Hard drive maker Seagate has been fined $300m – payable in 20 easy instalments – for flouting an export ban to Huawei. CC-licensed photo by Kenming Wang 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 at the Social Warming Substack at about 0845 UK time: it’s about the site we don’t seem to talk about any more. Free signup.
A selection of 10 links for you. Ticked off. I’m @charlesarthur on Twitter. On Mastodon: https://newsie.social/@charlesarthur. Observations and links welcome.
BuzzFeed News defined the 2010s • The Atlantic
Charlie Warzel worked at Buzzfeed News (which is being closed down) from 2013 to 2019:
»
One can attribute the site’s cultural relevance, the industry enthusiasm around the work, and even the rivalries and haters to BuzzFeed News’s unofficial mission: to report on the internet like it was a real place, and to tell stories in the honest, casual tone of the web. At the time I joined, this was, if not a new kind of journalism, certainly an updated model for seeking out stories—one that’s now been fully absorbed by the mainstream. At its simplest, it might have meant mining a viral tweet or Reddit thread for ideas, but more often than not, it meant bearing witness to the joy, chaos, and horrors that would pour across our timelines every day and using them as a starting point for real reporting. It meant realizing, as I and my colleagues did, during the on- and offline manhunt for the Boston Marathon bombers, that a new culture of internet vigilantism was beginning to take hold in digital communities and that the media no longer unilaterally shaped broad news narratives.
Reporting on the internet like it was a real place led some of my colleagues to peer around corners of our politics and culture. In 2015, Joseph Bernstein outlined the way that “various reactionary forces have coalesced into a larger, coherent counterculture”—a phenomenon bubbling up in message boards such as 4chan that he called a “Chanterculture.” To read the piece now is to see the following half decade—reactionary MAGA politics, Trump’s troll armies, our current digital culture warring—laid out plainly. The Chanterculture story is a BuzzFeed News archetype: Movements like this weren’t hard to see if you were spending time in these communities and taking the people in them seriously. Most news organizations, however, weren’t doing that.
People afflicted with Business School Brain who didn’t understand BuzzFeed News (including one of the company’s lead investors) often described it like a tech start-up. This was true only in the sense that the company had an amazing, dynamic publishing platform—a content-management system that updated almost daily with new features based on writer input. But the secret behind BuzzFeed News had nothing to do with technology (or even moving fast). The secret was cultural. Despite the site’s constant bad reputation as a click farm, I was never once told to chase traffic. No editor ever discussed referrals or clicks. The emphasis was on doing the old-fashioned thing: finding an original story that told people something new, held people to account, or simply delighted. The traffic would come.
«
It’s so hard to make money from news online. There’s a slew of sites – many documented here over the years – which have tried and, regrettably, failed. Seems like the best way to make a small amount of money in online news is to start with a large amount of money.
US resident? Claim your piece of the $725m Facebook settlement if you used the platform during the last 15 years • Business Insider via Yahoo News
Aaron McDade:
»
Anyone in the US who’s had a Facebook account in the last 15 years can now submit a claim for their share of a $725m settlement from the Cambridge Analytica privacy-class action lawsuit.
Current or former Facebook users can submit claims through a website for the lawsuit by the August 25 deadline. The exact payment amounts per person will depend on how many claims are submitted.
A hearing is scheduled for September 7 when a judge is expected to approve the final details of the settlement. Unless other appeals are filed to delay the case, the site says the court will approve the settlement, and payments “will be distributed as soon as possible.”
Anyone who lived in the US and had a Facebook account from May 24, 2007, through December 22, 2022, is eligible to submit a claim, even if the account is no longer active. The claim form, which takes a few minutes to fill out, asks for a name, address, and email associated with the account, as well as when it was last used if the account is no longer active.
The settlement can be paid out in a variety of ways. Claimants can choose from a prepaid Mastercard, direct deposit to a bank account, or digital payment apps like PayPal, Venmo, and Zelle.
According to the site’s FAQ section, users also can opt out of eligibility for the settlement, which would allow them to retain the rights to be involved in potential future lawsuits over the claims involved in the Cambridge Analytica case.
«
If you’re in the US, get your claim in! The longer you were a Facebook user, the more you could be in line for. Though it’s probably not going to pay for your next dream home. More like a few cups of coffee.
(I expect there’s a similar case in the works in the UK, but don’t have any knowledge of one.)
unique link to this extract
Apple tester claims to be ‘blown away’ by AR/VR headset, says there was giant development leap • MacRumors
Juli Clover:
»
Leaker Evan Blass, who has provided accurate insight into Apple’s plans in the past, claims to know a person who has had opportunities to “demo” the headset. Blass said that over the course of the last few months, the tester has gone from “lamenting its ‘underwhelming’ capabilities” to being “blown away” by the experience and the hardware.
“The leap they’ve made since [late last year] is giant,” the Apple tester told Blass. “I was so skeptical; now I’m blown away in a ‘take my money kind of way,'” they said. Blass shared the details on his Twitter account, which is private.
Apple has been working on the AR/VR headset for years now, and its debut has been pushed back multiple times as the company has aimed to solve development issues with the design and the software. Apple is now ready to preview it, and is expected to do so at the Worldwide Developers Conference.
Back in March, The New York Times reported that several Apple employees it had spoken to were skeptical about the headset’s potential for success. The employees have questioned whether the headset is a “solution in search of a problem” and if it is “driven by the same clarity” as other Apple devices.
Apple CEO Tim Cook in April said that with everything the company has done, there have always been “loads of skeptics.” It comes with the territory of doing “something that’s on the edge,” Cook said.
The AR/VR headset is shaping up to be similar to the Apple Watch in terms of early functionality. It will be expensive at over $3,000, and limited in usefulness to begin with.
«
I’m suspicious of the existence of this tester altogether, but especially that they’ve gone from “blah” to “wow” in the course of a few months. It wouldn’t be impossible for someone to be feeding lines that they think the public discourse wants. I still think it would be crazy for Apple to introduce a headset at WWDC: the time, the economy, everything’s all wrong for it. I’ve seen far lesser products be hyped to the heavens with a certainty of a launch.. and then not appear.
unique link to this extract
Alphabet merges AI-focused groups DeepMind and Google Brain • CNBC
Jennifer Elias:
»
Alphabet is merging an internal Google Research team called Brain with DeepMind, a move designed to bring two groups focused on artificial intelligence closer together as the battle for AI heats up.
Google acquired DeepMind in 2014 for a reported $500m and has until now run it as an independent unit out of the UK. DeepMind has been one of Alphabet’s “other bets,” performing futuristic work, such as teaching computer systems to beat top-ranked players of the Chinese board game Go.
“Combining all this talent into one focused team, backed by the computational resources of Google, will significantly accelerate our progress in AI,” Alphabet CEO Sundar Pichai said in blog post Thursday.
Jeff Dean, who currently leads Google’s AI efforts, will be promoted and given the title of chief scientist at Google, reporting to Pichai. He’ll head up the “most critical and strategic” technical projects related to AI, the first of which will be a series of powerful, multimodal AI models.
The move marks Google’s latest reorganization in response to the rapid developments in AI, following OpenAI’s launch of the chatbot ChatGPT late last year. CNBC previously reported that Google reshuffled its Assistant organization to prioritize the company’s AI chatbot Bard.
“The pace of progress is now faster than ever before,” Pichai wrote. “To ensure the bold and responsible development of general AI, we’re creating a unit that will help us build more capable systems more safely and responsibly.”
DeepMind has been able to operate separately from Google’s core research, enabling it to move quicker on breakthroughs such as AlphaFold, which can predict 3D models of protein structures. The two divisions, DeepMind and Google Research, have also reportedly had tensions in the past, leading DeepMind to seek more independence.
DeepMind CEO Demis Hassabis will lead the development of “the most capable and responsible general AI systems,” Pichai said. That research, he added, “will help power the next generation of our products and services.”
«
Doesn’t sound like this is going to bring peace and light. More like they’re putting them into a sack to see who emerges victorious. Dean, from Brain, will report to Pichai, while Hassabis works on general AI systems? Sparks will fly.
unique link to this extract
‘What time is’ SEO: Google direct answers cut out publishers • Press Gazette
Dominic Ponsford:
»
Google appears to have called time on an SEO tactic that has delivered millions of dubious clicks for publishers: the “what time is” story.
Stories like “what time is Eastenders on tonight” supercharged the traffic growth of websites such as the Daily Mirror and the Telegraph a decade ago.
Written for Google, these stories take a trending search term and manage to confect a news story out of answering it. They tend to reverse the conventional architecture of a news story by burying the relevant information near the bottom of a story, so readers spend more time on the page before arriving at a simply-told answer.
The most notorious form of this article purports to answer a question and then fails to do so. It might have the headline “when will Film X be released on Netflix” only to reveal in the final paragraph that the release date has yet to be revealed.
Google sought to stop surfacing these sort of articles in search with its “helpful content” algorithm update of August last year.
«
Perhaps it will do the same on “all we know about the [unreleased product]” stories, but that may be too much to hope for. I tried one of the examples shown in the story – “grand national 2023 time” (the Grand National is an annual very big horse race; think Kentucky Derby big) – and got informative answers in the excerpted text on DuckDuckGo. Seems fair to publishers.
unique link to this extract
‘I’ve never hired a writer better than ChatGPT’: how AI is upending the freelance world • Forbes
Rashi Shrivastava:
»
Melissa Shea hires freelancers to take on most of the basic tasks for her fashion-focused tech startup, paying $22 per hour on average for them to develop websites, transcribe audio and write marketing copy. In January 2023, she welcomed a new member to her team: ChatGPT. At $0 an hour, the chatbot can crank out more content much faster than freelancers and has replaced three content writers she would have otherwise hired through freelancing platform Upwork.
“I’m really frankly worried that millions of people are going to be without a job by the end of this year,” says Shea, cofounder of New York-based Fashion Mingle, a networking and marketing platform for fashion professionals. “I’ve never hired a writer better than ChatGPT.”
Shea has not posted a job on Upwork since she discovered ChatGPT (though she still has five freelancers working for her). After it was released in November 2022, ChatGPT amassed more than 100 million users, sparked an AI arms race at companies like Microsoft, Google and Amazon and has given rise to a flurry of AI startups. And for small businesses looking to trim costs, the free tool can automate swaths of their operations, providing a cheaper alternative to freelance workers. Built on recent advances in generative AI, ChatGPT and its image-based sibling DALL-E 2 can carry out work that spans most of the freelancing spectrum, from writing articles and compiling research to designing graphics, coding and decrypting financial documents.
Now, freelancers who are less experienced and don’t offer specialized skills stand to lose their gigs, according to five clients Forbes interviewed. But rather than steering clear of the AI tool that could make them obsolete, more and more freelancers are relying on ChatGPT to do some if not all their work for them. Clients on job marketplaces like Upwork and Fiverr are being flooded with nearly identical project proposals written by ChatGPT. A bitter side effect: it’s making clients dubious of the authenticity of work turned in by freelancers and causing transactional disputes and mistrust in the freelancing community.
«
If you’ve never hired a writer better than ChatGPT, you’ve really been hiring some terrible writers. Or perhaps more accurately, you’ve been commissioning some terrible work that good writers can’t be bothered with. Nevertheless: the AI tsunami is coming.
unique link to this extract
Space Elevator
Neal Agarwal:
»
Welcome to the space elevator, the only elevator that goes to space.
«
Very fun illustration of the scale that everything’s at when you’re heading to space. You’ll be surprised by how high a bird can fly. A helicopter too for that matter.
unique link to this extract
See the websites that make AI bots like ChatGPT sound so smart • Washington Post
Kevin Schaul, Szu Yu Chen and Nitasha Tiku:
»
Chatbots cannot think like humans: They do not actually understand what they say. They can mimic human speech because the artificial intelligence that powers them has ingested a gargantuan amount of text, mostly scraped from the internet.
This text is the AI’s main source of information about the world as it is being built, and it influences how it responds to users. If it aces the bar exam, for example, it’s probably because its training data included thousands of LSAT practice sites.
Tech companies have grown secretive about what they feed the AI. So The Washington Post set out to analyze one of these data sets to fully reveal the types of proprietary, personal, and often offensive websites that go into an AI’s training data.
To look inside this black box, we analyzed Google’s C4 [“Colossal Clean Crawled Corpus”] data set, a massive snapshot of the contents of 15 million websites that have been used to instruct some high-profile English-language AIs, called large language models, including Google’s T5 and Facebook’s LLaMA. (OpenAI does not disclose what datasets it uses to train the models backing its popular chatbot, ChatGPT)
The Post worked with researchers at the Allen Institute for AI on this investigation and categorized the websites using data from Similarweb, a web analytics company. About a third of the websites could not be categorized, mostly because they no longer appear on the internet. Those are not shown.
…Kickstarter and Patreon may give the AI access to artists’ ideas and marketing copy, raising concerns the technology may copy this work in suggestions to users. Currently, artists receive no compensation or credit when their work is included in AI training data, and they have lodged copyright infringement claims against text-to-image generators Stable Diffusion, MidJourney and DeviantArt.
The Post’s analysis suggests more legal challenges may be on the way: The copyright symbol — which denotes a work registered as intellectual property — appears more than 200 million times in the C4 data set.
«
One-third of the training sites aren’t on the internet any more? That seems like a lot to me. Sadly, The Overspill doesn’t appear in the training corpus; some interloper called “overspillsite.wordpress.com”, which mostly seems to be motivational messages. Pah.
When is a photo not a photo? The looming spectre of artificially generated photographs • Vanity Fair
Fred Ritchin is dean emeritus of the the school at New York’s International Center of Photography:
»
In 1984, when photographers were still using film, I began exploring the early use of computers to undetectably modify photographs. In an article in The New York Times Magazine I wrote that “in the not-too-distant future, realistic-looking images will probably have to be labeled, like words, as either fiction or nonfiction, because it may be impossible to tell them apart. We may have to rely on the image maker, and not the image, to tell us into which category certain pictures fall.”
…In a news cycle often dominated by conspiracy theories and fake news, legitimate, unaltered photographs—instead of confronting us with realities from which we cannot look away (as happened during the Vietnam War and the Civil Rights Movement)—will more easily and automatically be rejected out of hand. It is not a coincidence, for example, that no single iconic image emerged to initiate or sustain a societal discussion about the 20-year war in Afghanistan, the longest in US history. Or that Western support for Ukraine, in its response to Russia’s brutal and ongoing invasion, was galvanized largely by the persistent video and online dispatches of Ukraine’s media-savvy president, Volodymyr Zelenskyy, rather than by a series of iconic photographs.
Certainly, other factors have contributed to the photograph’s reduced role as witness: the disappearance of the newspaper front page, the billions of competing images on social media. But now, rather than the photograph, it is the occasional amateur video, posted online along with a fuller background narrative—such as of the footage of the murder of George Floyd by a Minneapolis police officer—that manages to mobilize meaningful, broad-based response.
«
Seagate to pay $300m penalty for shipping Huawei 7m hard drives • Reuters
Karen Freifeld:
»
Seagate Technology has agreed to pay a $300m penalty in a settlement with US authorities for shipping over $1.1bn worth of hard disk drives to China’s Huawei in violation of US export control laws, the Department of Commerce said on Wednesday.
Seagate sold the drives to Huawei between August 2020 and September 2021 despite an August 2020 rule that restricted sales of certain foreign items made with US technology to the company. Huawei was placed on the Entity List, a US trade blacklist, in 2019 to reduce the sale of US goods to the company amid national security and foreign policy concerns.
The penalty represents the latest in a string of actions by Washington to keep sophisticated technology from China that may support its military, enable human rights abuses or otherwise threaten US security.
Seagate shipped 7.4m drives to Huawei for about a year after the 2020 rule took effect and became Huawei’s sole supplier of hard drives, the Commerce Department said.
The other two primary suppliers of hard drives ceased shipments to Huawei after the new rule took effect in 2020, the department said. Though they were not identified, Western Digital Corp and Toshiba were the other two, the US Senate Commerce Committee said in a 2021 report on Seagate.
«
Seagate’s latest financials (on Thursday!) show that it usually has about a 30% gross margin, and 16% operating margin. So a $300m fine on $1.1bn of drives essentially takes away its gross margin, leaving it revenue-neutral, not really fined. And the payment is being made in $15m chunks, quarterly, for five years.
Anyway, what’s Huawei doing for hard drives now?
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 