Daily Shaarli

December 10, 2023

Pluralistic: “If buying isn’t owning, piracy isn’t stealing” (08 Dec 2023) – Pluralistic: Daily links from Cory Doctorow

Pluralistic: "If buying isn't owning, piracy isn't stealing" (08 Dec 2023)

20 years ago, I got in a (friendly) public spat with Chris Anderson, who was then the editor in chief of Wired. I'd publicly noted my disappointment with glowing Wired reviews of DRM-encumbered digital devices, prompting Anderson to call me unrealistic for expecting the magazine to condemn gadgets for their DRM:

https://longtail.typepad.com/the_long_tail/2004/12/is_drm_evil.html

I replied in public, telling him that he'd misunderstood. This wasn't an issue of ideological purity – it was about good reviewing practice. Wired was telling readers to buy a product because it had features x, y and z, but at any time in the future, without warning, without recourse, the vendor could switch off any of those features:

https://memex.craphound.com/2004/12/29/cory-responds-to-wired-editor-on-drm/

I proposed that all Wired endorsements for DRM-encumbered products should come with this disclaimer:

WARNING: THIS DEVICE’S FEATURES ARE SUBJECT TO REVOCATION WITHOUT NOTICE, ACCORDING TO TERMS SET OUT IN SECRET NEGOTIATIONS. YOUR INVESTMENT IS CONTINGENT ON THE GOODWILL OF THE WORLD’S MOST PARANOID, TECHNOPHOBIC ENTERTAINMENT EXECS. THIS DEVICE AND DEVICES LIKE IT ARE TYPICALLY USED TO CHARGE YOU FOR THINGS YOU USED TO GET FOR FREE — BE SURE TO FACTOR IN THE PRICE OF BUYING ALL YOUR MEDIA OVER AND OVER AGAIN. AT NO TIME IN HISTORY HAS ANY ENTERTAINMENT COMPANY GOTTEN A SWEET DEAL LIKE THIS FROM THE ELECTRONICS PEOPLE, BUT THIS TIME THEY’RE GETTING A TOTAL WALK. HERE, PUT THIS IN YOUR MOUTH, IT’LL MUFFLE YOUR WHIMPERS.

Wired didn't take me up on this suggestion.

But I was right. The ability to change features, prices, and availability of things you've already paid for is a powerful temptation to corporations. Inkjet printers were always a sleazy business, but once these printers got directly connected to the internet, companies like HP started pushing out "security updates" that modified your printer to make it reject the third-party ink you'd paid for:

Now, this scam wouldn't work if you could just put things back the way they were before the "update," which is where the DRM comes in. A thicket of IP laws make reverse-engineering DRM-encumbered products into a felony. Combine always-on network access with indiscriminate criminalization of user modification, and the enshittification will follow, as surely as night follows day.

This is the root of all the right to repair shenanigans. Sure, companies withhold access to diagnostic codes and parts, but codes can be extracted and parts can be cloned. The real teeth in blocking repair comes from the law, not the tech. The company that makes McDonald's wildly unreliable McFlurry machines makes a fortune charging franchisees to fix these eternally broken appliances. When a third party threatened this racket by reverse-engineering the DRM that blocked independent repair, they got buried in legal threats:

Everybody loves this racket. In Poland, a team of security researchers at the OhMyHack conference just presented their teardown of the anti-repair features in NEWAG Impuls locomotives. NEWAG boobytrapped their trains to try and detect if they've been independently serviced, and to respond to any unauthorized repairs by bricking themselves:

Poland is part of the EU, meaning that they are required to uphold the provisions of the 2001 EU Copyright Directive, including Article 6, which bans this kind of reverse-engineering. The researchers are planning to present their work again at the Chaos Communications Congress in Hamburg this month – Germany is also a party to the EUCD. The threat to researchers from presenting this work is real – but so is the threat to conferences that host them:

20 years ago, Chris Anderson told me that it was unrealistic to expect tech companies to refuse demands for DRM from the entertainment companies whose media they hoped to play. My argument – then and now – was that any tech company that sells you a gadget that can have its features revoked is defrauding you. You're paying for x, y and z – and if they are contractually required to remove x and y on demand, they are selling you something that you can't rely on, without making that clear to you.

But it's worse than that. When a tech company designs a device for remote, irreversible, nonconsensual downgrades, they invite both external and internal parties to demand those downgrades. Like Pavel Chekov says, a phaser on the bridge in Act I is going to go off by Act III. Selling a product that can be remotely, irreversibly, nonconsensually downgraded inevitably results in the worst person at the product-planning meeting proposing to do so. The fact that there are no penalties for doing so makes it impossible for the better people in that meeting to win the ensuing argument, leading to the moral injury of seeing a product you care about reduced to a pile of shit:

But even if everyone at that table is a swell egg who wouldn't dream of enshittifying the product, the existence of a remote, irreversible, nonconsensual downgrade feature makes the product vulnerable to external actors who will demand that it be used. Back in 2022, Adobe informed its customers that it had lost its deal to include Pantone colors in Photoshop, Illustrator and other "software as a service" packages. As a result, users would now have to start paying a monthly fee to see their own, completed images. Fail to pay the fee and all the Pantone-coded pixels in your artwork would just show up as black:

Adobe blamed this on Pantone, and there was lots of speculation about what had happened. Had Pantone jacked up its price to Adobe, so Adobe passed the price on to its users in the hopes of embarrassing Pantone? Who knows? Who can know? That's the point: you invested in Photoshop, you spent money and time creating images with it, but you have no way to know whether or how you'll be able to access those images in the future. Those terms can change at any time, and if you don't like it, you can go fuck yourself.

These companies are all run by CEOs who got their MBAs at Darth Vader University, where the first lesson is "I have altered the deal, pray I don't alter it further." Adobe chose to design its software so it would be vulnerable to this kind of demand, and then its customers paid for that choice. Sure, Pantone are dicks, but this is Adobe's fault. They stuck a KICK ME sign to your back, and Pantone obliged.

This keeps happening and it's gonna keep happening. Last week, Playstation owners who'd bought (or "bought") Warner TV shows got messages telling them that Warner had walked away from its deal to sell videos through the Playstation store, and so all the videos they'd paid for were going to be deleted forever. They wouldn't even get refunds (to be clear, refunds would also be bullshit – when I was a bookseller, I didn't get to break into your house and steal the books I'd sold you, not even if I left some cash on your kitchen table).

Sure, Warner is an unbelievably shitty company run by the single most guillotineable executive in all of Southern California, the loathsome David Zaslav, who oversaw the merger of Warner with Discovery. Zaslav is the creep who figured out that he could make more money cancelling completed movies and TV shows and taking a tax writeoff than he stood to make by releasing them:

Imagine putting years of your life into making a program – showing up on set at 5AM and leaving your kids to get their own breakfast, performing stunts that could maim or kill you, working 16-hour days during the acute phase of the covid pandemic and driving home in the night, only to have this absolute turd of a man delete the program before anyone could see it, forever, to get a minor tax advantage. Talk about moral injury!

But without Sony's complicity in designing a remote, irreversible, nonconsensual downgrade feature into the Playstation, Zaslav's war on art and creative workers would be limited to material that hadn't been released yet. Thanks to Sony's awful choices, David Zaslav can break into your house, steal your movies – and he doesn't even have to leave a twenty on your kitchen table.

The point here – the point I made 20 years ago to Chris Anderson – is that this is the foreseeable, inevitable result of designing devices for remote, irreversible, nonconsensual downgrades. Anyone who was paying attention should have figured that out in the GW Bush administration. Anyone who does this today? Absolute flaming garbage.

Sure, Zaslav deserves to be staked out over an anthill and slathered in high-fructose corn syrup. But save the next anthill for the Sony exec who shipped a product that would let Zaslav come into your home and rob you. That piece of shit knew what they were doing and they did it anyway. Fuck them. Sideways. With a brick.

Meanwhile, the studios keep making the case for stealing movies rather than paying for them. As Tyler James Hill wrote: "If buying isn't owning, piracy isn't stealing":

Enshitification Big Corpo Planned Obsolescence

Médias : les premières expériences 100 % IA | Les Echos

Médias : les premières expériences 100 % IA

Certains sites expérimentent l'utilisation de robots conversationnels pour écrire tout leur contenu et se revendiquent comme tels. A l'inverse, d'autres le cachent, provoquant de l'inquiétude.

Par Marina Alcaraz

Publié le 1 déc. 2023 à 13:43Mis à jour le 6 déc. 2023 à 17:59

Dans la présentation du site, il y a leurs parcours, leurs centres d'intérêt, leurs photos… Ils ont un style bien à eux, et des couvertures de sujets spécifiques. Carine Gravée, Vianney Garet, Nina Gavetière ont des noms, mais ne sont pas réels : ce sont des journalistes-robots créés de toutes pièces.

Certains sites réalisés entièrement par l'intelligence artificielle (IA) commencent à apparaître. Certes, pour l'heure, les initiatives restent limitées. Mais certains tentent de créer une niche, d'expérimenter un nouveau mode de création des contenus, en le revendiquant clairement… ou pas.

Magazine en kiosque

C'est par exemple le cas du magazine (papier et Web) « LHC - Les Heures Claires », qui se présente comme le premier magazine français généré à 99 % par l'IA, lancé il y a quelques semaines. Comme un support classique, il comporte des actualités, des interviews (avec des questions posées par un robot), des pages consacrées à la mode ou un horoscope.

A la manoeuvre, Rémy Rostan, ancien photographe. « Je suis toujours assez surpris par ce que propose ChatGPT », avoue-t-il. Le magazine sponsorisé par Easy Partner, cabinet de recrutement spécialisé dans le numérique, a vocation à être lancé en kiosque au printemps avec une fréquence mensuelle. « Je vise la communauté des technophiles et des curieux », explique Rémy Rostan, qui espère atteindre les 20.000 ventes.

Autres exemples : Tech Generation et Cuisine Generation, lancés au printemps par un consultant spécialisé en innovation chez Viseo, Ari Kouts. Il a connecté le site tech avec différents journaux spécialisés dans ce secteur, comme TechCrunch. Chaque « journaliste » fictif (qui a son style bien à lui) reprend donc des articles de presse sur des sujets d'actualité (la crise chez OpenAI, des déclarations de Musk…), les réécrit en donnant la source et ce, sans aucune intervention humaine. Au final, quelques incohérences, des maladresses mais des articles qui ressemblent à certains billets de blogs.

Dans la cuisine, les « chefs » imaginent plusieurs recettes « et bon nombre sont plausibles et même bonnes, même si les temps de cuisson sont approximatifs », estime Ari Kouts. C'est plus à titre d'expérience que le consultant a lancé ces « médias », sans volonté de les monétiser. « Cela permet aussi de rappeler l'intérêt de l'analyse, de l'enquête journalistique que des robots ne peuvent pas faire », assure-t-il.

Les deux sites ont une petite audience (autour de 3.000 visites par mois) et ressortent quelquefois dans Google News ! Même si la probabilité est faible, dans ce cas, puisqu'il s'agit d'une expérimentation un peu comme un jeu, « les sources primaires pourraient empêcher ce type de pratiques en invoquant le parasitisme, c'est-à-dire s'approprier la valeur d'un article », indique Julien Guinot-Deléry, avocat chez Gide.

Craintes des professionnels

Mais il existe aussi des sites dont le mode de production a été passé sous silence. « Dans un groupe de travail de la Commission paritaire des publications et agences de presse, la crainte qu'il y ait des sites avec une forte composante d'IA a été évoquée », dit un professionnel. « On a tous ce risque en tête », appuie Pierre Pétillault, directeur de l'Alliance de la presse d'information générale.

Dans une étude récente, Newsguard a identifié foule de sites avec des articles réécrits avec l'IA (presque 600 à fin novembre !), sans supervision humaine. Et dans nombre de cas, ils bénéficient de publicité programmatique. Aux Etats-Unis, « Sports Illustrated » ou « TheStreet » (Arena Group) ont été pointés par une enquête du média Futurism. Des articles auraient été écrits par des IA et de faux profils de journalistes créés (avec les images achetées sur un site proposant des photos générées par IA), ce qui a provoqué la colère des journalistes. Le groupe de médias s'est défendu, indiquant avoir acheté certains papiers à une agence.

Artificial intelligence Enshitification Press