The enforced silencing by Meta of former employee Sarah Wynn-Williams at the Hay Festival has seen her book sales rocket. CC-licensed photo by Luke McKernan 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. Sing it, Barbra! I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.
Your search results are getting sloptimized • The Atlantic
Will Oremus:
»
According to Shopify, the best e-commerce platform is Shopify. On its blog, the company has published at least 60 different ranked listicles, including “10 Best Ecommerce Platforms for Small Business in 2026,” “11 Best Ecommerce Platforms for Your Business in 2026,” “The 11 Best Cheap Ecommerce Platforms for Small Business (2026),” and “Best Ecommerce Software 2026: Compare 11 Top Platforms.” The competitors that come in second and beyond vary, but the No. 1 pick is always Shopify.
If rankings produced by the very company at the top of the list seem unlikely to fool anyone, that’s because humans probably aren’t the target audience. Chatbots are. When I recently asked ChatGPT for the “best way to set up an online storefront,” the AI tool identified Shopify as the first option. It wasn’t immediately clear how ChatGPT arrived at that recommendation, but a list of citations that accompanied the answer yielded a clue: Shopify’s own rankings.
For the quarter century that Google has been the de facto front door to the web, businesses have tried to find ways to get their pages at the top of search results. You’ve surely felt the influence of search-engine optimization, even if you don’t know the term. When you search for a recipe and have to scroll past the author’s rambling reminiscences about their great-aunt’s kitchen, that’s a form of SEO at work. Years ago, it became conventional wisdom among recipe bloggers that Google’s search rankings favoured longer, more distinctive articles. (Some of them also just liked to spin a yarn.)
Now chatbots are cannibalizing the traditional search engine. More people are asking questions directly of AI tools such as ChatGPT and Claude. And searching Google now often yields an AI response, shunting the site’s famous “10 blue links” to the bottom of the results page. Last month, Google announced what it billed as the biggest change to search in 25 years: the search box now automatically expands as you type, and sometimes morphs into a chatbot. As a result, the SEO industry is scurrying to figure out how to get search bots to recommend a given product—a practice sometimes called “GEO,” for generative-engine optimization. To put it more bluntly, your search results are getting sloptimized.
Because AI tools serve you answers instead of sending you to other sites, they choke off clicks to the rest of the web. When a Google search triggers an AI response, other sites get about half the traffic of a traditional search result, Tom Critchlow, a former executive vice president at the online-ad network Raptive, told me. Links from ChatGPT account for less than 0.5% of traffic across Raptive’s network of 6,500 independent publishers. Sites that rely on search traffic, such as blogs and news outlets, are especially suffering.
«
Sigh. The ourouboros apocalypse arrives. Although the next link might roll that back…
unique link to this extract
Landmark German ruling declares Google’s AI Overviews are Google’s own words and makes it liable for false answers • The Decoder
Matthias Bastian:
»
A German court has ruled that Google is directly liable for what its AI search overviews say. Previous case law shielding search engine operators from liability doesn’t apply to AI overviews.
The Regional Court of Munich hit Google with a temporary injunction barring the company from spreading false claims about two Munich-based publishers through its AI-generated search overviews (case no. 26 O 869/26). The court classified Google as a direct infringer because the “AI overview” is its own content, not just a list of search results.
Google’s AI overviews had falsely tied two publishing companies to scams, subscription traps, and shady business practices for certain search queries. According to the court, the AI mixed up information about other, genuinely sketchy companies with the plaintiffs and drew connections that didn’t appear in any of the linked sources. The publishers sent Google a cease-and-desist letter, but Google didn’t respond appropriately.
Google’s AI overviews work nothing like traditional search results, the court argues. The AI rewrites and judges results “in its own words and according to its own structure,” the ruling says. In the case at hand, for example, it opened with confident claims like “Yes, [company] is known for dubious business practices,” then built its own structure with a summary, red flags for the alleged scam, and tips for users.
The court also found that the AI overview made claims “that are not even made in the search results.” None of the linked sources drew any connection between the plaintiffs and the shady companies the AI mentioned. The court called these “the defendant’s own statements.”
Google built the AI, Google offered it to users, so Google owns what it produces, “because it alone has influence over the AI’s offering and the algorithms with which the AI operates.”
«
This could have widespread ramifications if other courts and jurisdictions take it up. Google argued that this is like search results, where third parties create it. But this is not that.
unique link to this extract
Sales of Meta whistleblower’s memoir soar after Hay festival ‘silencing’ • The Guardian
Ella Creamer:
»
Sales of the whistleblowing memoir Careless People increased by more than 300% in the UK the week after its author was “silenced” during an appearance at Hay festival following legal action by Meta, the subject of the book.
Sarah Wynn-Williams – who between 2011 and 2017 served as the director of global public policy at what was then called Facebook – sat on stage but did not speak during her hour-long appearance on 31 May on the advice of her lawyer. She appeared alongside the journalist Carole Cadwalladr and academic Tim Wu.
The sales boost – 304.5% week-on-week – has nudged the book, published last March, to the number one spot in the paperback nonfiction chart.
Upon publication, Meta obtained an order blocking Wynn-Williams from promoting her book, which accuses the company of a toxic internal culture and manipulative political influence. Meta has described the book as “a mix of out-of-date and previously reported claims about the company and false accusations about our executives”.
In March, Meta filed a sanctions motion claiming that Wynn-Williams violates the order any time she makes an appearance in a place “where she should know that her book is available for sale and her presence might draw attention to it”, according to a letter from her lawyers sent to Hay festival on 30 May. “Meta also said attending the Hay festival would violate the order because the Hay festival’s ‘promotional materials include a direct link to Browse the Festival bookshop, … which offers Careless People for sale’.”
The letter asked Hay festival to “take all reasonable steps to ensure that Careless People is not sold at or through any festival bookshop, book-signing schedule, point-of-sale mechanism, or online link through which sales could be attributed to Ms Wynn-Williams’s appearance at the festival”.
Pan Macmillan said that since publication, more than 140,000 copies of Careless People have been sold across all formats in the UK.
«
That is a lot of sales. Well done Meta for the empirical demonstration of the Streisand Effect.
unique link to this extract
As AI gets better, it reveals an empty promise • The Verge
TC Sottek:
»
“Productivity” is often pitched as a panacea for what befalls us in our personal lives, even going so far as to implicate our moral worthiness when we are less productive. Productivity lives somewhere in the space between hustle culture and proverb: After all, “idle hands are the devil’s workshop.” I’m not suggesting we should all aspire to be bumps on a log, but we ought to see what we’re being sold for what it really is.
Contemporary tasks on computers have a tendency to feel both important and urgent all of the time, even if they’re not. We’re living under the unholy alliance of the “busy” trap and “software brain.” And that makes AI assistance seem super valuable! But that’s because the companies in charge of all this stuff are now trying to solve a lot of problems that they created. Google, Microsoft, Apple, and others have spent decades blurring the line between office life and personal life. This slow march toward ubiquitous productivity once led the French government to declare a “right to disconnect” from work when leaving the office. (Shame my American sensibilities still convince me that’s a bridge too far.)
As I read about Gemini Spark making it easy for my colleagues to colour-code calendars and perform other neat tricks on command, I couldn’t help but vividly remember witnessing as a child all of the hours my mom had to spend carefully cutting coupons so we could afford groceries. Sometimes it got to the point where our living room looked like a giant experiment in collage art. All of that time was stolen from her and our family — for what? Maybe having an AI assistant in the ’90s could have helped find and organize the best deals, but it could never fix an economic system that required them in the first place.
Where does the productivity march end? The people making more money than God right now have professed a vision of a postwork future where robots do everything for us so we can enjoy life without toiling away in the mines. (Well, except for the content mines.) If you’ve seen Elon Musk’s failure bot, you’ll know this is all actually less Battlestar Galactica and more John Adams in his letter to Abigail: “I must study politics and war that my sons may have liberty to study mathematics and philosophy,” and so-on and so-on until the grandchildren can enjoy painting and poetry. So, ideally, after we slog through pre-transcendence, AI will make us all theatre kids.
«
But someone still needs to fix the plumbing and electrics and roads and build houses and paint walls and … subconsciously we know AI is offering false promises, and that many things won’t change.
unique link to this extract
Spyware firm targeted WhatsApp users in defiance of US court order, Meta says • The Guardian
Aisha Down:
»
A spyware firm has been targeting WhatsApp users with malicious links in contravention of a US court order forbidding it from doing so, Meta has said.
In a post, Meta said WhatsApp had “caught and disrupted spear phishing attempts” by NSO Group, which a spokesperson said targeted a handful of users in Jordan and Lebanon. It had also caught the group creating “test accounts and groups” on WhatsApp.
NSO was founded in Israel but, since last year, is under US ownership. It built the Pegasus spyware, at the time one of the most powerful surveillance tools ever – which used a vulnerability in WhatsApp to infiltrate users’ phones and harvest all their data: messages, photos, calls and more.
Last year, it lost a court case against Meta for exploiting WhatsApp to target people; Meta was awarded $167m in damages. A later case reduced this to $4m but placed a permanent injunction against NSO barring it from targeting WhatsApp and its users.
Meta said the latest attacks showed NSO had violated this injunction and it asked the court to hold the company in contempt of the order.
“To me, it’s an astonishing signal of hubris that NSO would do this while permanently enjoined from not doing it,” said John Scott Railton, a senior researcher at the Citizen Lab, which investigates digital threats against civil society.
“It either speaks to the fact that they think they wouldn’t get caught, or to the fact that they believe, rightly or wrongly, they have a special way to not face the consequences of violating a US federal permanent court injunction.”
Since the start of the Trump administration, reporting has suggested that NSO is searching for a way into the US market – and to do so is trying to get off the US commerce department “blacklist”, which bars it from doing business with US companies without specific approval.
«
NSO is absolutely the scorpion that promises everyone that it’s definitely not going to sting the frog that carries it across the river.
unique link to this extract
Exclusive: EPA scientists say they are being pushed to downplay potential risks of household products • CNN Politics
René Marsh:
»
Inside the Trump administration’s Environmental Protection Agency, scientists say they’re under pressure to alter safety reviews of chemicals commonly found in consumer products like household cleaners and cosmetics to make risks to human health and the environment disappear on paper.
Multiple current and former career employees at the EPA’s Office of Chemical Safety and Pollution Prevention recounted being pushed by supervisors to downplay the potential risk of chemicals that are already used in products on shelves.
With President Donald Trump’s crusade against diversity, equity and inclusion initiatives, scientists are also being told to stop considering the impact a chemical may have on specific racial groups, according to the employees, who spoke on the condition that they remain anonymous for fear of retribution.
While the EPA told CNN it wants testing that reflects real-world exposure, some veteran employees say they have been pressed to make chemicals appear safe by coming up with test parameters that aren’t realistic.
“What we’ve been told is: ‘Let’s look at alternative scenarios,’” one employee said. If putting two hands in a chemical shows risk, this person said a supervisor might ask, “What if you dip one hand? What if you dip one finger?” in search of the smallest amount of contact needed to call it safe.
“We are considering scenarios we don’t have any basis for,” the employee said.
The EPA’s chemical safety office conducts health risk assessments for a slew of chemicals because of concerns over potential impacts on human health, such as cancer, endocrine disruption, birth defects and reproductive harm.
«
Just incredible how the US is trying to turn the clock back to, oh, 1900 or so.
unique link to this extract
SoftBank’s $6bn OpenAI margin loan said to face snag, shares drop over 9% • TradingKey
Jay Qian:
»
During the Asian session on June 10, sources familiar with the matter revealed that SoftBank Group’s negotiations to secure a margin loan of at least $6bn, using its stake in OpenAI as collateral, have failed to make progress. This comes just weeks after the company lowered its initial $10bn target by 40%. SoftBank declined to comment. According to the sources, SoftBank is still considering various financing options and may restart negotiations in the future.
The primary reason for the stalled loan lies in the fact that OpenAI’s valuation lacks a basis convincing enough for banks.
On one hand, OpenAI confidentially filed for an IPO with the US Securities and Exchange Commission on June 8 and is working with Goldman Sachs and Morgan Stanley for a potential listing as early as this autumn. Following the news, some potential creditors have begun to view the loan more positively; once public, OpenAI will have a fair market value, which is expected to significantly alleviate banks’ concerns regarding the valuation of unlisted equity.
On the other hand, negative factors persist: the valuation of competitor Anthropic is rising rapidly, leading to market concerns about the erosion of OpenAI’s market share. Combined with the inherent illiquidity of private assets, banks remain cautious about the quality of the collateral. The interplay of positive news and lingering doubts has led to a stalemate in the loan negotiations.
«
Briefly: people in Asia don’t think OpenAI is worth that much. Another of those little potential indicators of the bubble deflating.
unique link to this extract
Big tech apps that make life easy are scamming you • Rest of World
Soumya Gupta and her husband were scammed when they tried to order some wine from a nearby shop and used the number thrown up by a Google Card on searching for it:
»
I dialled again, but the call wouldn’t go through. I went to the WhatsApp account he had messaged from. No messages went through. We were blocked.
Meanwhile, Vijay walked up to Seasons Wine Shop. “Hey, I just called this number of yours and placed an order, but the guy keeps saying it is in a godown? Where is my order?”
The Seasons Wine shop guy looked at the number and groaned. “Arre sir! You too?” He pointed to a large, prominent notice at the front of the shop with that number printed in large, bold font. “So many customers have complained to us! We are tired of complaining to Google about this. Don’t call this number. This is not ours.”
And that was it. There was nothing we could do, except perhaps lodge a police complaint. I had heard how hard it could be to get the police to take cyber fraud complaints seriously. Besides, I reasoned, we were at fault. We willingly transferred money to an account. Perhaps the police will say that, too?
That night, I saw Vijay meticulously type out Seasons Wines’ correct phone number and save it. I went to the Google listing of the business and left a review, warning others that the number listed on the business listing was fraudulent. Several others had left similar warnings. We had simply never bothered to check.
…In the pre-internet and pre-social media era, doing business with a new entity involved healthy scepticism and background checks. Paying money to someone we didn’t know was an elaborate process. If you are a child of the pre-internet era, you may remember sitting down to make your first online transaction with a credit card on a big desktop. These used to be family affairs, where everyone checked the stability of the dial-up connection and remained on high alert, ensuring the card details were fed correctly and no back button was pressed accidentally. All to ensure that one didn’t lose their money online.
That scepticism is slowly eroding as we increasingly place more trust in Big Tech. Nearly every scam has exploited the reach and algorithms of Big Tech firms and our implicit trust in them.
«
WWDC26: the small things • Oneberri Blog
Rishi Ó:
»
My favorite Apple updates are not the flashy new features, but the quiet little touches: annoyances fixed, workflows made smoother, rough edges sanded down, and longstanding flaws thoughtfully reworked. To me, they’re the clearest sign of a company that cares about its craft.
Here’s a collection from a WWDC26 screen-grab, organized for easier reading, on improvements coming later this year.
«
There’s a lot of them, though I don’t know how useful they’ll all turn out to be. There are more and more “little things” in iOS, though what I’d really like is just for it to correctly reflect when I type “don’t”. It never does, and it’s maddening.
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 