Month: December 2023

When I link to books, I typically point readers to IndieBound — now Bookshop — because it lets people buy a copy from a local store instead of a rainforest-themed monolith. Though this works for U.S.-based readers, it has not given Canadians the same access.

Enter IndieBookstores, which does basically the same thing but in Canada. It is built on BookManager, which seems to be fine software for bookstores but not great for individual users just trying to find a title.

Caroline O’Donovan, Washington Post:

Google settled a class-action lawsuit on Thursday brought by users who alleged the search giant captured and tracked their data while in “Incognito” mode, a Chrome browser setting that is supposed to protect users’ privacy.

Previously, a federal judge in California had scheduled a 2024 trial date in the case, which has been put on hold while the details of the settlement are finalized, according to a Thursday court filing.

Simon Sharwood, the Register:

The plaintiffs initially suggested damages of $5 billion, with around $5,000 paid to each of a million potential complainants. Sadly, document 1,089 [the term sheet] doesn’t mention the agreed settlement.

This lawsuit has always seemed pretty dumb to me if you know what “Incognito Mode” is supposed to mean, something which Google spells out when you enable it in Chrome. However, I do think it is telling how much this relies on the fine print of how Google itself defines “incognito” compared to the word’s actual meaning, in a way that sort of reminds me of Tesla’s “Autopilot” and “Full Self-Driving” features. I am not saying this misinterpretation should be worth five billion dollars; all I am saying is that “incognito” is a bad word to describe the actual functionality it offers.

I have previously linked to Bloomberg’s annual “jealousy list” of articles from other outlets they wish they had published; it is mostly an excuse to ask about the Supermicro fiasco.

This year, though, no snark and no Bloomberg. This year, Rest of World — a website you really ought to be reading, if I may be so bold — published its own list of great stuff from other publications. I like the international focus of the articles this list, even though most of them are from U.S. outlets.

In addition to the aforelinked Zeke Faux story, my favourites from this list are John Herrman’s look at Temu, Rebecca Tan and Regine Cabato’s investigation of A.I. training for the Washington Post, and Vox’s Izzie Ramirez explaining why everything you buy is a little worse now. Happy reading.

Alastair McCready and Allegra Mendelson, South China Morning Post (likely paywalled):

Cambodia’s scam sector gained infamy last year after reports seeped out of massive human trafficking, forced labour and systematic torture. The Philippines’ role as a major scam hub was also reaffirmed in late June when a raid on a Chinese-run scam complex in southern Manila freed more than 2,700 people, while Laos is host to large-scale scam operations at the notorious Golden Triangle Special Economic Zone.

But in Myanmar, these brutal criminal enterprises continue unchecked. They are abetted by the country’s ongoing domestic turmoil, brought about by the 2021 military coup, and operated by alliances of Chinese criminals and a local paramilitary group, beyond the reach of outside law enforcement, civil society and the media.

Teele Rebane, et al., CNN:

In three short years, according to UN and FBI investigators, transnational crime organizations have exploited developments in technology and the civil war in Myanmar to build a billion-dollar industry to scam people across the world out of their life savings.

This huge scam operation relies on army of modern-day slaves, assembled by what the UN has called one of the largest human trafficking events in Asia in recent history.

It’s known as a “pig butchering” scam — a type of confidence fraud in which victims are lured by scammers often impersonating young women on the internet. The scammers then spend weeks building a relationship with their victim, introducing them to cryptocurrency and encouraging them to invest on a fake platform.

In his book “Number Go Up”, Zeke Faux reported on the mass scam operations in Cambodia.

There are a bunch of stories about what these operations are like for scammers and the criminals they are coerced into working for. It is a consistently sad and brutal story, but it is one that is important to keep studying and repeating as a cautionary tale for the multiple layers of potential victims.

The CNN article is good, but it appears to draw heavily from other sources for which it gives no credit. Most obvious is the SCMP story; it also seems to me based on the details provided that CNN spoke to the same U.S. victim as first reported by Cezary Podkul and Cindy Liu of ProPublica last September. I recommend reading them all, as each has different details the others lack, though the ProPublica story is the most comprehensive.

Dan Goodin, Ars Technica:

Researchers on Wednesday presented intriguing new findings surrounding an attack that over four years backdoored dozens if not thousands of iPhones, many of which belonged to employees of Moscow-based security firm Kaspersky. Chief among the discoveries: the unknown attackers were able to achieve an unprecedented level of access by exploiting a vulnerability in an undocumented hardware feature that few if anyone outside of Apple and chip suppliers such as ARM Holdings knew of.

“The exploit’s sophistication and the feature’s obscurity suggest the attackers had advanced technical capabilities,” Kaspersky researcher Boris Larin wrote in an email. “Our analysis hasn’t revealed how they became aware of this feature, but we’re exploring all possibilities, including accidental disclosure in past firmware or source code releases. They may also have stumbled upon it through hardware reverse engineering.”

The post on Kaspersky’s site has more technical information about the chain of exploits used here.

As you might recall, Russian intelligence officials claimed Apple assisted the NSA to build this malware — something which Apple has denied and, it should be noted, no proof has been provided for Apple’s involvement or the NSA’s. It does not appear there is any new evidence which would implicate Apple. But it is notable that it relied on an Apple-specific TrueType specification, and bypasses previously undisclosed hardware memory protections. To be clear, neither of those things increases the likelihood of Apple’s alleged involvement in my mind. It does show how disused or seemingly irrelevant functions remain vulnerable and can be used by sophisticated and likely state-affiliated attackers.

Benjamin Mullin and Tripp Mickle, New York Times:

Apple has opened negotiations in recent weeks with major news and publishing organizations, seeking permission to use their material in the company’s development of generative artificial intelligence systems, according to four people familiar with the discussions.

This is very different from the way existing large language models have been trained.

Kali Hays, of Business Insider, in November:

Most tech companies seemed to agree that being required to pay for the huge amounts of copyrighted material scraped from the internet and used to train large language models behind AI tools like Meta’s Llama, Google’s Bard, and OpenAI’s ChatGPT would create an impossible hurdle to develop the tech.

“Generative AI models need not only a massive quantity of content, but also a large diversity of content,” Meta wrote in its comment. “To be sure, it is possible that AI developers will strike deals with individual rights holders, to develop broader partnerships or simply to buy peace from the threat of litigation. But those kinds of deals would provide AI developers with the rights to only a minuscule fraction of the data they need to train their models. And it would be impossible for AI developers to license the rights to other critical categories of works.”

If it were necessary to license published materials for training large language models, it would necessarily limit the viability of those models to those companies which could afford the significant expense. Mullin and Mickle report Apple is offering “at least $50 million”. Then again, large technology companies are already backing the “A.I.” boom.

Mullin and Mickle:

The negotiations mark one of the earliest examples of how Apple is trying to catch up to rivals in the race to develop generative A.I., which allows computers to create images and chat like a human. […]

Tim Bradshaw, of the Financial Times, as syndicated by Ars Technica:

Apple’s latest research about running large language models on smartphones offers the clearest signal yet that the iPhone maker plans to catch up with its Silicon Valley rivals in generative artificial intelligence.

The paper, entitled “LLM in a Flash,” offers a “solution to a current computational bottleneck,” its researchers write.

Both writers frame this as Apple needing to “catch up” to Microsoft — which licenses generative technology from OpenAI — Meta, and Google. But surely this year has demonstrated both how exciting this technology is and how badly some of these companies have fumbled their use of it — from misleading demos to “automated bullshit”. I have no idea how Apple’s entry will fare in comparison but it may, in retrospect, look wise for it to dodge this kind of embarrassment and the legal questions of today’s examples.

Jay Graber announced the new branding — a butterfly — on Bluesky’s blog:

Like a butterfly emerging from its chrysalis, we are starting to open up. Posts on Bluesky have been public from the start through the open protocol, but today we’re making them publicly accessible through the app. We’re unfolding a little bit at a time, and are excited to bring you along on this journey of metamorphosis!

This justification might be a little try-hard, but I love the logo: a blue butterfly. It is so nice and so simple that I was surprised when I could not think of another tech company which had used that kind of visual before. And then I remembered MSN. Bluesky’s take is better, of course, and I appreciate how it retains a connection to Twitter. It feels to me like a newer and brighter version of Twitter than it does a knockoff.

This announcement gives me an excuse to share some thoughts on the state of the big three Twitter replacements, beginning with Bluesky. It is my favourite of the three. It feels light and nimble, like it was specifically designed for quick snippets of text — because that is what it is. I appreciate the ability to create and share recipes for algorithmic feeds, and I like how it permits you to tailor filtering to your preferences. Its first biggest struggle right now is that it remains available by invitation only, which has slowed its growth, but perhaps that will prove to be a wise decision in the long-term. The other problem for me is the lack of a real Mac app.

Mastodon is my second-favourite, though it is the one I use most because there are proper apps for it available on all the platforms I use. I also have more followers there which helps when I want answers to a question. The main reason I like it less than Bluesky is because it feels more complicated. Every action is spread across multiple networks, which means adding a post to my favourites means copying the post URL to the search box of my Mastodon instance, and then hitting the star icon. It is one of many hiccups of the existing federated architecture, but it does not seem to me like it is an inherent problem — just an issue with the current implementation.

In distant last place is Threads. It may be the most popular, but it is the one I find myself checking least, and the only one I would not miss if it disappeared tomorrow. In contrast to Bluesky’s lightness, Threads feels cumbersome, like every action necessitates the physical movement of gears in some distant building. Publishing, viewing a thread, or blocking someone takes several seconds each time. The algorithmic home timeline rarely serves me things I am interested in, and seems to favour engagement bait and brands and influencer nonsense and brands, and also brands. It is a shame because it will sometimes show me threads from photographers who post great work, so I tap the heart and hope to see more — but never do. I hide and block with glee. None of this seems to suggest to Threads what I would like to see more or less of.

Colin Devroe:

In 1991, Geoffrey A. Moore described the challenges of introducing new technology products as Crossing the Chasm. The chasm is this very real gap between the earliest adopters and the early majority adopters of any new technology. By crossing the chasm, the momentum gained usually enables the technology to find market fit.

Most protocols, standards, products and services experience this gap in adoption. Even the internet followed the lifecycle described by Moore. I think it can be said that ActivityPub, though fairly well implemented in a variety of services for several years now, is about to cross that chasm in 2024.

I would like to see this happen — I am more optimistic than I may come across — but I am skeptical because Threads — the best chance for widespread adoption of ActivityPub — will make fediverse integration an option, not mandatory. That is probably the right call. It just means users will be required to dig around in their account settings to change a preference for something they may not understand, which I do not expect many people to do.

The other reason I am not holding my breath is that “ActivityPub” and “fediverse” are clunky and confusing names. They sound technical. I have long argued that “RSS” has a similar impediment. But my expectation of the role of naming in a product’s viability has softened in the past year because of the stunning success of products called “Substack” and “ChatGPT”.

David Pierce, the Verge:

I’m convinced we’ll be better off with a hundred different apps for Snapchat or Instagram or X instead of just one, a dozen companies competing to build the best moderation tools, and an app store filled with different ways for me to follow and be followed by other people on the internet. It doesn’t make sense that we have a dozen usernames, a dozen profiles, a dozen sets of fans and friends. All that stuff should belong to me, and I should be able to access it and interact with it anywhere and everywhere.

This sounds right to me. Any platform operator should have standards and expectations for what they will permit, but control over what users see should be separated from those centralized positions. It should be a marketing advantage for client software which can surface the best posts from anywhere and filter out the worst. There is no reason that kind of algorithmic sorting needs to have a large Californian business at its core.

Dave Karpf:

But the philosophical issues are secondary to the pragmatic ones. Pragmatically, it’s really quite simple. Content moderation is costly. It is a first-order revenue sink, not a revenue-generator. (I say “first-order” because if you skimp on content moderation, eventually you’re going to have a cascade of problems that cost you a lot of money <*cough* ElonyouidiotNilayPatelwarnedyou *cough*>.) But the KPIs for the content moderation team are never going to be “look how much new business we brought in for the company.” When content moderation is going well, you don’t hear much about it. That’s kind of the goal.

This is why every tech CEO loves the libertarian approach to speech issues. Tech libertarianism holds that someone else (or no one at all) should expend resources on setting and enforcing boundaries for how your product is used. The essence of the position is “I shouldn’t have to spend money on any of this. And I shouldn’t ever face negative consequences for not spending money on this.”

Karpf’s apt metaphor is the amount of mouse poop permitted in boxed cereal — which, if you are interested, is not a specific FDA category, but on a related note, is an average of below nine milligrams per kilogram of wheat, and less than one in a sample of fifty grams of cornmeal — was well considered even before Substack co-founder Hamish McKenzie published a defence of why platforming the views of literal Nazis is good and necessary. There is a vast gulf between acknowledging that some people who have Nazi views may be using the platform inadvertently and what McKenzie wrote, which is that Nazis ought to be welcomed on Substack to broadcast their views short of specific encouragement of violent acts. He may as well be advocating for the right of mice to defecate in cereal.

Substack is not the last bastion of free speech on the web, despite what it may claim. It is just one platform of many, and there is always the freedom to host your own. Substack is proud that writers can “own all your content” and “own your subscriber list”. It should mean a much lower tolerance for Nazis because it means those writers can be kicked off but they get to take their list elsewhere. They get to keep their audience; they lose nothing except time. Instead, Substack is pretending it is some bulwark against encroaching censorship, and the best way it can demonstrate that is by allowing Nazis to generate income and take a 10% commission from those subscribers.

In the nick of time to beat its self-imposed “late 2023” deadline, Apple has announced two automakers which will support the new version of CarPlay: Aston Martin and Porsche. This news is not relevant to my tax bracket, and it was delivered in a series of articles from — at least — Cool Hunting, British GQ, and Car and Driver. The Cool Hunting article, by Josh Rubin, is probably the best of the bunch — even though it reads like marketing copy directly from Apple — primarily because there are quotes from Alan Dye, an Aston Martin executive, and Porsche’s “Vice President of Style”. Nice.

I am still uncertain about the transformation of the entire dashboard into a control interface with moving touch targets. I can adjust the heated seats in my car while on the highway because I know exactly where the physical button is, and I can feel when my finger is on it and when it is pressed. None of that applies to a touch screen. There is not any great data on how safe these systems are compared to physical knobs and switches, but a 2020 study of just forty people found CarPlay and Android Auto distracted drivers to the point where their response times were worse than substance-impaired drivers.

Chance Miller, 9to5Mac:

In a statement to 9to5Mac, Apple has announced that it will soon halt sales of its flagship Apple Watch models in the United States.

The Apple Watch Series 9 and Apple Watch Ultra 2 will no longer be available to purchase from Apple starting later this week.

The move comes following an ITC ruling as part of a long-running patent dispute between Apple and medical technology company Masimo around the Apple Watch’s blood oxygen sensor technology.

Nilay Patel on Threads:

No shade to Chance or 9to5 which are both great but it is pretty intensely irresponsible for Apple to not have a full press release on this extremely material market moving news — an unnamed spox giving a background statement saying the Apple Watch won’t be on sale anymore to one enthusiast outlet is uhhhh

The “enthusiast” description seems unfair: 9to5Mac is not merely an Apple fan blog; its writers regularly produce original reporting and are routinely critical of Apple. But the substance of Patel’s post is absolutely true. It is deeply weird that Apple issued an anonymous statement to a niche publication announcing it will soon stop selling a flagship product, which is usually the kind of thing it would want to tell the Wall Street Journal or CNBC.

John Voorhees, MacStories:

If I had to guess what’s going on here, I’d say it’s a high-stakes game of corporate chicken. Masimo got a ruling from the ITC that gave it leverage, so they asked for a big licensing deal. The Biden administration probably doesn’t want to deal with the dispute or look like it’s bailing out a big tech company, so I bet it told the parties to work things out, assuming Apple would pay up. Whether it ultimately will, only Apple knows, but it’s decided to force the Biden administration’s hand on the veto. […]

Mark Gurman, Bloomberg:

Work within Apple suggests that the company believes software changes — rather than a more complicated hardware overhaul — will be enough to bring the device back to store shelves. But the patents at the heart of the dispute are mostly related to hardware, including how light is emitted into the skin to measure the amount of oxygen in a person’s blood.

Gurman says sales of refurbished Series 8 models — “refurbished” being the only way that model remains available from Apple — will also be halted because it, too, has blood oxygen capabilities.

Michael Grothaus, writing for Fast Company, has some pretty standard grievances with the increasing role of computers in photography:

As a middle-aged Gen Xer, these new camera technologies never cease to amaze me. But as someone who was also trained in analog photography — capturing images on actual strips of celluloid that I then developed in a dark room manually — these new technologies also leave me feeling that the gap between photographer and photograph is now so wide the two almost aren’t connected anymore. And as that gap broadens with additional generative AI processes, I fear that the photographer will one day become nothing more than a biological tripod for the camera. Once that happens, the capturing device becomes the creator.

This is a common fear for any technology where the behind-the-scenes processes are simplified and less intimidating. If anything, I would argue that the “gap between photographer and photograph” has narrowed, not widened. Photographers now get to shoot, edit, and publish images — often all by themselves — in a fraction of the time and with less expense as film development used to require.

Still, I am hopeful. In the past year, there’s been an upswing of professional photographers returning to analog film cameras.

This claim probably seems unlikely to you; it sure does to me. Its foundation is the flimsiest substance you can imagine — something which seems to be pretty standard for Fast Company. Look, I know I have an unhealthy obsession with the absolute nonsense from this publisher but, in fairness, they should have better stuff.

In this case, here is Grothaus’ proof of the aforementioned upswing:

But recently, there’s been a reversal. In 2023, Leica expects to sell 5,000 analog M series cameras — 10 times more than it sold nearly a decade earlier. To put that surging number into further context, Leica expects to sell around 12,000 digital M series cameras this year. The fact that analog cameras will be selling at 40% of the volume of the company’s digital cameras is staggering, especially considering that Leica is still innovating in digital photography on almost every front.

That is it: Grothaus’ sole evidence for a film photography “comeback”, as the headline puts it. The forecasts of one niche company — which caters to people that can afford its infamously pricey cameras, many of whom have a reputation of treating them as objets d’art and shelving them — surely does not speak to a broader trend. Still, Leica is expecting to sell a lot more film cameras this year compared to eight years prior, so perhaps there is a specific reason for that. And, indeed, there is: Leica re-released the M6 last year, and it is plausible lots of people would buy it. But who?

Here is the next paragraph:

As for why film photography is seeing a resurgence among professionals, each photographer probably has their own reasons. […]

Who says the primary market for this camera is among professional photographers? Not Grothaus; there is no evidence that professionals are increasingly shooting film, whether for clients or personally. There are also people with money to burn and display cabinets to fill, and I guess some hobbyists too.

[…] But I can guess at some of them, because they would be my own. Analog photography gives the photographer a connection to their photographs that digital photography does not provide. You assume a physical role in birthing the photograph through everything from selecting the film stock, to manually adjusting optical filters, to mixing the perfect solution in a warm water tub to develop the film under the red light of the dark room. There’s an intimacy in the analog process and a period of anticipation that makes seeing the final result exciting.

Speaking of assumptions unproven by the evidence, it is hard for me to believe most photographers developed their own film even at the peak of celluloid relevance. There is no obligation that one must develop their own film, even if they are a professional photographer. In fact, there is good reason to take images to a professional developer who knows exactly what they are doing and can help a photographer achieve specific effects.

But put a pin in the rest of this. I will come back to it later.

Then we get to the penultimate paragraph:

I don’t expect celluloid photography to make a mainstream comeback. Digital photography is too convenient for most people. Gen Z has really never known anything else. For Gen A, a photograph will have always been a malleable thing. It will be expected that a camera can apply any expression to a subject’s face or even entirely reposition them if needed.

This paragraph demonstrates the hollowness of the whole article. Because Grothaus is so focused on the specific sales forecast of a desirable German camera brand, he misses a really interesting story in how much younger generations are, in fact, interested in film photography. A New York Times article last year highlighted the sales growth of film cameras driven, it claimed, by younger buyers. A CBC News story from earlier this year also profiled surging interest from teenagers and young adults.

There is a fascinating narrative there — albeit, one with the side effect of increasing the cost of my own hobby over the past fifteen years thanks to those damn kids. But it is also a story which has been reported constantly; old things are new again is a well-worn trope. Specifically, you can find plenty articles proclaiming the revitalization of film photography from every year since at least 2010; here is a sampling:

  • That year, Kodak proclaimed a “very real resurgence for film” after noting strong sales of colour and black-and-white negatives.

  • In 2011, Ashlea Halpern framed it as an “analog renaissance”.

  • In 2012, Jenna Wortham wrote for the New York Times that “film photography is having another moment in the sun”.

  • In 2013, it was Bloomberg’s turn to highlight two people in the photo industry calling it a “resurgence”.

  • In 2014, Fstoppers noted renewed interest in film.

  • From 2015, Format magazine highlighted artists contributing to the “resurgence in the practice of film photography”.

  • Then, in 2016, the resurgence was framed as a question by Shutterbug.

  • In 2017, Time magazine excitedly covered the relaunch of Kodak Ektachrome and pointed to year-over-year film growth.

  • In 2018, the Guardian reported that “[l]arge numbers of still photographers, professional and amateur alike, are turning their backs on digital technology”.

  • In 2019, HP — of all companies — published a rather good exploration of how film “is making a comeback among professionals and amateurs alike”.

  • A 2020 headline at Fstoppers sounds like a threat.

  • Then, just one year later and from the exact same author, the headline became a question: “Are We Nearing the End of the Film Bubble?”

  • Finally, last year, Axios reported “hobbyists and young people are increasingly taking up the once bygone medium” causing “huge price increases”.

The thing is that all of these stories are believable; it is plausible film photography has experienced a constant increase in attention over the past thirteen years — at least. Some of these articles claim it is driven by professionals, while others credit younger buyers.

The headline of that CBC story calls it “like vinyl, but for photos”, and I think that is an apt comparison: in the U.S., L.P. album sales revenue in the 2020s is the highest it has been since the late 1980s, even accounting for inflation, but it is still a tiny fraction of spending on music. Likewise, an increased number of people shooting film is notable, but significant only in relation to how few people were using an analog camera ten or fifteen years ago and a “tumultuous” series of years for the digital camera industry thanks to smartphones and pandemic-related shortages.

In any case, I think it is fantastic any time some new generation rediscovers things considered irrelevant or obsolete. You can see this everywhere: young people update fashion of an older era with a new twist, listen to music from their parents’ and grandparents’ generations, and try older technologies for the first time. I remember doing the same when I was younger; I learned how to use a microfiche reader at the library when I was a teenager, and I still think that is a fabulous piece of tech.

I want to go back to the part earlier about the reasons people may choose to shoot film instead of digital, as there is something I wish to highlight:

[…] Analog photography gives the photographer a connection to their photographs that digital photography does not provide. […]

You will recall that Grothaus then explained how exciting it is to develop your own film, which is a completely different subject. I, too, enjoyed it at one point.

But I agree that there is a difference in the photographer’s connection to the image in film and digital formats — not less, per se, just different. A film photo is inherently a little nostalgic, a little bit fuzzier, a little bit more like a memory. It requires patience and being more okay with chance.

A digital image is more like reliving a specific moment in stark detail. It encourages connection if the photographer desires to edit it in nitpicking detail, in ways which are far more complicated and difficult for film.

These are two specific mediums which can happily coexist. They can encourage different behaviour from their photographers and subjects alike. If there is some kind of film renaissance happening — and it seems that there has been for over a decade — I welcome the opportunity for more people to experience a different way of capturing light. Digital has clearly won in the sense that you are likely reading these words on a digital camera right now. But that does not mean film must ever go away. This obsession with framing it as a resurgence is, I think, unnecessary when it is far more exciting to treat it as an existing tool of many. Young photographers now have options for creative expression — including brand new ones like generative imagery — that generations from twenty or more years ago did not. That should be celebrated.

Figma CEO Dylan Field:

Figma and Adobe have reached a joint decision to end our pending acquisition. It’s not the outcome we had hoped for, but despite thousands of hours spent with regulators around the world detailing differences between our businesses, our products, and the markets we serve, we no longer see a path toward regulatory approval of the deal.

It is not worth being too upset for Figma. Adobe’s press release confirms it will be paying Figma the termination fee of one billion dollars.

It seems unwise to treat this as a case of a smaller company not being allowed to grow up because of big government regulators. Adobe was clearly trying to buy its first major competitor since Macromedia — which it also acquired. Technology companies, even very big ones like Adobe, will certainly be allowed to merge and acquire, but it is right for competition authorities to ask questions about what impact it will have for individuals. It is hard to see how Figma would be more beneficial as a subsidiary instead of a competitor.

Glenn Fleishman wrote a wonderful history of Yahoo Pipes:

The interactive, web-based visual programming of Pipes — inspired by earlier interfaces — became a primary influence on a generation of UI/UX designers and products that followed. Its remix ethos matched the time, and led people to dream bigger. It seemed like it held some version of the possible—but with no business model behind it and no internal commitment to advance the product, it became a poster child for unrealized potential, like so much else at Yahoo.

I was pretty young but I remember putting together a handful of workflows in Pipes. I do not think I fully appreciated how novel it was at the time, though I recall missing it deeply when it was shut down.

What a great trip down memory lane, and presented in a very cool way. There is no “Chaos” slider but it also did not make my CPU work so hard it became sentient and asked me to wait a few minutes for a single webpage to load.

Joseph Cox, 404 Media:

A marketing team within media giant Cox Media Group (CMG) claims it has the capability to listen to ambient conversations of consumers through embedded microphones in smartphones, smart TVs, and other devices to gather data and use it to target ads, according to a review of CMG marketing materials by 404 Media and details from a pitch given to an outside marketing professional. Called “Active Listening,” CMG claims the capability can identify potential customers “based on casual conversations in real time.”

Pretty quick turnaround in the “egg on my face” department, I have to say, but it is great that the privacy reporter named Cox is investigating the privacy nihilists named Cox.

It is not immediately clear if the capability CMG is advertising and claims works is being used on devices in the market today, but the company notes it is “a marketing technique fit for the future. Available today.” 404 Media also found a representative of the company on LinkedIn explicitly asking interested parties to contact them about the product. One marketing professional pitched by CMG on the tech said a CMG representative explained the prices of the service to them.

It seems Cox Media Group has been selling this product for a couple of months with some of those pitched claiming CMG representatives said they get voice data from Apple and Google. That seems implausible, to say the least, and not just from Apple — why would Google be okay with listening in on users for targeted advertising purposes, but not use it for its own ads? It would be shocking if the substance of CMG’s pitch turns out to be true. Given that CMG is the only one making these claims and is not speaking about it at all beyond its marketing, it seems wise to approach this with overwhelming skepticism.

Still, you are thinking it and I am thinking it: what if this is really as described? It reads like a parody; the top of CMG’s marketing page says “It’s True. Your Devices Are Listening to You”, which must feel icky even to the most data-hungry marketer. It gets worse as you scroll down the page. One section asks rhetorically “is this legal?”, which is not generally the kind of question someone ought to be thinking as they consider your product. But it answers itself in the affirmative, insisting “consumers usually give consent when accepting terms and conditions” — a rationale which may be technically correct but is ethically indefensible.

I reached out to CMG with a few questions of my own and I will post an update in the unlikely event a representative replies.

Update: I received a vague statement from CMG which did not answer anything I asked. The company says “CMG businesses do not listen to any conversations or have access to anything beyond a third-party aggregated, anonymized and fully encrypted data set that can be used for ad placement” and it is “committed to ensuring our marketing is clear and transparent”.

Update: MindSift, a marketing firm with a suspiciously light website, also bragged about being able to target ads based on things people are saying in the real world. I am not convinced anybody is actually doing this. It feels to me like marketing bullshit which ad tech providers pretend to have; they can get away with it because many people already believe this is happening, and there is no way to prove it is not. Still, these businesses clearly believe marketers want to hear they can target advertising based on conversations in people’s homes. That does not say good things about the advertising and marketing industry.

Dropbox recently added some features, marketed as “A.I.”, which it turned on by default for people outside the European Economic Area, U.K., and — depending on which part of the page you read — maybe also Canada. In those named regions, it must be enabled by users. So you can imagine how, when people discovered in their Dropbox settings, a toggle in the “on” position for a “third-party A.I.” preference, many were immediately concerned all their stuff was being used to train third-party models.

Not only is this a setting which, according to Dropbox, is only relevant if you use its “A.I.” features, it is also protective of your work.

Simon Willison:

Here’s copy from that Dropbox preference box, talking about their “third-party partners” — in this case OpenAI:

Your data is never used to train their internal models, and is deleted from third-party servers within 30 days.

It’s increasing clear to me like people simply don’t believe OpenAI when they’re told that data won’t be used for training.

What’s really going on here is something deeper then: AI is facing a crisis of trust.

This is a very good article, comparing mistrust of machine learning to the ad-tech-is-listening-through-your-microphone fiction. It is the kind of thing where the available evidence — or lack thereof — for the theory is matched against decades of the deregulation of industry and consequential corporate malfeasance.

Willison suggests OpenAI and other machine learning companies should provide more information about how they train their models, which I think would be helpful. But it does not go far enough because, as Willison writes, people do not believe the overwhelming evidence against the microphone spying theory, either. How much of this is human nature and how much is a learned behaviour is well beyond something I can write about, but it does seem to be something which can be coaxed in one direction or another by setting expectations. There ought to be a legal privacy framework that would give people the confidence their data is not being misused, and there ought to be a way to audit the entire chain. Does your Dropbox-hosted document really get removed from OpenAI’s servers within a month after you used a summarization feature? It would be nice if there was some way to confirm that — for example, I feel confident that a connection between web services has been severed when I delete the relevant OAuth token.

This is important to me. Longtime readers know I am someone who cares deeply about privacy. It is understandable why people were scared: there is deservedly little trust in corporations, this toggle was poorly labelled, and Dropbox did not effectively communicate this addition. This is why it is vital for privacy legislation to be comprehensive and to give people confidence. Not only is it, you know, effective, it should also reduce the spread of these kinds of theories when they are unfounded and give relevant authorities the ability to investigate when there is a legitimate problem.

People should be able to trust systems and institutions, and systems and institutions need to give people reasons to trust them. The correct thing to do is to repair these critical parts of society, not tear them down or prevent them from being built in the first place.

Brian X. Chen tried Meta’s recently updated Ray-Ban “smart” glasses for the New York Times:

To inform people that they are being photographed, the Ray-Ban Meta glasses include a tiny LED light embedded in the right frame to indicate when the device is recording. When a photo is snapped, it flashes momentarily. When a video is recording, it is continuously illuminated.

As I shot 200 photos and videos with the glasses in public, including on BART trains, on hiking trails and in parks, no one looked at the LED light or confronted me about it. And why would they? It would be rude to comment on a stranger’s glasses, let alone stare at them.

Chen interviewed Chris Gilliard, who raised concerns about the implications of these glasses in places where privacy is expected. That seems like a correct worry for the mainstreaming of spy glasses. Glasses with cameras are not new, and you can buy them on Amazon for about $50. But Meta is reframing the creepiness of them as a virtue, as is the company’s business model. What I do not get is what Chen’s experiences are supposed to indicate.

I have shot thousands of photos in public places and, yes, people do not notice — and I was using a conspicuous digital camera for many of those images. People do not have their guard up on trains and in parks in the same way they would in, say, a locker room. It is not newsworthy that nobody asked Chen about the little LED indicator on a hike, and not just because it is “rude to comment on a stranger’s glasses” — or sunglasses, as is the case here. It is a completely different context from the actual places where it is a problem that someone is wearing a discreet camera. It is therefore difficult to see legal and publishable experiences as a useful proxy for illegal and creepy uses.

Jeff Johnson:

It’s a momentous year for my Safari extension StopTheMadness. The fifth anniversary of StopTheMadness was back in April. And today is the zeroth anniversary of StopTheMadness Pro! Actually, Tuesday was the zeroth anniversary, but that’s a story for another day. The story for this day is that my secret project is finally revealed, the next generation of StopTheMadness.

I have been testing this new version for about a month and it is terrific. It remains the extension I most often turn to when a website is user-hostile or misbehaving — a piece of software which should not need to exist yet is invaluable for the modern web.

Google announced yesterday that it would soon begin storing users’ Maps Timeline — which uses the optional Location History toggle to produce a record of places one has visited — locally on users’ devices. It sounds like a privacy enhancement at first, but Ron Amadeo of Ars Technica read the announcement more carefully and finds this description is misleading, and the way controls are presented to users is overly complicated. (Update: Amadeo’s interpretation of this change, according to an update from Google PR in the article, was not correct. Part of the reason seems to be that Google is in the middle of rebranding Location History to Timeline. This is all still confusing to me, but in a new and different way.)

The part of me which avoids cynicism wants to believe Google is doing its best and any confusion is purely accidental. That is hard to believe, though, given how often Google has run into legal trouble over location data and, consequently, how many in-house lawyers must have reviewed this update. This kind of granular adjustment is often worse for users than grouped preferences — consider how often users are forced to individually deselect checkboxes on cookie consent forms, for example. Without strong prophylactic privacy laws, corporations like Google will continue to dazzle people into believing they have meaningful control over their privacy.