chip stocks tumble as AI sell-off deepens • Financial Times
Emily Herbert, Song Jung-a and David Keohane:
»
Global semiconductor stocks tumbled as fears over the durability of the AI boom intensified ahead of results from some of Silicon Valley’s biggest companies this week.
US chip and memory stocks extended the previous day’s losses, dragging the Nasdaq 100 index down as much as 1.8% on Tuesday and taking its losses since its early June high briefly into correction territory at more than 10%. It later recovered to close 1% lower on the day.
Memory group Sandisk, which fell 11% on Monday, plunged more than 14% on Tuesday. The stock has lost more than 50% of its value this month but nevertheless remains more than four times higher since the start of 2026.
Western Digital and Micron fell 6.9% and 8.9% respectively and Advanced Micro Devices dropped 8.1%.
Semiconductor stocks, which powered global markets higher in the first half of the year as investors bet they would be the biggest beneficiaries of vast AI spending, have borne the brunt of a brutal sell-off in recent weeks as many of those bets have unwound.
“Worries around funding uncertainties, capex increases and Big Tech free cash flow have taken centre stage,” said Venu Krishna, head of US equities strategy at Barclays.
…South Korea’s Kospi led declines in Asia earlier on Tuesday, falling more than 10% and prompting a short halt in trading, after investors dumped shares in chipmakers SK Hynix and Samsung Electronics.
SK Hynix fell 14.6% on Monday, while its larger rival, Samsung Electronics, dropped 13.4%. Both companies’ share prices have dropped more than a third in July’s chipmaker sell-off — although still remain 138% and 83% higher in the year to date.
«
The circular funding model is just about holding together, but some people have decided they’ve seen enough. Is this just a hiccup, or is it going to turn into the full vomiting camel?
unique link to this extract
eBay and ex-executives to pay $55.7m to couple sent cockroaches in harassment campaign • Reuters via The Guardian
»
Ebay and three former top executives have agreed to pay $55.7m to resolve a lawsuit by a Massachusetts couple who fell victim to a bizarre stalking and harassment campaign carried out by several of its employees in retaliation for their coverage of the e-commerce company in a newsletter they publish.
David and Ina Steiner announced the settlement late on Monday, bringing an end to a case they filed in 2021 in Boston federal court after several former eBay employees were prosecuted for targeting them in a cyberstalking campaign that involved sending the couple cockroaches, fly larvae and a bloody Halloween pig mask.
As part of the settlement, eBay agreed to pay $46.15m to the couple; fund $6m in charitable contributions to non-profit organizations; and issue a “strongly-worded” statement about the conduct of three of its top former executives.
“We believe this resolution sends a clear message that corporations and their executives cannot engage in this type of misconduct without facing significant consequences,” Christopher Murphy, the Steiners’ attorney, said in a statement.
Ebay said in a statement that what happened to the Steiners was “wrong, reprehensible and should never have happened”.
“We continue to extend our deepest apologies to the Steiners,” the company said. “This agreement is consistent with our commitment to fairly compensate the Steiners and fulfills our efforts to make things right.”
Seven former eBay workers pleaded guilty and received sentences of as high as 57 months in prison for participating in the scheme, which also involved several of them traveling from California to Natick, Massachusetts, to surveil the Steiners and try to install a GPS tracking device on their car.
«
This case keeps on going. Ebay paid a $3m fine to the FBI in 2024, for its activities in 2019 in retaliation for what the pair wrote. Ebay was “criminally and financially responsible for emotionally, psychologically, and physically terrorizing” the couple, the FBI said at the time.
unique link to this extract
Smart rings from Pebble and Stream are looking like the AI gadget to beat • The Verge
David Pierce:
»
nothing has convinced me of the future of talking to my computer quite like a demo I got recently from Mina Fahmi. Fahmi is the CEO of Sandbar, which is making a smart ring called the Stream that is scheduled to ship later this year. As I sat in a conference room in Sandbar’s New York City offices, Fahmi gave me a rundown of the device’s features, including the latest one: dictation to all your devices. He opened up a new note in Apple Notes, held the ring up to his mouth, pressed the device’s touchpad with his thumb, and chatted quietly to himself for a few seconds. A moment after he finished, two complete — and correctly transcribed — sentences appeared in the note.
The Stream itself is basically just a microphone, a battery, and a Bluetooth connection to your phone. You could describe a lot of AI gadgets the same way, now that seemingly everyone wants to sell you a way to record your conversations and meetings to extract summaries and action items. Companies like Plaud and Pocket and FoCase and Mobvoi and so many others are all happy to sell you these portable recorders — bonus points if it sticks to your phone to record phone calls.
Fahmi and Sandbar are trying to do something different. The Stream is a more introverted device, focused on capturing your voice rather than everyone else’s. (Another of the ring’s promised features is a whisper mode — Fahmi spoke to the thing so quietly I couldn’t hear a word he said, and the Stream still transcribed perfectly. Or at least Fahmi said it did.) Fahmi says the idea for the device came because he likes to walk a lot and got annoyed having to dig out his phone every time he had an idea or wanted to ask his AI assistant for something.
The Stream ring, accordingly, has three jobs: to start a chat with an AI assistant, to take notes, and to dictate text. Right now, most of those things happen in the Stream companion app. But that’s not the right long-term answer, nor is it Sandbar’s long-term plan. Ultimately, devices like these should be connected to all the other tools you use, so that you can chat with whatever assistant you want, take notes straight to your existing app, and dictate into any text box anywhere.
«
A ring, or a watch? Both are likely candidates for an AI device, though all this walking around whispering to yourself is even stranger than the usual walking around chatting to yourself.
unique link to this extract
Despite AI hype, Google’s data shows workers aren’t automating themselves away • Ars Technica
Kyle Orland:
»
Anyone following the AI space is by now familiar with lofty claims that AI models will soon be better than humans at everything and capable of replacing vast swaths of the human workforce. In a new study from Google Research, though, a team that looked at how workers are actually using Gemini “[did] not find evidence… to support the claims that AI is about to cause massive automation and displacement of white-collar work…”
The paper, released last week, introduces the “AI & Economy ATLAS,” an Activity, Task, Landscape, and Adoption Study of 15 million anonymized AI interactions across the Gemini App, Google’s AI Mode, and the Gemini API. Their initial review of the data finds that, while AI sees some significant use across a wide variety of occupations, that use “remains shallow and overwhelmingly collaborative in nature, with end-to-end task automation limited in scope.”
To come to these conclusions, Google researchers used an automated classifier to sort work-based AI interactions using the Bureau of Labor Statistics’ Standard Occupational Classifications and O*NET’s more detailed database of specific work interactions. While this method required some probabilistic classification of “inherently uncertain” interactions, verification by human reviewers found it to be a reliable gauge of how Gemini prompts were being used for work.
Unsurprisingly, white-collar jobs in fields like computers, finance, and arts and entertainment were some of the ones where the volume of Gemini use was overrepresented (when compared to their prevalence across the US economy). Financial/market analysts, software developers, and systems administrators were some of the relatively heaviest users of AI for job-related tasks, while salespeople, transportation workers, and food preparation/service workers were heavily underrepresented in the AI use data.
«
Not really a surprise, is it. Going to be a while before plumbers and hairdressers are AI-powered robots.
unique link to this extract
Substackers say new AI detection tool is a “witch hunt” • 404 Media
Emanuel Maiberg:
»
Substack contributors are rejecting the company’s decision to flag AI generated writing on the newsletter publishing platform, with several creators saying the new AI detection features are a “witch hunt.” Some Substack users who use AI to help them write or edit articles say the new policy unfairly discriminates against them, while other users who don’t use AI in their writing at all are worried that Substack will mistakenly flag their writing as being AI generated and tarnish their reputation.
“I’m not going to apologize for using AI in the creation process. I wrote for 20 years without AI, I could do it again if I wanted to,” Mack Collier, who has a small Substack called Backstage Pass, wrote. “My output would fall, my posts would be less structured, and it would be more obvious that they needed a good editor. Using AI improves the overall quality of my writing. That’s why I use it.”
“These detectors are notoriously, wildly inaccurate,” Alice Lemee, a ghostwriter and digital writing coach, said in a video she made about Substack’s new AI detection features. “All it takes is one false accusation for a writer to have their reputation almost irreversibly tarnished.”
Substack announced the AI detection features on Tuesday in a post from CEO Christ Best, who said that “It’s getting harder to tell what’s real on the internet” and that “when content made by no one takes over parts of the internet that are supposed to be human, it pollutes the commons and makes it hard to discover and hear human voices.”
Best explained that Substack has partnered with the AI detection tool Pangram, which is now built in to Substack and allows any user to scan a piece of writing. Pangram then produces a percentage-based score determining how much of the writing was AI or human generated.
«
The people who are most upset about this are the ones who use AI the most, and their opinions can really discounted on the basis that Substack isn’t meant to be a place for AI writing. And they also tend to be the ones who charge for access, and aren’t producing anything worth having.
unique link to this extract
Hiker trekked ten miles to safety after impaling himself on pole • The New York Times
Remy Tumin:
»
David Cifaldi and two of his friends had been hiking for about seven hours up Montana’s tallest mountain on July 20 when he slipped on a rock and fell. He saw one of his trekking poles on the ground, but the second one was nowhere to be found.
Then he rolled over. Something wasn’t right.
“Oh,” he recalled. “This thing is sticking out of me.”
Mr. Cifaldi, 32, had impaled himself on the pole. He immediately went into nurse mode: Mr. Cifaldi is a wound nurse at St. Vincent Regional Hospital in Billings, Mont.
The pole was jutting at an angle into his left latissimus dorsi muscle, below his armpit, and out through his back. He wasn’t in pain, there was no blood, he was able to take deep breaths and could feel the trekking pole through his skin. He was two to three miles from the summit of Granite Peak, elevation of about 12,800 feet.
The three friends had started their ascent at 2 a.m. with plans to camp overnight at Froze-to-Death Plateau and then summit the mountain. Mr. Cifaldi had made it to the plateau a number of times, but never to the summit. His friend Jesse Ross, who had reached the top successfully, was leading Mr. Cifaldi and another friend, Brad Reich, up the mountain. They made it to the plateau by sunrise.
“As you get up there, you feel like you’re in, like, ‘The Lord of the Rings,’ or something like that,” Mr. Cifaldi said in an interview on Monday. “It’s just mountains and green.”
Everyone was cruising at that point. The weather was perfect — mid-60s and sunny, no wind — and the summit was in reach. Then Mr. Cifaldi lost his footing.
While one friend sent an SOS signal on a GPS phone, the other took pictures of the wound so Mr. Cifaldi could assess it. “I was pretty confident it was not in my chest or lung,” he said. Hoping to avoid an expensive lift off the mountain via helicopter, he walked around and found he wasn’t lightheaded and told his friends, “I think we can get out of here.”
And so they went.
«
Hoping to avoid an expensive lift off the mountain via helicopter. A classic of American health care. And an amazingly lucky person.
unique link to this extract
How the 1980s drum sound came about by accident • Far Out Magazine
Reuben Cross:
»
In 1979, while [drummer Phil Collins] was working alongside his former Genesis bandmate Peter Gabriel on his third studio album, recording engineer Hugh Padgham noticed that the microphone suspended from the ceiling of the live room to communicate with the band had accidentally been picking up the sound of Collins’ drums. Due to the microphone’s heavy compression, it was filtering out parts of the recording through its application of a noise gate.
The unusual sound that they had captured fascinated Gabriel, and he decided to try and recreate the sound with Collins once more, with the drum beat that opens his 1980 album Peter Gabriel (also known as Melt or PG3) on the track ‘Intruder’ making use of this newly discovered gated reverb sound. While this isn’t exactly one of the best-known uses of it, Collins’ love for the sound led him to use it on later releases, and with the drum fill that comes almost four minutes into ‘In the Air Tonight’, the gated reverb drum sound was propelled into the spotlight.
Drums in pop and rock music prior to this had always had the same dry sound that they would have had in the studio, but it was this moment of accidental innovation that drastically changed that, with several artists choosing to adopt the same recording and production technique to emulate the drum sound that Gabriel and Collins had achieved first on ‘Intruder’. Throughout the rest of the ‘80s, it felt almost impossible to hear a song on the radio that wasn’t using gated reverb on the drums, with artists such as Daryl Hall & John Oates and Prince getting a lot of mileage out of the sound.
This then led future technologies to automatically include the gated sound as a recording preset, allowing artists to easily reproduce the same effect on their own recordings with a much simpler press of a single button. However, the sound became so oversaturated within pop music that it only seemed to be a short-lived fad that fell out of fashion at the turn of the decade. While recordings in the ‘90s slowly began to favour dry drums once again, the gated reverb sound had already made its mark on the history of pop music and inadvertently defined the sound of the ‘80s for better or worse.
«
Noticing little things that others would throw away as intrusive or spoiled is a hallmark of discovery. In a way, the gated drum was like the music business’s penicillin in a petri dish.
unique link to this extract
AI powers citizen-led earthquake response for Venezuela • Rest of World
Daniella Fernández and Andrea Paola Hernández:
»
“Who knew when the government was going to respond?” Jorge Bastidas, a 31-year-old Venezuelan programmer who lives in Buenos Aires, told Rest of World. “We decided to take action.”
Bastidas and his team of six created Desaparecidos Terremoto Venezuela, a website for citizens to report, identify, and reunite with missing family and friends. Using Claude Opus 4.8, and facial recognition software donated by Mexican company Lab-Co, Bastidas designed the site to load quickly, without requiring users to register or download an app. It received more than 30,000 missing-person reports in the first two days, he said.
Without AI, “it would have taken me about 24 hours without rest to build something that took me three hours,” Bastidas said.
The earthquakes killed more than 3,800 people, and destroyed homes and hospitals. Telecommunications infrastructure, including a submarine fiber-optic cable and mobile network towers, was severely damaged. Power outages added to the challenge of accessing the internet.
With a near-total absence of state support, Venezuelan citizens and diaspora members like Bastidas stepped in. Within hours of the quakes, volunteer developers in Miami, San Francisco, Chile, Argentina, and elsewhere used AI to quickly build websites to report missing persons, coordinate donation drives, assess structural damage to buildings, map hospital capacity, locate shelters, and reunite pets with owners.
Venezuelan developer Samuel Mariña in California created Ayuda en Camino, or “help on the way,” a website for aid organizations, volunteers, donation centers, and people in need. He built it in four hours, with Replit. With limited connectivity in the affected areas, a WhatsApp assistant can check donations and requests without accessing the site, Mariña told Rest of World.
The internet was “drowning in disconnected information. It was hard to know what people had versus what people needed,” he said. “We wanted to centralize the information and make sure the aid wasn’t wasted, and reached where it was needed.”
«
All-new Apple Home Hub reportedly launching as soon as October • MacRumors
Joe Rossignol:
»
Apple plans to release an all-new smart home hub at some point between October and early next year, according to Bloomberg’s Mark Gurman.
Siri AI is expected to be a core feature of the hub, which is designed to be placed in rooms throughout your house, such as a kitchen, living room, or bedroom. The report said the device will be available in tabletop and wall-mounted versions, with Apple apparently believing that many customers will buy more than one.
The hub will essentially combine a HomePod-like speaker with a 7in display. The report said the device runs an entirely new operating system based on the Apple TV’s tvOS platform, with the interface described as tvOS blended with watchOS and iOS. There will be a grid of icons, widgets, and apps, along with customizable clock faces.
With built-in facial recognition, the hub can identify who is looking at the display and show personalized information, the report said. As a user approaches the hub, or moves away from it, the size of the interface elements will change.
Other features mentioned include FaceTime video calls, home security monitoring, HomeKit accessory management, photo galleries, a calendar, music playback, and an intercom system for households with multiple Apple home devices.
«
Seven inches diagonally, and square? Or.. some other shape? The last hardware Apple launched was the Vision Pro. This might at least sell better than that, though it’s hard to know what the use case really is. Maybe people asking chatbots questions all the time is going to be more of a thing in the future?
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 