Apple Aims to Source All U.S. iPhones From India by End of 2026 ⇥ ft.com

Michael Acton, Stephen Morris, John Reed, and Kathrin Hille, Financial Times:

Apple plans to shift the assembly of all US-sold iPhones to India as soon as next year, according to people familiar with the matter, as President Donald Trump’s trade war forces the tech giant to pivot away from China.

The push builds on Apple’s strategy to diversify its supply chain but goes further and faster than investors appreciate, with a goal to source from India the entirety of the more than 60mn iPhones sold annually in the US by the end of 2026.

Obviously this is not fulfilling the stated aim of U.S. tariffs to onshore manufacturing of all kinds of products, including iPhones. But it seems unlikely to substantially lessen our global dependence on supply chains in China, either — in addition to individual parts sourced there, these reporters say factories in China are still responsible for “pre-assembled component sets” which are brought to India for final iPhone assembly.

Google Ends Support for Early Nest Thermostats, Degrading Their Functionality ⇥ arstechnica.com

Ryan Whitwam, Ars Technica:

As Google points out, these products have had a long life, and they’re not being rendered totally inoperable. Come October 25, 2025, these devices will no longer receive software updates or connect to Google’s cloud services. That means you won’t be able to control them from the Google Home app or via Assistant (or more likely Gemini by that point). The devices will still work as a regular dumb thermostat to control temperature, and scheduling will remain accessible from the thermostat’s screen.

Google stopped selling the second-generation Nest Thermostat in 2015. A ten-year lifespan may be acceptable by consumer electronics standards, it is not impressive for a thermostat. And, while a typical programmable thermostat might slowly fail on its own accord, it is disconcerting that Google can simply pull the plug on its smart features.

This is, however, a fairly graceful degradation of its functionality, all things considered. Nests will apparently continue to work like a programmable thermostat. Also, after Ecobee ended support for its earliest thermostat models last year, it sounded like newer models’ HomeKit integration would allow smart features to keep working when Ecobee discontinues those, too. It would be better if Google could have done something similar with the Nest. It is not encouraging for the smart home fantasy.

Sponsor: Magic Lasso Adblock: YouTube Ad Blocker for Safari ⇥ magiclasso.co

Do you want to block all YouTube ads in Safari on your iPhone, iPad, and Mac?

Then download Magic Lasso Adblock — the ad blocker designed for you.

Magic Lasso Adblock: YouTube Ad Blocker for Safari

As an efficient, high performance, and native Safari ad blocker, Magic Lasso blocks all intrusive ads, trackers, and annoyances — delivering a faster, cleaner, and more secure web browsing experience.

Magic Lasso Adblock is easy to set up, doubles the speed at which Safari loads, and also blocks all YouTube ads; including all:

  • video ads

  • pop up banner ads

  • search ads

  • plus many more

With over 5,000 five star reviews, it’s simply the best ad blocker for your iPhone, iPad, and Mac.

And unlike some other ad blockers, Magic Lasso Adblock respects your privacy, doesn’t accept payment from advertisers and is 100% supported by its community of users.

So, join over 350,000 users and download Magic Lasso Adblock today.

Making ‘Careless People’ Care ⇥ pluralistic.net

Cory Doctorow:

The important thing about Facebook’s carelessness is that it wasn’t the result of the many grave personality defects in Facebook’s top executives – it was the result of policy choices. Government decisions not to enforce antitrust law, to allow privacy law to wither on the vine, to expand IP law to give Facebook a weapon to shut down interoperable rivals – these all created the enshittogenic environment that allowed the careless people who run Facebook to stop caring.

The corollary: if we change the policy environment, we can make these careless people – and their successors, who run other businesses we rely upon – care. They may never care about us, but we can make them care about what we might do to them if they give in to their carelessness.

This is a prudent conclusion to draw from “Careless People”. It is tempting to wonder if things would be much different with others in charge, but what Meta represents today is not an individualized problem.

Apple and Meta Are Mad Over Minimal E.U. Fines ⇥ ec.europa.eu

The European Commission:

Today, the European Commission found that Apple breached its anti-steering obligation under the Digital Markets Act (DMA), and that Meta breached the DMA obligation to give consumers the choice of a service that uses less of their personal data. Therefore, the Commission has fined Apple and Meta with €500 million and €200 million respectively.

The two decisions come after extensive dialogue with the companies concerned allowing them to present in detail their views and arguments.

A Financial Times article last month suggested these fines would be “minimal”, but I suppose that is entirely relative. Apple and Meta could have been fined up to 10% of their annual worldwide revenue — earnings which total hundreds of billions of dollars each — so €500 million does, in fact, seem to be “minimal”. And that speaks to the power and size of these corporations. Call it a warning.

Jon Brodkin, Ars Technica:

“The European Commission is attempting to handicap successful American businesses while allowing Chinese and European companies to operate under different standards,” Meta Chief Global Affairs Officer Joel Kaplan said.

[…]

“Today’s announcements are yet another example of the European Commission unfairly targeting Apple in a series of decisions that are bad for the privacy and security of our users, bad for products, and force us to give away our technology for free,” Apple said. “We have spent hundreds of thousands of engineering hours and made dozens of changes to comply with this law, none of which our users have asked for. Despite countless meetings, the Commission continues to move the goal posts every step of the way.”

I am halfway through “The Big Myth” and it is striking to read so many of the same complaints from detractors of regulation no matter when or for what issue. Pick any of the past couple-hundred years and it is apparently a regulatory sweet spot, according to industry representatives at the time. These complaints feel derivative.

That does not mean all new regulation is justified, well-considered, or without negative consequences. However, it is not as though either Apple or Meta stumbled into the exact structure they did to comply with their interpretation of the law on paper while evading the Commission’s expectations. Nobody expects either company to “give away [their] technology for free”, in the words of an unnamed Apple spokesperson. One cannot, for example, run iOS apps on non-Apple system, and the company builds the cost of software and updates into each product’s price.

The Anti-Pirate Code ⇥ fedi.rib.gay

Melissa Lewis on Bluesky:

TIL: The 2000s piracy PSA used a font designed by the fantastic Just van Rossum, whose brother Guido created the Python programming language.

[…]

I reached out to van Rossum, and Xband Rough is indeed an “illegal clone” of FF Confidential; it’s just been around forever and is ubiquitous. I have no idea whether the PSA used the original, and felt too shy to ask!

“Rib”:

Naturally, it would be hilarious if the anti-piracy campaign actually turned out to have used this pirated font, so I went sleuthing and quickly found a PDF from the campaign site with the font embedded (https://web.archive.org/web/[…]150605_8PP_brochure.pdf).

So I chucked it into FontForge and yep, turns out the campaign used a pirated font the entire time!

They are more like guidelines, you could say.

On the ‘Streaming Is More Expensive Than Cable’ Complaint ⇥ birchtree.me

Matt Birchler:

A social media post flew by today where someone was complaining that they stopped paying $180 for cable and now they’re paying $200 for streaming apps. I see this sort of complaint a lot, and I find myself feeling like I’m taking crazy pills whenever I do: I had cable, I know how bad it was, and the new streaming world seems so much better to me, even if it is creeping up in price.

I always appreciate when people bring evidence to a discussion instead of basing their analysis entirely on vibes. Here, Birchler demonstrates the combined cost of streaming services is, for lots of people in the U.S., less than cable television.

I remember when the cost of cable was a routine complaint. You paid for all these channels, but you only watched a handful of things on three of them — and, so, would it not be better if we could just buy the channels we wanted?

I have nothing to support this, but I strongly suspect the rosy glasses reaction to cable in a streaming world is driven by two things. First, the lack of bundling feels worse. If you were told you could get a “complete streaming package” for $60 per month — including Apple TV Plus, Disney Plus, Hulu, Max, Netflix, Paramount Plus, Peacock, and Prime Video — that might sound like a better deal than subscribing to them individually for the same total cost. A single cable subscription made you feel like you were getting everything, while you now need multiple subscriptions if you like to watch a lot of stuff.

Second, as Birchler acknowledges, some stuff still is not available on streaming, and so some people still feel they must add a cable subscription, too. Because most of the big streaming services are owned by the same broadcasters as cable channels, it gives them an opportunity to double-dip.

Canadian Election Coverage and Meta’s Response to the Online News Act ⇥ thelogic.co

Martin Patriquin, the Logic:

More than 25 different Facebook accounts, some with Canadian-themed names, have promoted multiple fake news articles involving political actors, including a three-minute video of a deepfake CBC report fronted by anchor Rosemary Barton, in which a deepfake Musk, appearing on a deepfake Joe Rogan podcast, accuses a deepfake Carney of lying about a non-existent passive income scheme.

The posts are widespread on the Meta-owned platform, with 25 per cent of Canadians saying they’ve seen content on Facebook mimicking legitimate news in the last month, according to the Canadian Digital Media Research Network (CDMRN), which is part of a digital literacy coalition monitoring the federal election campaign.

Leyland Cecco, the Guardian:

More than a quarter of Canadians have been exposed to fake political content on social media that is “more sophisticated and more politically polarizing” as the country prepares to vote in a federal election, researchers have found, warning that platforms must increase protections amid a “dramatic acceleration” of online disinformation in the final weeks of the campaign.

In a new report released on Friday, Canada’s Media Ecosystem Observatory found a growing number of Facebook ads impersonating legitimate news sources were instead promoting fraudulent investment schemes, often involving cryptocurrency.

Rob Brown, reporter for CBC News in Calgary, interviewed Aengus Bridgman of the Media Ecosystem Observatory — previously linked — about Meta’s restriction of news links for Canadian users. This is one effect of the Online News Act, which required Google and Meta to pay news publishers for links on their platforms. Meta, in deciding not to, has effectively permitted only links which do not qualify as “news” to spread among Canadians on its platforms.

Meta may be a belligerent and frustrating company, but the Online News Act remains a blight. In a stopped-clock credit to the Conservatives, they pledge to reverse it, something which all parties ought to consider. We can do better with policies that do not, in practice, discourage the sharing of links to news stories.

The Paid A.I. Hype Guys ⇥ ft.com

Siddharth Venkataramakrishnan, Financial Times:

One hint that we might just be stuck in a hype cycle is the proliferation of what you might call “second-order slop” or “slopaganda”: a tidal wave of newsletters and X threads expressing awe at every press release and product announcement to hoover up some of that sweet, sweet advertising cash.

That AI companies are actively patronising and fanning a cottage economy of self-described educators and influencers to bring in new customers suggests the emperor has no clothes (and six fingers).

The (verified) X accounts producing threads of links to bad A.I. products, boosted by dozens of replies from other verified X accounts, are annoying spam, but seem relatively harmless. But the newsletters profiled here are curious.

One is Rowan Cheung’s Rundown AI newsletter, which Cheung bills as the “world’s most read daily AI newsletter”, which is different from the other newsletter in the article, Zain Kahn’s Superhuman AI, the “world’s biggest AI newsletter”. As alluded to by Venkataramakrishnan but not expanded upon, both attract some heavy-hitting sponsors. The Rundown was recently sponsored by Salesforce, HubSpot, Sana, and Writer, while Superhuman’s recent sponsors include companies like HubSpot, Sana, and Writer — no Salesforce.

While Cheung’s newsletter is mostly A.I. boosterism with sponsorships, there is a block near the end of each issue for a sibling product: the Rundown University. The first thing you need to know about it is that it is not a university, obviously. It is online training through individual “courses” offered at $50 apiece with individual lessons on using A.I. tools, some of which — like Gamma and Zapier — happen to have sponsored of the newsletter. Or, if you want access to all the “courses” plus workshops, tutorials, and some kind of group chat, you can get all that for just a penny shy of $1,000 per year. Just above the footer sits a carousel of inspirational quotes about A.I. from people like Sam Altman, Jeff Bezos, Mark Cuban, and Elon Musk. None are actually specific to Rundown “University”, just vibes about the importance of A.I. and its impact. It all feels a bit much.

I am fascinated by the knock-on effects of a hype cycle, and A.I. has produced some magnificent examples — including those above. It is illuminating to search Google for a phrase like intitle:"how" intitle:"is using ai to" and witnessing what is ostensibly breathtaking innovation across industries. A 2023 Guardian story says the “oldest surviving newspaper in the world” — untrue and in spirit only — is using ChatGPT to generate articles from council meeting minutes. According to the Harvard Business Review, construction companies are using large language models to, among other things, summarize documents. Stitch Fix is, according to a disguised ad in Vogue Business, using A.I. to detect trends.

There are countless examples of articles like these illustrating how companies are benefitting from the glow of hype while contributing to it. None of this is to say it is necessarily unearned — maybe detecting money laundering really is more reliable when you run it through A.I. processes. But I remember stories like these in the days when the hot new thing was called “machine learning” or, before that, “big data”. I would not be surprised if these technologies are truly beneficial yet it is hard not to feel the weight of hype and the marketing people behind it all.

Bluesky Verification Is Decentralized in Theory ⇥ steveklabnik.com

Steve Klabnik:

The underlying [verification] protocol is totally open, but you can make an argument that most users will use the main client, and therefore, in practice it’s more centralized than not. My mental model of it is the bundled list of root CAs that browsers ship; in theory, DNS is completely decentralized, but in practice, a few root CAs are more trusted than others. It’s sort of similar here, except there’s no real “delegation” involved: verifiers are peers, not a tree.

This core design allows Bluesky to adjust the way it works in the future fairly easily, they could allow you to decide who to trust, for example. We’ll see how this feature evolves over time.

The way Bluesky described verification made it sound like a centralized service, but it is not. I have corrected my post.

Bluesky Adds Verification Badges ⇥ bsky.social

In a post attributed to the Bluesky Team, the company announced a familiar verification method with a twist:

In 2023, we launched our first layer of verification: letting individuals and organizations set their domain as their username. Since then, over 270,000 accounts have linked their Bluesky username to their website. Domain handles continue to be an important part of verification on Bluesky. At the same time, we’ve heard from users that a larger visual signal would be useful in knowing which accounts are authentic.

Now, we’re introducing a new layer — a user-friendly, easily recognizable blue check. Bluesky will proactively verify authentic and notable accounts and display a blue check next to their names. Additionally, through our Trusted Verifiers feature, select independent organizations can verify accounts directly. Bluesky will review these verifications as well to ensure authenticity.

The domain-based verification Bluesky already implemented was a good start, but incomplete. My profile is connected to my personal domain but, even with that single line of text, there is no obvious indicator. You can see the problem on Bluesky’s own profile which tells you to “check username👆” to confirm it is, indeed, the real one.

Then there is the problem of public personalities. Maybe a musician or actor wants to confirm their profile is really them, but is domain-based verification really the best way for them? Does Pedro Pascal have a dot-com? What about a journalist who could “verify” an account connected to an employer, but they also want a separate personal account? Both ought to be verified, but only one is employer-connected. Then, when they leave that employer, they will need to change their username, and Bluesky does not automatically redirect moved profiles. (Followers, posts, and so forth are moved to the new profile, but any links to posts made under the previous profile are not redirected.)

A verification checkmark is not an offensive idea unto itself, but it comes with a new set of concerns. First, what does it mean? Bluesky says it is for “authentic and notable accounts”, but what that means is not clear. Twitter used to assign itself an authoritative role in determining which accounts fit similar criteria. In doing so, it helped make this badge a bizarre mark of status. Now, Meta and X let you simply buy a badge to confirm an account belongs to the person or organization claimed. If the goal is to avoid controversy with notable figures, surely self-verification and manual profile confirmation achieve those purposes. But a parody account is also legitimate in its own way, right?

Second, while it makes sense to have a quick identifier of accounts susceptible to impersonation, the badge also comes with risks. If such an account is hijacked, for example, its posts carry additional weight.

Bluesky seems like it is trying to solve some of these issues by allowing others to vouch for the authenticity of an account. In theory, this minimizes its own role in being the voice of authority. In its example, the New York Times’ account can verify the authenticity of others’ accounts. But Bluesky still plays a central role in this process — it appears to determine whether an account can become a Trusted Verifier, and its moderators confirm each verification request from third-party verifiers. I take it this latter requirement is necessary because it is not clear to me whether there are any limits to which accounts a Trusted Verifier can approve. Still, it means Bluesky is the final say in which accounts belong to the person or organization claimed.

Update: I was wrong, according to Bluesky technical advisor Jeremy Johnson:

anyone can publish a verify record, any app can choose who to display verifies from. And as this thing bakes a bit we’re going to make it easier to pick different verifiers in our app.

If you were to run your own app view then you definitely don’t have to care at all what Bluesky PBC does here, just collect all the app.bsky.graph.verification records and go ham.

It sounds like I could verify my own accounts — or, indeed, vouch for any account — but it is up to the client how it wants to display that verification attempt. Bluesky itself is only trusting itself and select others for now.

The Charm of Movie Mistakes ⇥ fxrant.blogspot.com

Todd Vaziri:

I’m particularly obsessed with moments that reveal the craft and artistry of the magic trick of a shot that slightly shatters the illusion of cinema. These revealing moments have been in movies since the dawn of cinema, and are everywhere (if you know exactly where to look).

I have always loved spotting “mistakes” in movies. Not from the perspective of ha ha, you messed up, but more because of exactly what Vaziri describes here: they show the little moments of craft. These ones are charming.

Sponsor: Magic Lasso Adblock: Incredibly Private and Secure Safari Web Browsing ⇥ magiclasso.co

Online privacy isn’t just something you should be hoping for – it’s something you should expect. You should ensure your browsing history stays private and is not harvested by ad networks.

Magic Lasso Adblock: No ads, no trackers, no annoyances, no worries

By blocking ad trackers, Magic Lasso Adblock stops you being followed by ads around the web.

As an efficient, high performance and native Safari ad blocker, Magic Lasso blocks all intrusive ads, trackers, and annoyances on your iPhone, iPad, and Mac. And it’s been designed from the ground up to protect your privacy.

Users rely on Magic Lasso Adblock to:

  • Remove ad trackers, annoyances, and background crypto-mining scripts

  • Browse common websites 2.0× faster

  • Block all YouTube ads, including pre-roll video ads

  • Double battery life during heavy web browsing

  • Lower data usage when on the go

With over 5,000 five star reviews; it’s simply the best ad blocker for your iPhone, iPad, and Mac.

And unlike some other ad blockers, Magic Lasso Adblock respects your privacy, doesn’t accept payment from advertisers and is 100% supported by its community of users.

So, join over 350,000 users and download Magic Lasso Adblock today.

Apple’s Rebranded Apple Ads Probably Forecasts More Advertising ⇥ digiday.com

Benjamin Mayo, 9to5Mac:

Apple today announced that it has officially rebranded Search Ads as ‘Apple Ads’, reflecting the expanding scope of Apple’s advertising business.

When Apple first launched ads in the App Store, they were only shown as promoted search results, hence the name ‘Search Ads’. But the company now offers advertising placements in many more places beyond just the Search tab, so the old name had become a bit anachronistic.

Seb Joseph and Krystal Scanlon, Digiday:

The timing matters. Apple has spent years building its defenses, designing its own chips, overhauling supply chains — but tariffs on Chinese imports, especially under Trump’s return to the campaign trail, are reintroducing uncertainty. Some products have dodged the impact, others haven’t. And with hardware margins already tight, Apple needs new ways to protect its bottom line.

I do not think it is nitpicking to dispute calling margins of 35% on hardware “tight”.

Still, Apple is at risk of being squeezed in both hardware and software — to the extent a corporation the size of Apple can actually be squeezed, of course. Its hardware made in China is now exempt from Trump’s escalating and volatile tariff demands, but that could obviously change if Tim Cook is insufficiently grateful. Meanwhile, its Google search deal with 100% margins is at risk of being disallowed.

The advertising market is an obvious but dispiriting place to find money. Apple already has a few spots available in the App Store, and in News and Stocks, and I could see the potential for placements in apps like Maps, Music, Podcasts, and Wallet.

This rebranding and the increasingly important Services revenue charts look like a forecast for more ads. I hope I am wrong. I do not use News or Stocks, but the ads in the App Store are already a gross exploitation of Apple’s position owning these platforms. It is the kind of move that encourages understandable regulatory scrutiny.

Microsoft Is Bringing Recall Back to Windows ⇥ arstechnica.com

Dan Goodin, Ars Technica:

Following months of backlash, Microsoft later suspended Recall. On Thursday, the company said it was reintroducing Recall. It currently is available only to insiders with access to the Windows 11 Build 26100.3902 preview version. Over time, the feature will be rolled out more broadly.

Kevin Beaumont:

After a bunch of discussions with a bunch of folks, it’s pretty clear there’s been zero published research on Copilot Recall in almost a year – all the news articles have just reprinted Microsoft’s talking points saying it is secure now.

Beaumont says there are improvements to the onboarding process and security, but also found there is still no way for applications to opt out of being recorded, and it still captures private information like Signal chats and credit card numbers. Oh, and Beaumont may have found the decryption key. I see the utility in a feature like this, but I am not sure that outweighs the risks for the user or anyone they communicate with.

Google Loses Online Advertising Monopoly Case ⇥ axios.com

Scott Rosenberg and Sara Fischer, Axios:

Google’s dominance of the online advertising and ad-tech markets violates U.S. antitrust laws, a federal court ruled Thursday, in a decision that could scramble the massive digital ad universe.

[…]

The court drew a distinction between the markets for advertising exchanges and ad servers, where it found Google has an illegal monopoly, and the general market for display ads online, where it found Google does not.

Emma Roth and Lauren Feiner, the Verge:

“We won half of this case and we will appeal the other half,” Lee-Anne Mulholland, Google’s vice president of regulatory affairs, said in a statement to The Verge. “The Court found that our advertiser tools and our acquisitions, such as DoubleClick, don’t harm competition. We disagree with the Court’s decision regarding our publisher tools. Publishers have many options and they choose Google because our ad tech tools are simple, affordable and effective.”

If the conclusion of this case is held after what I assume will be a lengthy and expensive appeals process, the corrections for this and Google’s illegal search monopoly will have a noticeable effect on the web.

Bluesky, Censorship, and Country-Based Moderation ⇥ fediversereport.com

Laurens Hof, the Fediverse Report:

There is a technical story, of how Bluesky and the AT Protocol (ATProto) do composable moderation for specific countries. But this is not just an interesting technology, it has implications of government censorship more broadly. Not only is the Turkish censorship of accounts even easier to sidestep, it also allows for new ways to highlight and create visibility for the content that the Turkish government wants to be hidden. To explain how that all works, first a closer look at how moderation works on Bluesky.

I understand why some would disagree with Bluesky’s approach and prefer the way of Mastodon. But I do not think Bluesky’s handling is invalid or necessarily inferior — just different.

MITRE Makes CVE Program Independent After U.S. Government Funding Expires ⇥ krebsonsecurity.com

Brian Krebs:

A critical resource that cybersecurity professionals worldwide rely on to identify, mitigate and fix security vulnerabilities in software and hardware is in danger of breaking down. The federally funded, non-profit research and development organization MITRE warned today [April 15] that its contract to maintain the Common Vulnerabilities and Exposures (CVE) program — which is traditionally funded each year by the Department of Homeland Security — expires on April 16.

[…]

Update, April 16, 11:00 a.m. ET: The CVE board today announced the creation of non-profit entity called The CVE Foundation that will continue the program’s work under a new, unspecified funding mechanism and organizational structure.

I do not love how I am learning how many critical things worldwide are dependent on an increasingly unstable U.S. government.

Lyft Texted a Rider a Transcript of the Conversation She Had in a Car ⇥ cbc.ca

Nicole Brockbank, CBC News:

Anvi Ahuja noticed a “freaky” new text message from a number she didn’t know right after getting back to her downtown Toronto apartment last month.

The text was a transcript of the conversation she’d just had with her roommates during their eight-minute Lyft ride home from a friend’s place.

[…]

The company confirms the incident took place, but has offered varying explanations.

None of Lyft’s explanations make sense to me. Bizarre and creepy.

Meta’s Poor Prebuttal to “Careless People”

I would have loved to have been a fly on the wall when Mark Zuckerberg and the rest of the Meta’s leadership team found out about Sarah Wynn-Williams’ book “Careless People”. This conversation could have taken place anywhere, but I imagine it was in a glass-walled meeting room which would allow one of these millionaire’s shouted profanities to echo. That feels right. Wynn-Williams, a former executive at Facebook, is well-placed to tell stories of the company’s indifference to user privacy, its growth-at-all-costs mentality, its willingness to comply with the surveillance and censorship requirements of operating in China, and alleged sexual harassment by Joel Kaplan. And, of course, its inaction in Myanmar that played a “determining role” in a genocide that killed thousands in one month in 2017 alone.

Based on some of the anecdotes in this book, I am guessing Zuckerberg, Kaplan, and others learned about this from a public relations staffer. That is how they seem to learn about a lot of pretty important things. The first public indication this book was about to be released came in the form of a preview in the Washington Post. Apparently, Meta had only found out about it days before.

It must have been a surprise as Meta’s preemptive response came in the form of a barely formatted PDF sent to Semafor, and it seems pretty clear the company did not have an advance copy of the book because all of its rebuttals are made in broad strokes against the general outline reported by the Post. Now that I have read the book, I thought it would be fun and educational to compare Meta’s arguments against the actual claims Wynn-Williams makes. I was only half right — reading about the company’s role in the genocide in Myanmar remains a chilling exercise.

A caveat: Wynn-Williams’ book is the work of a single source — it is her testimony. Though there are references to external documents, there is not a single citation anywhere in the thing. In an article critical of the book, Katie Harbath, one of Wynn-Williams’ former colleagues, observes how infrequently credit is given to others. And it seems like it, as with most non-fiction books, was not fact-checked. That is not to disparage the book or its author, but only to note that this is one person’s recollections centred around her own actions and perspective.

One other caveat: I have my own biases against Meta. I think Mark Zuckerberg is someone who had one good idea and has since pretended to be a visionary while stumbling between acquisitions and bad ideas. Meta’s services are the product of bad incentives, an invasive business model, and a caustic relationship with users. I am predisposed to embracing a book validating those feelings. I will try to be fair.

Anyway, the first thing you will notice is that most of the points in Meta’s response do not dispute the stories Wynn-Williams tells; instead, the company wants everyone to think it is all “old news” and it resents all this stuff being dredged up again. Yet even though this is apparently a four-hundred-page rehash, Meta is desperate to silence Wynn-Williams in a masterful gambit. But of course Wynn-Williams is going to write about things we already know a little bit about; even so, there is much to learn.

Most of Meta’s document is dedicated to claims about the company’s history with China, beginning with the non-news that it wanted to expand into the country:

SWW’s “New” Claim:

Facebook Had A Desire To Operate In China.

Old News:

Zuckerberg Addressed This In 2019 Televised Speech. Mark himself said in a televised address in 2019, “[He] wanted our services in China … and worked hard to make this happen. But we could never come to agreement on what it would take for us to operate there.’ That is why we don’t operate our services in China today.”

No, it is not just you: the link in this section is, indeed, broken, and has been since this document was published, even though the page title suggests it was available six months prior. Meta’s communications staff must have known this because they include a link to the transcript, too. No, I cannot imagine why Meta thought it made sense to send people to an inactive video.

At any rate, Zuckerberg’s speech papers over the lengths to which the company and he — personally — went to ingratiate himself with leaders in China. Wynn-Williams dedicates many pages of her book to examples, but only one I want to focus on for now.

But let me begin with the phrase “we could never come to an agreement on what it would take for us to operate there”. In the context of this speech’s theme, the importance of free expression, this sounds like Meta had a principled disagreement with the censorship required of a company operating in China. This was not true. Which brings me to another claim Meta attempts to reframe:

SWW’s “New” Claim:

Facebook Developed Censorship Tools For Use By Chinese Officials.

Old News:

2016 New York Times Report On Potential Facebook Software Being Used By Facebook In Regard To China; Noted It “So Far Gone Unused, And There Is No Indication That Facebook Has Offered It To The Authorities In China.” […]

This is not a denial.

Wynn-Williams says she spent a great deal of time reading up on Facebook’s strategy in China since being told to take it over on a temporary basis in early 2017. Not only was the company okay with censoring users’ posts on behalf of the Chinese government, it viewed the capability as leverage and built software to help. She notices in one document “the ‘key’ offer is that Facebook will help China ‘promote safe and secure social order'”:

I find detailed content moderation and censorship tools. There would be an emergency switch to block any specific region in China (like Xinjiang, where the Uighurs are) from interacting with Chinese and non-Chinese users. Also an “Extreme Emergency Content Switch” to remove viral content originating inside or outside China “during times of potential unrest, including significant anniversaries” (like the June 4 anniversary of the Tiananmen Square pro-democracy protests and subsequent repression).

Their censorship tools would automatically examine any content with more than ten thousand views by Chinese users. Once this “virality counter” got built, the documents say that Facebook deployed it in Hong Kong and Taiwan, where it’s been running on every post.

Far from a principled “agreement on what it would take for us to operate [in China]”, Facebook was completely fine with the requirements of the country’s censorial regime until it became more of a liability. Then, it realized it was a better deal to pay lobbyists to encourage the U.S. government to ban TikTok.

SWW’s “New” Claim:

Facebook Tested Stealth App In China.

Old News:

2017 New York Times Headline: “In China, Facebook Tests The Waters With A Stealth App” “We have long said that we are interested in China, and are spending time understanding and learning more about the country in different ways,” Facebook said in a statement.”

Wynn-Williams provides plenty more detail than is contained in the Times report. I wish I could quote many pages of the book without running afoul of copyright law or feeling like a horrible person, so here is a brief summary: Facebook created an American shell company, which spun up another shell company in China. Meta moved one of its employees to an unnamed “Chinese human resources company” and makes her the owner of its Chinese shell company. A subsidiary of that company then launched two lightly reskinned Facebook apps in China, which probably violate Chinese data localization laws. And I need to quote the next part:

“Are Mark and Sheryl okay with it?” I ask.

He [Kaplan] admits that they weren’t aware of it.

I am unsure I believe Zuckerberg did not know about this arrangement.

As far as I can find, most of these details are new. The name of the Chinese shell company’s subsidiary was published by the Times, but the only reference I can find — predating this book — to the other shell companies is in an article on Sohu. I cannot find any English-language coverage of this situation.

Nor, it should be said, can I find any discussion of the legality of Facebook’s Chinese operations, the strange leadership in the documentation for the shell companies, the other apps Facebook apparently released in China, and the fallout after the Times article was published. Meta’s anticipatory response to Wynn-Williams’ book pretends none of this is newsworthy. I found it captivating.

SWW’s “New” Claim:

Chinese Dissident Guo Wengui Was Removed Due To Pressure From The Chinese Government.

Old News:

The Reasons Facebook Removed Guo Wengui From The Platform Were Publicly Reported In 2017;

Unpublished His Page And Suspended His Profile Because Of Repeated Violations Of Company’s Community Standards.

Wynn-Williams says this is false. She quotes the same testimony Facebook’s general counsel gave before a Senate Intelligence Committee hearing, but only after laying out the explicit discussions between Facebook and the Cyberspace Administration of China saying the page needed to be removed. At a time when Facebook was eager to expand into China, the company’s compliance was expected.

Then we get to Myanmar:

SWW’s “New” Claim:

Facebook Dragged Its Feet On Myanmar Services.

Old News:

Facebook Publicly Acknowledged Myanmar Response In 2018. The facts here have been public record since 2018, and we have said publicly we know we were too slow to act on abuse on our services in Myanmar […]

“Too slow to act” is one way to put it — phrasing that, while not absolving Meta, obscures a more offensive reality.

Myanmar, Wynn-Williams writes, was a particularly good market for Facebook’s Free Basics program, a partnership with local providers that gives people free access to services like Facebook and WhatsApp, plus a selection of other apps and websites. It is an obvious violation of net neutrality principles, the kind Zuckerberg framed in terms of civil rights in the United States, but seemed to find more malleable in India. Wynn-Williams’ description of the company’s campaign to save it in India, one that was ultimately unsuccessful, is worth reading on its own. Facebook launched Free Basics in Myanmar in 2016; two years later, the company pulled the plug.

From 2014 to 2017, Kevin Roose reported for the New York Times, Facebook’s user base in Myanmar — a country of around fifty million people — grew from two million to thirty million, surely accelerated by Free Basics. During that time period, Wynn-Williams writes, Facebook observed a dramatic escalation in hate speech directed toward the Rohingya group. As this began, Meta had a single contractor based in Dublin who was reviewing user reports. One person.

Also, according to Wynn-Williams, Facebook is not available in Burmese, local users “appear to be using unofficial Facebook apps that don’t offer a reporting function”, and there incompatibility between Unicode and Zawgyi. As a result, fewer reports are received, they are not necessarily in a readable format, and they are processed by one person several time zones away.

Leading up to the November 2015 election, Wynn-Williams says they doubled the number of moderators speaking Burmese — two total — plus another two for election week. Wynn-Williams says she worked for more than a year to get management’s attention in a candidate for Myanmar-specific policies, only for Kaplan to reject the idea in early 2017.

From the end of August to the end of September that year, 6,700 people are killed. Hundreds of thousands more are forced to leave the country. All of this was accelerated by Facebook’s inaction and ineptitude. “Too slow to act” hardly covers how negligent Facebook was at the scale of this atrocity.

SWW’s New Claim:

Facebook Offered Advertisers The Ability To Target Vulnerable 13-17 Year Olds.

Old News:

Claim Was Based On A 2017 Article By The Australian, Which Facebook Refuted. “On May 1, 2017, The Australian posted a story regarding research done by Facebook and subsequently shared with an advertiser. The premise of the article is misleading. Facebook does not offer tools to target people based on their emotional state.”

In fact, Wynn-Williams quotes the refutation, calling it a “flat-out lie”. She says this is one of at least two similar slide decks discussing targeting ads based on a teenager’s emotional state, and in an internal discussion, finds an instance where Facebook apparently helped target beauty ads to teenage girls when they deleted pictures of themselves from their accounts. She also writes that the deputy chief privacy officer confirmed it is entirely possible to target ads based on implied emotion.

After the “lie” is released as a statement anyway, Wynn-Williams says she got a phone call from a Facebook ad executive frustrated by how it undermined their pitch to advertisers for precise targeting.

SWW’s New Claim:

Donald Trump Was Charged Less Money For Incendiary Adverts Which Reached More People.

Old News:

Claim Was Based On A 2018 Article By Wired, Which Facebook Refuted.

The argument made by Meta’s Andrew Bosworth, in a since-deleted Twitter thread, was that Trump’s average CPM was almost always higher than that of the Clinton campaign. But while this is one interpretation of the Wired article — one the publication disputed — this is not the claim made by Wynn-Williams. She only says the high levels of engagement on Trump’s ads drove their prices down, but not necessarily less than Clinton’s ads.

Those were all of the claims Meta chose to preemptively dispute or reframe. The problem is that Wynn-Williams did make news in this book. There are plenty of examples of high-level discussions, including quoted email messages, showing the contemporaneous mindset to build its user base and market share no matter the cost. It is ugly. Meta’s public relations team apparently thought it could get in front of this thing without having all the details, but their efforts were weak and in vain.

Following its publication, Meta not only sought and received a legal demand that Wynn-Williams must stop talking about the book, it followed its “playbook for answering whistleblowers”. I am not a highly paid public relations person; I assume those who are might be able to explain if this strategy is effective. It does not seem as though it is, however.

This is all very embarrassing for Meta, yet seems entirely predictable. Like I wrote, I have my own biases against the company. I already thought it was a slimy company and I hate how much this confirms my expectations. So I think it is important to stay skeptical. Perhaps there are other sides to these stories. But nothing in “Careless People” seems unbelievable — only unseemly and ugly.