Meta’s Red Card

September 16, 2024 ⌥

It was only a couple of weeks ago when Mark Zuckerberg wrote in a letter to U.S. lawmakers about his regret in — among other things — taking officials at their word about Russian election meddling in 2020. Specifically, he expressed remorse for briefly demoting a single link to a then-breaking New York Post story involving data it had obtained from a copy of the hard drive of a laptop formerly belonging to Hunter Biden, the son of the now-U.S. president.

At the time, some U.S. officials were concerned about the possibility it was partly or wholly Russian disinformation. This was later found to be untrue. In his letter, Zuckerberg wrote “in retrospect, we shouldn’t have demoted the story” and said the company has “changed [its] policies and processes to make sure this doesn’t happen again”.

To be clear, the laptop story was not a hoax and social media platforms ultimately erred in the decisions they made, but their policies were not inherently unfair and were reasonably cautious in their handling of the story. Nevertheless, the phrase “Hunter Biden’s laptop” is now a kind of shibboleth among those who believe there is a mass censorship campaign by disinformation researchers, intelligence agencies, and social media companies. That group includes people like Jim Jordan, to whom Zuckerberg addressed his obsequious letter. Surely, he was no longer taking U.S. officials at their word, and would be shrugging off their suggestions for platform moderation.

Right?

Kevin Collier and Phil Helsel, NBC News:

Social media giant Meta announced Monday that it is banning Russian media outlet RT, days after the Biden administration accused RT of acting as an arm of Moscow’s spy agencies.

[…]

U.S. officials allege that in Africa, RT is behind an online platform called “African Stream” but hides its role; that in Germany, it secretly runs a Berlin-based English-language site known as “Red”; and that in France, it hired a journalist in Paris to carry out “influence projects” aimed at a French-speaking audience.

As of writing, the Instagram and Threads accounts for Red are still online, but its Facebook page is not. A June report in Tagesspiegel previously connected Red to RT.

But I could not find any previous reporting connecting African Stream to Russia before U.S. officials made that claim. Even so, without corroborating evidence, Meta dutifully suspended African Stream’s presence on its platforms, which appeared to be active as of Friday.

Meta should — absolutely — do its best to curtail governments’ use of its platforms to spread propaganda and disinformation. All platforms should do so. I also hope it was provided more substantial evidence of RT’s involvement in African Stream. By that standard, it was also reasonable — if ultimately wrong — for it to minimize the spread of the Post story in 2020 based on the information it had at the time.

For all Zuckerberg’s grovelling to U.S. lawmakers, Meta ultimately gets to choose what is allowed on its platforms and what is not. It is right for it to be concerned about political manipulation. But this stuff is really hard to moderate. That is almost certainly why it is deprioritizing “political” posts — not because they do not get engagement or that the engagement they do get is heated and negative, but because it risks allegations of targeted censorship and spreading disinformation. Better, in Meta’s view, to simply restrict it all. Zuckerberg has figured out Meta is just as valuable when it does not react to criticism.

What I am worried about is the rising tension between the near-global scope of social media platforms and the parallel attempts by governments to get them to meet local expectations. Many of these platforms are based in the U.S. and have uncomfortably exported those values worldwide. Meta’s platforms are among the world’s most-used, so it is often among the most criticized. But earlier this month, X was banned in Brazil. The U.S. is seeking to ban TikTok and, based on a hearing today, it may well succeed.

It is concerning these corporations have such concentrated power, but I also do not think it makes sense to either treat them as common carriers, nor for them to be moderated in other countries as they would in the United States. I am more supportive of decentralized social software based on protocols like ActivityPub. Those can be, if anything, even more permissive and even harder to moderate. That also makes them more difficult for governments to restrict them — something which I support, but I know is not seen universally as the correct choice. They minimize the control of a single party’s decisions and, with it, help reduce the kinds of catastrophes we have seen from the most popular days of Facebook and Twitter.

Surely there will be new problems with which to contend, and perhaps it will have been better for there to be monolithic decision-makers after all. But it is right to try something different, and I am glad to see support building in different expressions. It is an exciting time for the open web.

Even so, I still wish for a good MacOS Bluesky client.

September 16, 2024 ⌥

Apple’s Lawsuit Against NSO Group Has Been Dropped at the Company’s Request, but Meta’s Continues ⇥ forbiddenstories.org

Stephanie Kirchgaessner, the Guardian, in February:

NSO Group, the maker of one the world’s most sophisticated cyber weapons, has been ordered by a US court to hand its code for Pegasus and other spyware products to WhatsApp as part of the company’s ongoing litigation.

Phineas Rueckert and Karine Pfenniger, Forbidden Stories, in July:

In July 2020, about nine months after WhatsApp sued NSO in a California court, the Israeli government requested for files in NSO’s office in Israel to be seized as NSO faced a potential discovery process, the consortium found. This pre-trial procedure, typical to common law jurisdictions such as the U.S., can lead to sensitive, internal documents being produced in court through a subpoena. Israel feared that these documents, if included in the lawsuit through discovery, could reveal information about Israeli state secrets.

The state of Israel also maneuvered to keep the seizure under wraps. Israel filed a gag order on information pertaining to the seizure of files at NSO’s offices, effectively preventing Israeli media from publishing information about the seizure. This publication ban, which has been in place for more than a year, aimed to protect Israel’s national security and foreign relations. The gag order referred to the WhatsApp litigation, suggesting that it could have been issued in response to the lawsuit.

Joseph Menn, Washington Post:

Apple asked a court Friday to dismiss its three-year-old hacking lawsuit against spyware pioneer NSO Group, arguing that it might never be able to get the most critical files about NSO’s Pegasus surveillance tool and that its own disclosures could aid NSO and its increasing number of rivals.

Apple’s request (PDF) cites the co-published Guardian version of the above Forbidden Stories reporting in saying it does not believe information it would produce in this suit would be held securely, and it worries about how forthcoming NSO Group has been. It also downplays the effects of a successful suit — a win would, according to Apple, “no longer have the same impact as it would have had in 2021” because there are plenty of NSO Group competitors.

WhatsApp appears to be continuing its suit against NSO Group. On the same day Apple filed its request to dismiss its case, WhatsApp attorneys were scheduling depositions (PDF). I hope Meta does not shy away from fully litigating this. It is important to hold vendors like these accountable for the abuse of their software.

September 13, 2024 ⌥

A Cartoon Butt Clenching a Bar of Soap ⇥ arstechnica.com

Nate Anderson, Ars Technica:

The great irony of online advertising these days is that it’s often claimed to be “targeted,” mining our personal and demographic information to serve us the ads that we allegedly want to see. Wouldn’t I prefer to view ads “relevant to my interests”? Maybe. But I can say with confidence that after two decades of being “extremely online” for work, the number of ads I have voluntarily and enthusiastically clicked upon must number in the low double digits.

Instead, the engines powering these ad networks continue to bombard me with two kinds of ads: 1) those that are wholly irrelevant to my interests and 2) those that are relevant to my interests because they display the exact product I once looked at in some online store. Ad targeting companies may “know a lot about me,” but they don’t know me in any truly useful way.

Ad tech companies will never have enough information about us to ensure truly reliable targeting, but they have more than enough to be a privacy nightmare unlike anything the world has seen before.

Update: This felt familiar to me. Four years ago, a similar butt-themed ad — this time for pyjamas with a flap — followed Vice writer Kate Dries around the web.

September 13, 2024 ⌥

To Forget is an Ethical Act ⇥ emilygorcenski.com

Emily Gorcenski decided to prune her dormant Twitter/X account, one post at a time:

Revisiting every post came with emotional baggage. Many of the posts were cringe; several were from stupid internet arguments. Others were painful to watch, dredging up traumatic experiences or memories of loved ones who’ve passed. But reading them was also a unique and worthwhile experience. It gave me the opportunity to reflect on what I’ve learned from 10+ years of microblogging and, hopefully, has made me a better person. So that’s what this post is about: a little self-retrospective on what brought me to where I am and, by extension, a little clue of where I might be going. Here’s what I’ve learned.

I am sure there would be a lot to learn if I decided it would be worth going through my own fifteen-and-a-bit years of tweets, but I would not wish that upon my worst enemy. Kudos to Gorcenski for this thoughtful reflection.

September 12, 2024 ⌥

The U.S. Government Begins Making Its Case Against Google’s Ad Business ⇥ reuters.com

Google, fresh from being found to have an illegal monopoly in search, is now facing another trial over its dominance in internet advertising.

Jody Godoy, Reuters:

Prosecutors say Google has largely dominated the technological infrastructure that funds the flow of news and information on websites through more than 150,000 online ad sales every second.

Google used classic monopoly-building tactics of eliminating competitors through acquisitions, locking customers into using its products, and controlling how transactions occurred in the online ad market, Julia Tarver Wood, an attorney with the Justice Department’s antitrust division, said in an opening statement.

“Google is not here because they are big, they are here because they used that size to crush competition,” she said.

Allison Schiff, AdExchanger:

As one ad tech executive close to the trial told AdExchanger on background, learning more about Project Poirot and Project Bernanke was like discovering “all that stuff we hear about late at night at the bar was really real.”

Many people seem to have taken for granted that it is natural for a marketplace as vast and as borderless as the internet provides to be so largely dominated by a small handful of parties in any given domain. Perhaps this is actually how things ought to shake out. But perhaps it is also right for governments to test this theory. It may take a while for legal efforts to bear fruit — Google’s lawyer called it a “time capsule” — and even longer for the power of these giant businesses to wane. That does not mean it is misguided to understand whether the behaviour of giants like Apple, Google, and Meta has been legal.

September 12, 2024 ⌥

Wired’s Look at Apple’s Private Cloud Compute ⇥ wired.com

Lily Hay Newman, of Wired, interviewed Apple’s Craig Federighi about Private Cloud Compute, the company’s custom system for processing Apple Intelligence requests unable to be handled on-device. Federighi explains how it all works in more simplified language, which is helpful, but I am eager to hear from third-parties about the real privacy differences.

There is an interesting two-part media story to this article, too. First, this little error:

Apple says it is still committed to doing as much Apple Intelligence processing as possible on-device, and a brand new iPhone 16 with its A18 chip, for example, will be able to do more AI processing locally than an iPhone 15 with an A16 chip. […]

I would hope so — an iPhone 15 with an A16 chip is not compatible with Apple Intelligence. An iPhone 15 Pro and its A17 Pro chip would be a better comparison. I do not know whether this error is Apple’s or the reporter’s, but it has survived a full day since the article’s publication.

Next, this paragraph:

Those who do get access to Apple Intelligence will have the ability to do far more than they could with past versions of iOS, from taking advantage of writing tools to photo analysis. Federighi says that his family celebrated their dog’s recent birthday with an Apple Intelligence–generated Image Playground creation shared exclusively with WIRED. […]

This paragraph was edited — Wired appended a cheeky note to the article saying it “was updated with clarification on the Apple Intelligence-generated image Federighi created for his dog’s birthday and additional confirmation that she is a very good dog”. The original version read:

Those who do get access to Apple Intelligence will have the ability to do far more than they could with past versions of iOS, from writing tools to photo analysis. Federighi says that his family celebrated their dog’s recent birthday with an Apple Intelligence-generated GenMoji (viewed and confirmed to be very cute by WIRED). […]

“Cute” is subjective; I am less amused. The referenced article was also updated after it was published “to add the name of Federighi’s dog”. Who requested that change, I wonder?

September 11, 2024 ⌥

A.I. Corporate Bullshit ⇥ adactio.com

Nathan J. Robinson, of Current Affairs, reviewing “Corporate Bullshit” by Nick Hanauer, Joan Walsh, and Donald Cohen last year:

Over the last several decades, we have been told that “smoking doesn’t cause cancer, cars don’t cause pollution, greedy pharmaceutical companies aren’t responsible for the crisis of opioid addiction.” Recognizing the pattern is key to spotting “corporate bullshit” in the wild, and learning how to spot it is important, because, as the authors write, the stories told in corporate propaganda are often superficially plausible: “At least on the surface, they offer a civic-minded, reasonable-sounding justification for positions that in fact are motivated entirely by self-interest.” When restaurant owners say that raising the minimum wage will drive their labor costs too high and they’ll be forced to cut back on employees or close entirely, or tobacco companies declare their product harmless, those things could be true. They just happen not to be.

Via Cory Doctorow.

Jeremy Keith:

I’ve noticed a really strange justification from people when I ask them about their use of generative tools that use large language models (colloquially and inaccurately labelled as artificial intelligence).

I’ll point out that the training data requires the wholesale harvesting of creative works without compensation. I’ll also point out the ludicrously profligate energy use required not just for the training, but for the subsequent queries.

And here’s the thing: people will acknowledge those harms but they will justify their actions by saying “these things will get better!”

This piece is making me think more about my own, minimal use of generative features. Sure, it is neat that I can get a more accurate summary of an email newsletter than a marketer will typically write, or that I can repair something in a photo without so much manual effort. But this ease is only possible thanks to the questionable ethics of A.I. training.

Jake Evans, ABC News:

Facebook has admitted that it scrapes the public photos, posts and other data of Australian adult users to train its AI models and provides no opt-out option, even though it allows people in the European Union to refuse consent.

[…]

Ms Claybaugh [Meta’s global privacy policy director] added that accounts of people under 18 were not scraped, but when asked by Senator Sheldon whether public photos of his own children on his account would be scraped, Ms Claybaugh acknowledged they would.

This is not ethical. Meta has the ability to more judiciously train its systems, but it will not do that until it is pressured. Shareholders will not take on that role. They have been enthusiastically boosting any corporation with an A.I. announcement. Neither will the corporations themselves, which have been jamming these features everywhere — there are floating toolbars, floating panels, balloons, callouts, and glowing buttons that are hard to ignore even if you want to.

September 11, 2024 ⌥

On the iPhone 16 Pro’s Cameras ⇥ aows.co

Adrian Vila:

I said it a year ago, and I still think Apple made a mistake with the 120mm lens. The current lineup of 13mm, 24mm, and 120mm leaves a huge gap between the main and telephoto lenses, missing out on key and very useful focal lengths for everyday situations. I’d rather see a 75mm lens on a 48MP sensor, with the ability to reach 120mm using the fancy cropping the main sensor has.

Last year was my iPhone purchasing year. I went from the 52mm-equivalent “2×” camera available on the 12 Pro to a 77mm-equivalent “3×” camera. Even though I understand the difference between those two focal lengths, I was still surprised by how much of a radical change it was. Relatively tight interior spaces, like those in my house, became even tighter; food photography required me to stand up and back away from the table.

For my purposes, I agree with Vila — the 77mm-equivalent is most often about as tight as I want to go with my iPhone. Perhaps it makes sense to give people two extreme lenses in the ultra-wide and 120mm-equivalent “5×” telephoto; maybe the in-between focal lengths are just too subtle for some people. There are use cases for this more extreme telephoto, like at concerts or getting some nice compression in cities. But the iPhone is now missing a perfect focal length for portraits, and that seems like a glaring omission.

September 10, 2024 ⌥

Welcome Back, 42 Millimetre Apple Watch ⇥ mastodon.social

David Smith:

An interesting point of comparison. The Series 10 “Small” (42mm) Apple Watch is very nearly the same physical footprint of the “Large” (42mm) Series 1 Apple Watch. Though obviously with a much larger screen given the rounded shape and drastic reduction in bezel thickness over the last decade.

These dimensions are remarkably close: the smallest Series 10 is 0.4mm narrower and 0.8mm thinner than the largest Series 1. Even commentators of the original Apple watch — like Benjamin Clymer and Serenity Caldwell — said they would feel comfortable with either, and the larger model did not feel too big.

September 9, 2024 ⌥

Amazon Says It Will Keep Selling Taxi Signs Because It Is Not a Crime ⇥ cbc.ca

Sophia Harris, CBC News:

The taxi scam has several variations but typically ends the same way: the victim pays with a debit card, then the scammer secretly steals it and hands the victim a similar but fake card. Shortly thereafter, money disappears from the victim’s account.

[…]

CBC News bought a taxi sign from Amazon for $35. It has a magnetic strip on the bottom, so it easily sticks to the top of a car.

[…]

But Amazon told Way — and CBC News — the signs will remain on its site, because the company isn’t breaking any rules.

I cannot think of a legitimate reason someone would purchase a taxi sign off Amazon.

Just because something is not a crime, it does not mean it is okay. Amazon, like most other massive corporations, tends to argue it can regulate itself and does not need government intervention. Fine — prove it.

September 9, 2024 ⌥

Camera Control on the iPhone 16 Line ⇥ apple.com

Apple:

Camera Control — a result of thoughtful hardware and software integration — elevates the camera experience on the iPhone 16 lineup. It is packed with innovation, including a tactile switch that powers the click experience, a high-precision force sensor that enables the light press gesture, and a capacitive sensor that allows for touch interactions. […]

Later this year, Camera Control will unlock visual intelligence to help users learn about objects and places faster than ever before. […]

Nilay Patel and Alison Johnson, the Verge:

I ran into Apple’s Phil Schiller, and we chatted briefly about the Camera Control button. I wanted to know about the balance of using the button as a classic camera control versus the beginning of the camera itself becoming an input method for Apple Intelligence, and he told me that it was really both, which is fascinating.

The positioning of this question is what is fascinating to me — far more so than Schiller’s confirmation. Apple could have added a hardware camera button at any time in the iPhone’s history. It did not until it wanted to use the camera for things not directly concerning photography and videography. Oh, it has those features too, of course, but it also makes the buttons down the right-hand side of this year’s iPhone line into dedicated Apple Intelligence launchers.

The third-party case story is going to be interesting. Apple’s own cases have a button with “sapphire crystal, coupled to a conductive layer” to create a passthrough experience. Launch-day third-party options, like those from Otterbox, appear to simply have a cutout for that button. Will Apple sell accessory makers the dedicated button part, I wonder? I doubt it, but you never know.

September 6, 2024 ⌥

A.I. and Art ⇥ newyorker.com

Ted Chiang, the New Yorker:

It’s harder to imagine a program that, over many sessions, helps you write a good novel. This hypothetical writing program might require you to enter a hundred thousand words of prompts in order for it to generate an entirely different hundred thousand words that make up the novel you’re envisioning. It’s not clear to me what such a program would look like. Theoretically, if such a program existed, the user could perhaps deserve to be called the author. But, again, I don’t think companies like OpenAI want to create versions of ChatGPT that require just as much effort from users as writing a novel from scratch. The selling point of generative A.I. is that these programs generate vastly more than you put into them, and that is precisely what prevents them from being effective tools for artists.

Matt Muir, writer of the excellent Web Curios newsletter:

[…] Broadly speaking I agree with some of the points he makes, specifically about the requirement for art to have an element of intentionality which is necessarily absent from anything made by (current generative) AI being as all it is is maths, and maths cannot have intent. Equally, though, Chiang concedes that artists have made, are making, and will continue to make, work *in conjunction with* non-intentional systems, and that these works are perfectly capable of being considered as ‘art’. […]

Adi Robertson on Bluesky:

I can hazard lots of guesses why, but it’s striking that virtually none of the “can AI do art” conversation focuses on the most interesting examples I’ve seen, in which the interactive conversation between user and machine — rather than the end output — *is* the art.

Robertson points to the Are You the Asshole bot and the Hey Robot game as two examples, both of which are creative explorations of human–A.I. interaction. Whether those conversations are considered “art” is something I will leave others to decide because I spent a bachelor’s degree hearing hundreds of people asking that question and I lost my patience for it.

Robertson’s observation is a spiritual successor to my issue with Instagram bait art installations: neither are necessarily cheapening art, but I wish artists treated social media and, now, A.I. with less formalism and more conceptualism. Artists can eke compelling works out of any medium. In fact, the very suspicion of A.I.’s involvement in art seems likely to lend itself to surprising and moving works, with suitably talented artists.

September 6, 2024 ⌥

Grain2Pixel, a Photoshop Script for Converting Film Negative Scans ⇥ grain2pixel.com

I got a batch of film scans from the developer today and realized I needed a better process for converting them — better, that is, than the way I had been doing it, which was to flip the curves in Lightroom and then do all my corrections in reverse.

I played around with the Filmomat SmartConvert demo but I did not like the workflow enough to consider paying for it. I really like the results I got from Negative Lab Pro and I think the USD $99 price tag is reasonable. However, its main selling point — that DNG scans remain in DNG format — is also its drawback: your workflow is still going to be the reverse of what you expect because, under the hood, the image is still a negative.

That pushed me to trying Grain2Pixel which, from a getting-started perspective, is more cumbersome than the other two options, particularly as MacOS is alarmed you are trying to use unsigned software.

But once you get that sorted out and install the script, it makes quick work out of batch processing a folder of DNGs into TIFF images. Then you can import them into Lightroom and make corrections in positive colour. You do not need to worry too much about a loss of range — TIFF is plenty flexible in post, at least for my amateur purposes. I am very happy with the resulting images.

September 4, 2024 ⌥

The Surprisingly Big Business of Library Ebooks ⇥ newyorker.com

Danielle Deschamps, in the conclusion to a rather interesting chapter from “Contemporary Issues in Collection Management”, hosted by Open Education Alberta:

Ebook licensing agreements have become the widespread norm for library ebook access. Yet, between libraries and publishers, these agreements, the terms of which are set by publishers, have devolved to an extent that libraries are struggling to maintain their access to ecollections. Publishers perceive libraries as harming their bottom lines and libraries are in a particularly vulnerable place, without much negotiating power. However, there are several optional ways for public libraries to move forward, in effort of balancing their financial capacity while maintaining their ethical principle of respecting intellectual property rights. […]

The subsequent chapter specifically about ebook pricing is also a terrific primer.

Daniel A. Gross, writing in September 2021, in the New Yorker:

To illustrate the economics of e-book lending, the N.Y.P.L. sent me its January, 2021, figures for “A Promised Land,” the memoir by Barack Obama that had been published a few months earlier by Penguin Random House. At that point, the library system had purchased three hundred and ten perpetual audiobook licenses at ninety-five dollars each, for a total of $29,450, and had bought six hundred and thirty-nine one- and two-year licenses for the e-book, for a total of $22,512. Taken together, these digital rights cost about as much as three thousand copies of the consumer e-book, which sells for about eighteen dollars per copy. As of August, 2021, the library has spent less than ten thousand dollars on two hundred and twenty-six copies of the hardcover edition, which has a list price of forty-five dollars but sells for $23.23 on Amazon. A few thousand people had checked out digital copies in the book’s first three months, and thousands more were on the waiting list. (Several librarians told me that they monitor hold requests, including for books that have not yet been released, to decide how many licenses to acquire.)

If you want to know why publishers so aggressively fought the Internet Archive on its model of lending out scanned copies of physical books, this is the reason. Publishers have created a model which fundamentally upsets a library’s ability to function. There is no scarcity in bytes, so publishers have created a way to charge more for something limitless, weightless, with nearly no storage costs.

What the Internet Archive did was perhaps a legal long-shot, and I worry about the effects of this suit and the one over shellac 78s. But it is hard not to see publishers as the real villains in this mess. They are consolidating power and charging even legitimate libraries unreasonable amounts of money for electronic copies of books which the publishers and their intermediaries ultimately still control.

September 4, 2024 ⌥

RationalWiki Is Being Sued ⇥ rationalwiki.org

At the beginning of August, Nassim Haramein sued RationalWiki on charges of defamation, conspiracy, and invasion of privacy. Regardless of the merits of the suit — I write, trying not to fall afoul of an obviously litigious individual — RationalWiki is a small, volunteer-run operation and will need legal representation to avoid losing next week by default. The site is currently soliciting donations.

I think the world is better for having RationalWiki in it. If you have the means and would like to chip in, I am sure the administrators there would appreciate it.

Update: RationalWiki has been SLAPP-ed into settling. Donations will go toward a proper legal fund.

September 3, 2024 ⌥

Here Are Over Twenty Examples of Google Employees Trying to Avoid Creating Antitrust Evidence ⇥ 404media.co

Jason Koebler, 404 Media:

The chats show 22 instances in which one Google employee told another Google employee to turn chat history off. In total, the court has dozens of specific employees who have told others to turn history off in DMs or broader group chats and channels. The document includes exchanges like this (each exchange includes different employees) […]

These examples are equal parts amusing, blatant, and telling. I doubt this is isolated; there are probably similar policies standard at other companies. But apparently this was part of Google’s new employee training.

September 3, 2024 ⌥

The Economist: Big, Heavy Cars Are Killing More People ⇥ economist.com

The Economist:

So how big is too big? At what point do the costs of the heaviest vehicles — measured in lives lost — vastly exceed their benefits? To answer this question, The Economist compiled ten years’ worth of crash data from more than a dozen states. Like the data compiled by Messrs Anderson and Auffhammer, our figures come from reports filed by police officers, who are tasked with recording information about car crashes when called to the scene. Although all states collect such data, we focus on those that collect the most detailed figures and share them with researchers. The resulting dataset, which covers more than a third of America’s population, provides us with a sample that is both big and representative.

The results? According to the Economist, “if the heaviest tenth of vehicles in America’s fleet were downsized […] road fatalities in multi-car crashes — which totaled 19,081 in 2023 — could be reduced by 12%, or 2,300, without sacrificing the safety of any cars involved”.

Andre Mayer and Emily Chung, reporting for CBC News in June:

But the ubiquity of SUVs and trucks isn’t an accurate reflection of what people want to drive, say industry analysts.

The trend has been greatly influenced by a combination of savvy marketing, government regulations that incentivize bigger vehicles and limited supply of more modest ones.

Indeed, much of it is driven by one simple economic fact.

“Smaller cars are less profitable,” said Stephanie Brinley, associate director at U.S.-based transportation consultancy S&P Global Mobility.

People are guided to purchase an SUV or truck in the United States and Canada because most cities oblige us to own a vehicle of some kind, but inexpensive cars are not generally available, and other people drive oversized SUVs and trucks which makes us scared of driving anything smaller. Repeat until around 80% of new vehicle sales are various kinds of SUVs and trucks.

This forced market is dangerous for everybody except for those who are inside a large SUV or truck. It means headlights from oncoming traffic at eye level. It means small roads are less navigable and parking spaces need to be made larger. It means roads feel more dangerous so fewer people feel comfortable walking or cycling. It means more people are seriously injured and die. All because these vehicles are more profitable, many cities are inaccessible by other means, automakers have artificially constrained their wares, and people feel roads are competitive instead of cooperative.

Update: The U.S. National Highway Traffic Safety Administration is proposing rules to comply with international pedestrian protection standards, even for larger vehicles. A good first step on a long and difficult path, and one which will have effects here as the Canadian auto market is more-or-less an extension of our southern neighbour’s.