Skip to main content

Home/ QN2019/ Group items tagged management

Rss Feed Group items tagged

Aurialie Jublin

[Écologie] Le monde du logiciel est en train de se détruire... Manifeste pour... - 0 views

  • Tout est lent, et cela ne va pas dans le bon sens. Certaines voix s’élèvent. Je vous invite notamment à lire “Le désenchantement du logiciel”. Tout est insupportablement lent, tout est ÉNORME, tout finit par devenir obsolète… La taille des sites web explose. Un site web est aussi gros que le jeu Doom. Le phénomène ne touche pas que le Web mais aussi l’IoT, le mobile… Le saviez-vous ? Il faut 13% de CPU pour faire clignoter un curseur…
  • Tout grossit : la taille des applications, les données stockées, la taille des pages web, la mémoire des téléphones… Les téléphones ont maintenant 2 Go de mémoire, échanger une photo de 10 Mo par mail est maintenant classique… À la limite, cela ne serait peut-être pas un problème si tous les logiciels étaient utilisés, efficaces et performants… Mais cela n’est pas le cas,
  • Alors tout le monde se cale sur une lenteur. Tout est uniformément lent. On se cale sur cela et tout va bien. Être performant aujourd’hui, c’est arriver à atteindre un ressenti utilisateur qui correspond à cette lenteur uniforme. On élague les choses qui pourraient être trop visibles. Une page qui met plus de 20 secondes à se charger, c’est trop lent. Par contre, 3 secondes c’est bien. 3 secondes ? Avec les multicoeurs de nos téléphones/PC et les data centers partout dans le monde, le tout relié par des supers technologies de communication (4G, fibre…),c’est un peu bizarre non ? Si on regarde la débauche de ressources pour le résultat obtenu, 3 secondes, c’est énorme. D’autant plus que les bits circulent dans nos processeurs avec des unités de temps du niveau de la nanoseconde. Donc oui, tout est uniformément lent.
  • ...29 more annotations...
  • Il est nécessaire de revenir à plus d’efficience, de « challenger » les besoins en matériel, de redéfinir ce qu’est la performance. Tant que l’on se satisfera de cette lenteur uniforme avec des solutions pour ne pas ralentir plus (comme l’ajout de matériel), nous n’avancerons pas. La dette technique, notion largement assimilée par les équipes de développement, n’est malheureusement pas adaptée à ce problème (on y reviendra). Nous sommes sur une dette de ressources matérielles et de mauvaise adéquation entre le besoin utilisateur et la solution technique. On parle ici d’efficience et non pas uniquement de performance. L’efficience est une histoire de mesure du gaspillage. L’ISO définie l’efficience avec comme domaine : Time behaviour, Resource utilization et Capacity. Pourquoi ne pas pousser plus ces concepts ?
  • Développez du code Java pour un serveur J2EE ou pour un téléphone Android, ce n’est pas pareil. Des structures spécifiques existent par exemple pour traiter des données en Android mais les structures classiques sont toujours utilisées. Les développeurs ont perdu le lien avec le hardware. C’est malheureux car c’est passionnant (et utile) de savoir comment fonctionne un processeur. Pourquoi : abstraction et spécialisation (nous verrons cela plus loin). Car en perdant ce lien, on perd une des forces du développement. Ce lien est important chez les hackers ou chez les développeurs d’informatique embarquée mais malheureusement de moins en moins présent chez les autres développeurs.
  • Les pratiques devops pourraient répondre à cette perte de lien. Là, c’est pareil, nous n’allons pas jusqu’au au bout : généralement le devops va se focaliser à bien gérer le déploiement d’une solution logicielle sur une infrastructure mixte (matérielle et un peu logicielle). Il faudrait aller plus loin en remontant par exemple les métriques de consommation, en discutant sur les contraintes d’exécution… plutôt que de “scaler” juste parce que c’est plus simple.
  • Certaines technologies ont une utilité mais sont maintenant systématiquement utilisées. C’est le cas par exemple des ORM qui sont devenus systématiques. Aucune réflexion n’est faite sur son intérêt en début des projets. Résultat : on a rajouté une surcouche qui consomme, qu’il faut maintenir et des développeurs qui n’ont plus l’habitude d’effectuer des requêtes natives. Cela ne serait pas un problème si chaque développeur connaissait très bien le fonctionnement des couches d’abstraction : comment fonctionne HIBERNATE par exemple ? On s’appuie hélas de façon aveugle sur ces frameworks.
  • Le besoin d’abstraction est lié à un autre défaut: nous attendons toujours des outils miracles. La silver bullet qui améliorera encore plus nos pratiques. Le langage idéal, le framework pour aller plus vite, l’outil de gestion miracle des dépendances… C’est la promesse à chaque fois d’un nouveau framework : gagner du temps en développement, être plus performant… Et on y croit, on fonce. On abandonne les frameworks sur lesquels nous avions investi, sur lesquels on avait passé du temps… et on passe au nouveau. C’est le cas actuellement des frameworks JS. L’histoire du développement est pavé de framework oubliés, non maintenus, abandonnés… Nous sommes les champions pour réinventer ce qui existe déjà. Si on le gardait suffisamment longtemps, on aurait le temps de maîtriser un framework, de l’optimiser, de le comprendre. Mais ce n’est pas le cas. Et que l’on ne me dise pas que si on n’avait pas continuellement réinventé la roue, on aurait encore des roues en pierre… Innover serait d’améliorer les frameworks existants .
  • Sur les langages, c’est la même rengaine. Attention, je ne préconise pas de rester sur l’assembleur et sur le C… C’est le cas par exemple dans le monde Android, pendant plus de 10 ans les développeurs ont pu travailler sur des outils et des frameworks Java. Et comme cela, par magie, le nouveau Langage de la communauté est Kotlin. On imagine l’impact sur les applications existantes (si elles doivent changer), il faut recréer des outils, retrouver des bonnes pratiques… Pour quel gain?
  • Sincèrement, on ne voit aucun ralentissement sur les cycles de renouvellement des technologies. C’est toujours un rythme frénétique. Nous trouverons bien le Graal un jour. Le problème est ensuite l’empilement de ses technologies. Comme aucune ne meurt vraiment et que l’on en maintient toujours des parties, on développe d’autres couches pour s’adapter et continuer à maintenir ces bouts de code ou ces librairies. Le problèmen’est pas le code legacy,, c’est la glue que l’on développe autour qui pêche.
  • Au final, on passe du temps à résoudre des problèmes techniques internes, on recherche des outils pour résoudre les problèmes que l’on ajoute, on passe son temps à s’adapter à ses nouveaux outils, on ajoute des surcouches (voir chapitre précédent…) … et on n’a pas améliorer la qualité intrinsèque du logiciel ou les besoins auxquels on doit répondre.
  • Au final, le rythme frénétique des changements ne nous permet pas de nous stabiliser sur une technologie. J’avoue qu’en tant que vieux développeur que je suis, j’ai été découragé par le changement Java vers Kotlin pour Android. C’est peut-être pour certains de véritables challenges, mais quand je repense au temps que j’ai passé sur l’apprentissage, sur la mise en place d’outils
  • Ensuite, comme aucune solution ne meurt vraiment, que de nouvelles arrivent… on se retrouve avec des projets avec des multitudes de technologies à gérer avec les compétences associées aussi… On s’étonne ensuite que le marché du recrutement de développeur soit bouché. Pas étonnant.. Il y a énormément de développeurs mais il est difficile de trouver un développeur React avec 5 ans d’expérience qui connaîsse le Go. Le marché est fractionné, comme les technologies. C’est peut-être bon pour les développeurs car cela crée de la rareté et cela fait monter les prix, mais pas pour le projet !
  • Cependant, cet apprentissage continu des technologies pose le problème d’apprentissage de domaines transverses : accessibilité, agilité, performance… En effet, qu’est-ce qui nous prouve que les outils et les langages que nous allons choisir ne vont pas changer dans 4 ans ? Rust, Go… dans 2 ans ? Rien ne tend à donner une tendance.
  • On ne se remet pas en question car on veut s’amuser. Le fun est important, car si l’on s’ennuie dans son boulot, on va déprimer. Par contre, on ne peut pas, sous prétexte de vouloir du fun tout le temps, changer nos outils continuellement. Il y a un déséquilibre entre l’expérience du développeur et l’expérience de l’utilisateur. On veut du fun, mais qu’est-ce que cela va réellement apporter à l’utilisateur ? Un produit plus « joyeux » ? Non, nous ne sommes pas des acteurs.
  • On ne se pose pas de question sur l’éthique de notre domaine, sur sa durabilité… Cela vient peut-être du fait que notre domaine n’a pas réellement de code éthique (comme par exemple les médecins ou les avocats). Mais sommes-nous en tant que développeurs réellement libres si l’on ne peut pas avoir une autocritique ? Nous sommes peut être asservis à une cause portée par d’autres personnes ? Le problème n’est pas simple mais nous avons dans tous les cas une responsabilité. Sans code éthique, c’est le plus fort et le plus malhonnête qui est le plus fort. Le buzz et les pratiques pour manipuler les utilisateurs sont de plus en plus répandus. Sans Dark Pattern ton produit ne sera rien. Les plus gros (GAFA…) n’en sont pas arrivés là pour rien.
  • Est-ce que la solution est politique ? Il faut légiférer pour mieux gouverner le monde du logiciel. On le voit avec les dernières réponses législatives aux problèmes concrets : RGPD, notification des cookies… la source du problème n’est pas résolue. Peut-être parce que les politiques ne comprennent que très mal le monde du logiciel.
  • Car si cette structuration n’est pas faite, les développeurs vont perdre la main sur ce qu’ils font. Or le manque d’éthique de la profession est critiqué à l’extérieur. Rachel Coldicutt (@rachelcoldicutt) directrice de DotEveryOne, un think tank britannique qui promeut une technologie plus responsable, encourage à former des diplômés non-informaticiens qui traiteraient de ces problèmes (Voir plus précisément dans l’article d’Internet Actu). Pour poursuivre sur ce dernier article, cela serait dans la droite ligne de l’informatique, domaine issu du monde militaire où les ingénieurs et développeurs seraient formés à suivre des décisions et des commandements.
  • Le monde du logiciel s’intègre dans un système organisationnel classique. Grands groupes, sous-traitances via des ESN, web agencies… Tous suivent les mêmes techniques de gestion des projets informatiques. Et tout le monde va « dans le mur ». Aucune analyse sérieuse n’est faite sur le coût global d’un logiciel (TCO), sur son impact sur la société, sur son bénéfice, sa qualité… C’est la rapidité de release(Time to Market), la surcharge featurale (fonctionnelle), la productivité immédiate, qui comptent. Premièrement car les gens externes à ce monde ne connaissent que trop peu la technicité du logiciel et son monde. Il est virtuel donc simple (sic). Mais ce n’est pas le cas. Les écoles de commerce et autres usines de managers n’ont pas de cours de développement. Comment peuvent-ils bien diriger et piloter des projets ?
  • On continue a vouloir chiffrer des projets informatiques comme de simples projets alors que des mouvements comme le no estimate propose des approches innovantes. Les projets continuent d’échouer : le chaos report rapporte que simplement 30% des projets réussissent bien. Et face à cette mauvaise gouvernance, les équipes techniques continuent de se battre sur les technologies. Dommages collatéraux : la qualité, l’éthique, l’environnement… et au final l’utilisateur. Cela ne serait pas si critique si le logiciel n’avait pas un impact aussi fort sur le monde. Software eats the world… et oui, on le « bouffe »…
  • Si ces décisions absurdes arrivent, ce n’est pas uniquement la faute du développeur mais bien de l’organisation. Et qui dit organisation dit management (sous-différente forme). Si l’on revient au livre de Morel, il parle de piège cognitif dans lesquels les managers et les techniciens tombent souvent. C’est le cas de la navette Challenger qui a été quand même lancée malgré la connaissance du problème d’un joint défectueux. Les managers ont sous-évalué les risques et les ingénieurs ne les ont pas prouvés. Chacun a reproché à l’autre de ne pas fournir assez de preuves scientifiques. C’est souvent ce qui se passe dans les entreprises : des warnings sont levés par certains développeurs mais le management ne les prend pas assez au sérieux.
  • En même temps, on revient aux causes précédentes (silver bullet, on s’amuse…), il est nécessaire d’avoir une vraie ingénierie et une vraie analyse des technologies. Sans cela, les équipes techniques seront toujours non-écoutées par le management. Des outils et benchmark existent mais ils sont encore trop peu connus. Par exemple, Technologie Radar qui classe les technologies en terme d’adoption.
  • Ce phénomène de décision absurde est renforcé par le tissu complexe du développement logiciel : Les sociétés historiquement hors du numérique sous-traitent à des entreprises du numérique, les ESN sous-traitent aux freelances… Le partage de responsabilité technique / management est encore plus complexe et les décisions absurdes plus nombreuses.
  • Mais cela ne s’arrête pas là. On peut aussi voir l’usage de l’open-source comme une sorte de sous-traitance. Idem pour l’usage de framework. On est juste consommateur passif, on se déleste de plein de problématiques (qui ont un impact sur les ressources, la qualité…).
  • C’est d’autant plus facile que le domaine est passionnant et que la pratique des sides-projects, du temps passé sur les projets open-source hors des horaires de bureau est chose commune… La recherche de “fun” et le temps passé bénéficient alors plus aux organisations qu’aux développeurs. Difficile dans ce cas de chiffrer le coût réel d’un projet. Et pourtant, cela ne serait pas un problème si on arrivait à des logiciels « au top ».
  • Le développeur n’est ici plus un artisan du code, mais plutôt un pion dans un système critiquable du point de vue humain. Cela n’est pas visible, tout va bien et on s’amuse. En apparence seulement, car certains domaines du développement logiciel vont plus loin et rendent beaucoup plus visible cette exploitation : Le domaine du jeux-vidéo où les heures explosent.
  • Et donc, toutes ces maladresses (logiciel trop lourd, sans qualité…) se retrouvent chez les utilisateurs. Comme on doit releaser au plus vite les logiciels, que l’on ne tente pas de résoudre les inefficiences internes, et que l’on ne met pas plus de ressource pour faire de la qualité, on arrive à des logiciels médiocres. Mais on a tellement d’outils de monitoring et de suivi des utilisateurs pour détecter ce qui se passe directement chez eux qu’au final, on pense que ce n’est pas grave. Cela serait une bonne idée si les outils étaient bien utilisés. Or la multitude d’informations récoltées (en plus des bugs remontés par les utilisateurs) n’est que faiblement utilisée. Trop d’information, difficulté de cibler la vraie source du problème… on s’y perd et au final, c’est l’utilisateur qui trinque. Tous les logiciels sont maintenant en bêta-test. A quoi bon faire de la sur-qualité, autant attendre que l’utilisateur le demande. Et on revient ainsi au premier chapitre : un logiciel uniformément lent … et médiocre.
  • Heureusement, on est sauvé par la non-sensibilisation des utilisateurs au monde du logiciel. C’est un monde effectivement virtuel et magique qu’ils ont l’habitude d’utiliser. On leur a mis en main les outils mais sans notice explicative. Comment évaluer la qualité d’un logiciel, les risques sur l’environnement, les problèmes de sécurité… si l’on n’a pas des notions d’informatique, même rudimentaires ?
  • L’informatique du 21ème siècle est ce que l’agroalimentaire était pour les consommateurs au 20ème siècle. Pour des raisons de productivité, on a poussé des solutions médiocres avec un calcul court-termiste : mise sur le marché de plus en plus rapide, profit en hausse constante… agriculture intensive, malbouffe, pesticides… avec des impacts importants sur la santé, sur l’environnement… Les consommateurs savent maintenant (de plus en plus) les conséquences désastreuses de ces dérives, l’industrie agroalimentaire doit donc se réinventer, techniquement, commercialement et éthiquement. Pour le logiciel, quand les utilisateurs comprendront les tenants et les aboutissants des choix techniques, l’industrie du logiciel devra gérer les mêmes problèmes. En effet, le retour au bon sens et aux bonnes pratiques n’est pas une chose simple pour l’agroalimentaire. Dans l’IT, on commence à le voir avec ses conséquence sur la vie privée des utilisateurs (mais nous n’en sommes qu’aux balbutiements).
  • On va dans la mauvaise direction. L’industrie de l’informatique a déjà effectué dans les années 70 des erreurs avec des impacts non-négligeables. L’exclusion des femmes de l’informatique en fait partie. Non seulement cela a été fatal pour certaines industries mais on peut se poser la question de comment on peut maintenant adresser des réponses à seulement 50% de la population informatique, avec une représentativité très faible. Le chemin est maintenant difficile à retrouver.
  • Mais l’impact du monde IT ne s’arrête pas là. La source et le modèle d’une grosse partie de l’informatique sont issus de la Silicon valley. Si l’on écarte les gagnants de la Silicon Valley, les populations locales subissent la montée des prix, le déclassement, la pauvreté…
  • Dans sa lancée vers le progrès technique, le monde du logiciel crée aussi sa dette environnementale…
  •  
    "Le monde du logiciel va mal et si l'on n'agit pas, on risque de le regretter. Environnement, qualité, exclusion… Software Eats The World (Le logiciel mange le monde…) ? Oui un peu trop. "
Aurialie Jublin

Mastercard, Microsoft to Advance Digital Identity Innovations - 0 views

  • “Today’s digital identity landscape is patchy, inconsistent and what works in one country often won’t work in another. We have an opportunity to establish a system that puts people first, giving them control of their identity data and where it is used,” says Ajay Bhalla, president, cyber and intelligence solutions, Mastercard. “Working with Microsoft brings us one step closer to making a globally interoperable digital identity service a reality, and we look forward to sharing more very soon.”
  •  
    "Mastercard (NYSE: MA) and Microsoft (Nasdaq "MSFT" @microsoft) today announced a strategic collaboration to improve how people manage and use their digital identity. Currently, verifying your identity online is still dependent on physical or digital proof managed by a central party, whether it's your passport number, your proof of address, driver's license, user credentials or other means. This dependence places a huge burden on individuals, who have to successfully remember hundreds of passwords for various identities and are increasingly being subjected to more complexity in proving their identity and managing their data. Working together, Mastercard and Microsoft aim to give people a secure, instant way to verify their digital identity with whomever they want, whenever they want."
Aurialie Jublin

Amazon's system for tracking its warehouse workers can automatically fire them - MIT Te... - 0 views

  •  
    "A world where people are monitored and supervised by machines isn't confined to the realms of sci-fi. It's here now. Tough conditions: There have been many reports over recent years about unpleasant conditions workers face at Amazon warehouses. Employees are under pressure to pack hundreds of boxes per hour, and face being fired if they aren't fast enough. What's new: Documents obtained by The Verge show that it's far more common for people to be fired due to lack of productivity than outsiders realize. Roughly 300 people were fired at a single facility between August 2017 and September 2018 for that reason. And crucially, the documents show that much of the firing process is automated. The automation aspect: Amazon tracks every individual workers' productivity, and automatically generates warnings or even terminations without any input from supervisors, the company said. Managers can override the process, but it didn't say how regularly they do. Amazon says that the termination process can be appealed. Wider implications: Uber drivers have complained of "management by algorithm." This is not the first "fired by a computer" case but, as automation marches on, these stories are likely to become a lot more common."
Aurialie Jublin

Microsoft Bug Testers Unionized. Then They Were Dismissed - Bloomberg - 0 views

  • In California, Uber, Lyft, TaskRabbit, and a half-dozen other companies are lobbying to defang a court ruling that could make it difficult to avoid reclassifying such workers as employees. And in Washington, the Republican-dominated National Labor Relations Board has made moves to undo an Obama-era precedent that could make big employers legally liable for contract workers even if they have only indirect control over them.The GOP takeover in Washington is one reason the Temporary Workers of America, a union of bug testers for Microsoft Corp., gave up on what had been, for people in the software world, an almost unheard of unionization victory, says the group’s founder, Philippe Boucher.
  • Boucher and his ex-colleagues are among a growing population of tech workers, including many Uber drivers, Amazon.com warehouse loaders, and Google software engineers, who lack the rights and perks of those companies’ full-fledged employees.
  • Google parent Alphabet Inc. now has fewer direct employees than it does contract workers, some of whom write code and test self-driving cars.
  • ...2 more annotations...
  • “Companies are deciding they don’t want to make long-term commitments to people, and they’re using a variety of devices to shift that work out,” says David Weil, dean of Brandeis University’s social policy and management school who oversaw federal wage-and-hour enforcement during the Obama presidency.
  • To help demonstrate that Microsoft was a joint employer, the union provided documents such as an email appearing to show a Lionbridge manager sharing performance metrics with Microsoft counterparts and a list of Microsoft managers who worked in the same office and oversaw Lionbridge employees’ work—at least one of whom listed his management of contractors on his LinkedIn résumé.
Aurialie Jublin

An Apology for the Internet - From the People Who Built It - 1 views

  • There have always been outsiders who criticized the tech industry — even if their concerns have been drowned out by the oohs and aahs of consumers, investors, and journalists. But today, the most dire warnings are coming from the heart of Silicon Valley itself. The man who oversaw the creation of the original iPhone believes the device he helped build is too addictive. The inventor of the World Wide Web fears his creation is being “weaponized.” Even Sean Parker, Facebook’s first president, has blasted social media as a dangerous form of psychological manipulation. “God only knows what it’s doing to our children’s brains,” he lamented recently.
  • To keep the internet free — while becoming richer, faster, than anyone in history — the technological elite needed something to attract billions of users to the ads they were selling. And that something, it turns out, was outrage. As Jaron Lanier, a pioneer in virtual reality, points out, anger is the emotion most effective at driving “engagement” — which also makes it, in a market for attention, the most profitable one. By creating a self-perpetuating loop of shock and recrimination, social media further polarized what had already seemed, during the Obama years, an impossibly and irredeemably polarized country.
  • The Architects (In order of appearance.) Jaron Lanier, virtual-reality pioneer. Founded first company to sell VR goggles; worked at Atari and Microsoft. Antonio García Martínez, ad-tech entrepreneur. Helped create Facebook’s ad machine. Ellen Pao, former CEO of Reddit. Filed major gender-discrimination lawsuit against VC firm Kleiner Perkins. Can Duruk, programmer and tech writer. Served as project lead at Uber. Kate Losse, Facebook employee No. 51. Served as Mark Zuckerberg’s speechwriter. Tristan Harris, product designer. Wrote internal Google presentation about addictive and unethical design. Rich “Lowtax” Kyanka, entrepreneur who founded influential message board Something Awful. Ethan Zuckerman, MIT media scholar. Invented the pop-up ad. Dan McComas, former product chief at Reddit. Founded community-based platform Imzy. Sandy Parakilas, product manager at Uber. Ran privacy compliance for Facebook apps. Guillaume Chaslot, AI researcher. Helped develop YouTube’s algorithmic recommendation system. Roger McNamee, VC investor. Introduced Mark Zuckerberg to Sheryl Sandberg. Richard Stallman, MIT programmer. Created legendary software GNU and Emacs.
  • ...45 more annotations...
  • How It Went Wrong, in 15 Steps Step 1 Start With Hippie Good Intentions …
  • I think two things are at the root of the present crisis. One was the idealistic view of the internet — the idea that this is the great place to share information and connect with like-minded people. The second part was the people who started these companies were very homogeneous. You had one set of experiences, one set of views, that drove all of the platforms on the internet. So the combination of this belief that the internet was a bright, positive place and the very similar people who all shared that view ended up creating platforms that were designed and oriented around free speech.
  • Step 2 … Then mix in capitalism on steroids. To transform the world, you first need to take it over. The planetary scale and power envisioned by Silicon Valley’s early hippies turned out to be as well suited for making money as they were for saving the world.
  • Step 3 The arrival of Wall Streeters didn’t help … Just as Facebook became the first overnight social-media success, the stock market crashed, sending money-minded investors westward toward the tech industry. Before long, a handful of companies had created a virtual monopoly on digital life.
  • Ethan Zuckerman: Over the last decade, the social-media platforms have been working to make the web almost irrelevant. Facebook would, in many ways, prefer that we didn’t have the internet. They’d prefer that we had Facebook.
  • Step 4 … And we paid a high price for keeping it free. To avoid charging for the internet — while becoming fabulously rich at the same time — Silicon Valley turned to digital advertising. But to sell ads that target individual users, you need to grow a big audience — and use advancing technology to gather reams of personal data that will enable you to reach them efficiently.
  • Harris: If you’re YouTube, you want people to register as many accounts as possible, uploading as many videos as possible, driving as many views to those videos as possible, so you can generate lots of activity that you can sell to advertisers. So whether or not the users are real human beings or Russian bots, whether or not the videos are real or conspiracy theories or disturbing content aimed at kids, you don’t really care. You’re just trying to drive engagement to the stuff and maximize all that activity. So everything stems from this engagement-based business model that incentivizes the most mindless things that harm the fabric of society.
  • Step 5 Everything was designed to be really, really addictive. The social-media giants became “attention merchants,” bent on hooking users no mater the consequences. “Engagement” was the euphemism for the metric, but in practice it evolved into an unprecedented machine for behavior modification.
  • Harris: That blue Facebook icon on your home screen is really good at creating unconscious habits that people have a hard time extinguishing. People don’t see the way that their minds are being manipulated by addiction. Facebook has become the largest civilization-scale mind-control machine that the world has ever seen.
  • Step 6 At first, it worked — almost too well. None of the companies hid their plans or lied about how their money was made. But as users became deeply enmeshed in the increasingly addictive web of surveillance, the leading digital platforms became wildly popular.
  • Pao: There’s this idea that, “Yes, they can use this information to manipulate other people, but I’m not gonna fall for that, so I’m protected from being manipulated.” Slowly, over time, you become addicted to the interactions, so it’s hard to opt out. And they just keep taking more and more of your time and pushing more and more fake news. It becomes easy just to go about your life and assume that things are being taken care of.
  • McNamee: If you go back to the early days of propaganda theory, Edward Bernays had a hypothesis that to implant an idea and make it universally acceptable, you needed to have the same message appearing in every medium all the time for a really long period of time. The notion was it could only be done by a government. Then Facebook came along, and it had this ability to personalize for every single user. Instead of being a broadcast model, it was now 2.2 billion individualized channels. It was the most effective product ever created to revolve around human emotions.
  • Step 7 No one from Silicon Valley was held accountable … No one in the government — or, for that matter, in the tech industry’s user base — seemed interested in bringing such a wealthy, dynamic sector to heel.
  • Step 8 … Even as social networks became dangerous and toxic. With companies scaling at unprecedented rates, user security took a backseat to growth and engagement. Resources went to selling ads, not protecting users from abuse.
  • Lanier: Every time there’s some movement like Black Lives Matter or #MeToo, you have this initial period where people feel like they’re on this magic-carpet ride. Social media is letting them reach people and organize faster than ever before. They’re thinking, Wow, Facebook and Twitter are these wonderful tools of democracy. But it turns out that the same data that creates a positive, constructive process like the Arab Spring can be used to irritate other groups. So every time you have a Black Lives Matter, social media responds by empowering neo-Nazis and racists in a way that hasn’t been seen in generations. The original good intention winds up empowering its opposite.
  • Chaslot: As an engineer at Google, I would see something weird and propose a solution to management. But just noticing the problem was hurting the business model. So they would say, “Okay, but is it really a problem?” They trust the structure. For instance, I saw this conspiracy theory that was spreading. It’s really large — I think the algorithm may have gone crazy. But I was told, “Don’t worry — we have the best people working on it. It should be fine.” Then they conclude that people are just stupid. They don’t want to believe that the problem might be due to the algorithm.
  • Parakilas: One time a developer who had access to Facebook’s data was accused of creating profiles of people without their consent, including children. But when we heard about it, we had no way of proving whether it had actually happened, because we had no visibility into the data once it left Facebook’s servers. So Facebook had policies against things like this, but it gave us no ability to see what developers were actually doing.
  • McComas: Ultimately the problem Reddit has is the same as Twitter: By focusing on growth and growth only, and ignoring the problems, they amassed a large set of cultural norms on their platforms that stem from harassment or abuse or bad behavior. They have worked themselves into a position where they’re completely defensive and they can just never catch up on the problem. I don’t see any way it’s going to improve. The best they can do is figure out how to hide the bad behavior from the average user.
  • Step 9 … And even as they invaded our privacy. The more features Facebook and other platforms added, the more data users willingly, if unwittingly, released to them and the data brokers who power digital advertising.
  • Richard Stallman: What is data privacy? That means that if a company collects data about you, it should somehow protect that data. But I don’t think that’s the issue. The problem is that these companies are collecting data about you, period. We shouldn’t let them do that. The data that is collected will be abused. That’s not an absolute certainty, but it’s a practical extreme likelihood, which is enough to make collection a problem.
  • Losse: I’m not surprised at what’s going on now with Cambridge Analytica and the scandal over the election. For long time, the accepted idea at Facebook was: Giving developers as much data as possible to make these products is good. But to think that, you also have to not think about the data implications for users. That’s just not your priority.
  • Step 10 Then came 2016. The election of Donald Trump and the triumph of Brexit, two campaigns powered in large part by social media, demonstrated to tech insiders that connecting the world — at least via an advertising-surveillance scheme — doesn’t necessarily lead to that hippie utopia.
  • Chaslot: I realized personally that things were going wrong in 2011, when I was working at Google. I was working on this YouTube recommendation algorithm, and I realized that the algorithm was always giving you the same type of content. For instance, if I give you a video of a cat and you watch it, the algorithm thinks, Oh, he must really like cats. That creates these feeder bubbles where people just see one type of information. But when I notified my managers at Google and proposed a solution that would give a user more control so he could get out of the feeder bubble, they realized that this type of algorithm would not be very beneficial for watch time. They didn’t want to push that, because the entire business model is based on watch time.
  • Step 11 Employees are starting to revolt. Tech-industry executives aren’t likely to bite the hand that feeds them. But maybe their employees — the ones who signed up for the mission as much as the money — can rise up and make a change.
  • Harris: There’s a massive demoralizing wave that is hitting Silicon Valley. It’s getting very hard for companies to attract and retain the best engineers and talent when they realize that the automated system they’ve built is causing havoc everywhere around the world. So if Facebook loses a big chunk of its workforce because people don’t want to be part of that perverse system anymore, that is a very powerful and very immediate lever to force them to change.
  • Duruk: I was at Uber when all the madness was happening there, and it did affect recruiting and hiring. I don’t think these companies are going to go down because they can’t attract the right talent. But there’s going to be a measurable impact. It has become less of a moral positive now — you go to Facebook to write some code and then you go home. They’re becoming just another company.
  • Step 12 To fix it, we’ll need a new business model … If the problem is in the way the Valley makes money, it’s going to have to make money a different way. Maybe by trying something radical and new — like charging users for goods and services.
  • Parakilas: They’re going to have to change their business model quite dramatically. They say they want to make time well spent the focus of their product, but they have no incentive to do that, nor have they created a metric by which they would measure that. But if Facebook charged a subscription instead of relying on advertising, then people would use it less and Facebook would still make money. It would be equally profitable and more beneficial to society. In fact, if you charged users a few dollars a month, you would equal the revenue Facebook gets from advertising. It’s not inconceivable that a large percentage of their user base would be willing to pay a few dollars a month.
  • Step 13 … And some tough regulation. Mark Zuckerberg testifying before Congress on April 10. Photo: Jim Watson/AFP/Getty Images While we’re at it, where has the government been in all this? 
  • Stallman: We need a law. Fuck them — there’s no reason we should let them exist if the price is knowing everything about us. Let them disappear. They’re not important — our human rights are important. No company is so important that its existence justifies setting up a police state. And a police state is what we’re heading toward.
  • Duruk: The biggest existential problem for them would be regulation. Because it’s clear that nothing else will stop these companies from using their size and their technology to just keep growing. Without regulation, we’ll basically just be complaining constantly, and not much will change.
  • McNamee: Three things. First, there needs to be a law against bots and trolls impersonating other people. I’m not saying no bots. I’m just saying bots have to be really clearly marked. Second, there have to be strict age limits to protect children. And third, there has to be genuine liability for platforms when their algorithms fail. If Google can’t block the obviously phony story that the kids in Parkland were actors, they need to be held accountable.
  • Stallman: We need a law that requires every system to be designed in a way that achieves its basic goal with the least possible collection of data. Let’s say you want to ride in a car and pay for the ride. That doesn’t fundamentally require knowing who you are. So services which do that must be required by law to give you the option of paying cash, or using some other anonymous-payment system, without being identified. They should also have ways you can call for a ride without identifying yourself, without having to use a cell phone. Companies that won’t go along with this — well, they’re welcome to go out of business. Good riddance.
  • Step 14 Maybe nothing will change. The scariest possibility is that nothing can be done — that the behemoths of the new internet are too rich, too powerful, and too addictive for anyone to fix.
  • García: Look, I mean, advertising sucks, sure. But as the ad tech guys say, “We’re the people who pay for the internet.” It’s hard to imagine a different business model other than advertising for any consumer internet app that depends on network effects.
  • Step 15 … Unless, at the very least, some new people are in charge. If Silicon Valley’s problems are a result of bad decision-making, it might be time to look for better decision-makers. One place to start would be outside the homogeneous group currently in power.
  • Pao: I’ve urged Facebook to bring in people who are not part of a homogeneous majority to their executive team, to every product team, to every strategy discussion. The people who are there now clearly don’t understand the impact of their platforms and the nature of the problem. You need people who are living the problem to clarify the extent of it and help solve it.
  • Things That Ruined the Internet
  • Cookies (1994) The original surveillance tool of the internet. Developed by programmer Lou Montulli to eliminate the need for repeated log-ins, cookies also enabled third parties like Google to track users across the web. The risk of abuse was low, Montulli thought, because only a “large, publicly visible company” would have the capacity to make use of such data. The result: digital ads that follow you wherever you go online.
  • The Farmville vulnerability (2007)   When Facebook opened up its social network to third-party developers, enabling them to build apps that users could share with their friends, it inadvertently opened the door a bit too wide. By tapping into user accounts, developers could download a wealth of personal data — which is exactly what a political-consulting firm called Cambridge Analytica did to 87 million Americans.
  • Algorithmic sorting (2006) It’s how the internet serves up what it thinks you want — automated calculations based on dozens of hidden metrics. Facebook’s News Feed uses it every time you hit refresh, and so does YouTube. It’s highly addictive — and it keeps users walled off in their own personalized loops. “When social media is designed primarily for engagement,” tweets Guillaume Chaslot, the engineer who designed YouTube’s algorithm, “it is not surprising that it hurts democracy and free speech.”
  • The “like” button (2009) Initially known as the “awesome” button, the icon was designed to unleash a wave of positivity online. But its addictive properties became so troubling that one of its creators, Leah Pearlman, has since renounced it. “Do you know that episode of Black Mirror where everyone is obsessed with likes?” she told Vice last year. “I suddenly felt terrified of becoming those people — as well as thinking I’d created that environment for everyone else.”
  • Pull-to-refresh (2009) Developed by software developer Loren Brichter for an iPhone app, the simple gesture — scrolling downward at the top of a feed to fetch more data — has become an endless, involuntary tic. “Pull-to-refresh is addictive,” Brichter told The Guardian last year. “I regret the downsides.”
  • Pop-up ads (1996) While working at an early blogging platform, Ethan Zuckerman came up with the now-ubiquitous tool for separating ads from content that advertisers might find objectionable. “I really did not mean to break the internet,” he told the podcast Reply All. “I really did not mean to bring this horrible thing into people’s lives. I really am extremely sorry about this.”
  • The Silicon Valley dream was born of the counterculture. A generation of computer programmers and designers flocked to the Bay Area’s tech scene in the 1970s and ’80s, embracing new technology as a tool to transform the world for good.
  •  
    Internet en 15 étapes, de sa construction à aujourd'hui, regards et regrets de ceux qui l'ont construit... [...] "Things That Ruined the Internet" les cookies 1994 / la faille Farmville 2007 / le tri algorithmique 2006 / le "like" 2009 / le "pull to refresh" 2009 / les "pop-up ads" 1996 [...]
Aurialie Jublin

Draft for a Mutualism and Self-Management Manifesto | Commonfare.net - 0 views

  •  
    "In view of the European meeting of recovered factories and mutual societies, to be held in RiMaflow between next 12 and 14 April, we propose, in Italian and in English, a text on conflicting mutualism whose drafting was shared among all the realities that wanted to participate. The ambition is that this manifesto will concretely become a plural but common mobilization tool with which to face the challenges of the new phase. (The conference working group). "
Aurialie Jublin

Open source sustainability | TechCrunch - 0 views

  • Engineers around the world pour their sweat and frankly, their hearts into these passion projects that undergird all software in the modern internet economy. In exchange, they ask for nothing in return except for recognition and help in keeping their projects alive and improving them. It’s an incredible movement of decentralized voluntarism and represents humanity at its best.
  • Two years later, a new crop of entrepreneurs, open source maintainers, and organizations have taken Eghbal up on that challenge, developing solutions that maintain the volunteer spirit at the heart of open source while inventing new economic models to make the work sustainable. All are early, and their long-term effects on the output and quality of open source are unknown. But each solution offers an avenue that could radically change the way we think of a career in open source in the future.
  • Eghbal’s report two years ago summarized the vast issues facing open source maintainers, challenges that have remained essentially unchanged in the interim. It’s a quintessential example of the “tragedy of the commons.” As Eghbal wrote at the time, “Fundamentally, digital infrastructure has a free rider problem. Resources are offered for free, and everybody (whether individual developer or large software company) uses them, so nobody is incentivized to contribute back, figuring that somebody else will step in.” That has led to a brittle ecosystem, just as open source software reached the zenith of its influence.
  • ...22 more annotations...
  • Eghbal pointed to OpenSSL as an example, a library that powers a majority of encrypted communications on the web. Following the release of the Heartbleed security bug, people were surprised to learn that the OpenSSL project was the work of a very small team of individuals, with only one of them working on it full-time (and at a very limited salary compared to industry norms).
  • No one wants open source to disappear, or for maintainers to burnout. Yet, there is a strong cultural force against commercial interests in the community. Money is corrupting, and dampens the voluntary spirit of open source efforts. More pragmatically, there are vast logistical challenges with managing money on globally distributed volunteer teams that can make paying for work logistically challenging.
  • Kyle Mitchell, a lawyer by trade and founder of License Zero, says that there is an assumption that “Open source will continue to fall from the sky like manna from heaven and that the people behind it can be abstracted away.” He concludes: “It is just really wrong.”That view was echoed by Henry Zhu, who is the maintainer of the popular JavaScript compiler Babel. “We trust startups with millions of VC money and encourage a culture of ‘failing fast,’ yet somehow the idea of giving to volunteers who may have showed years of dedication is undesirable?” he said.
  • Mitchell believes that one of the most important challenges is just getting comfortable talking about money. “Money feels dirty until it doesn’t,” he said. “I would like to see more money responsibility in the community.” One challenge he notes is that “learning to be a great maintainer doesn’t teach you how to be a great open source contractor or consultant.” GitHub works great as a code repository service, but ultimately doesn’t teach maintainers the economics of their work.
  • Perhaps the greatest debate in sustaining open source is deciding who or what to target: the individual contributors — who often move between multiple projects — or a particular library itself.
  • Patreon is a crowdsourced subscription platform, perhaps best known for the creatives it hosts. These days though, it is also increasingly being used by notable open source contributors as a way to connect with fans and sustain their work. Aboukhadijeh launched his page after seeing others doing it. “A bunch of people were starting up Patreons, which was kind of a meme in my JavaScript circles,” he said. His Patreon page today has 72 contributors providing him with $2,874 in funding per month ($34,488 annually).
  • That may seem a bit paltry, but he explained to me that he also supplements his Patreon with funding from organizations as diverse as Brave (an adblocking browser with a utility token model) to PopChest (a decentralized video sharing platform). That nets him a couple of more thousands of dollars per month.
  • Aboukhadijeh said that Twitter played an outsized role in building out his revenue stream. “Twitter is the most important on where the developers talk about stuff and where conversations happen…,” he said. “The people who have been successful on Patreon in the same cohort [as me] who tweet a lot did really well.”
  • For those who hit it big, the revenues can be outsized. Evan You, who created the popular JavaScript frontend library Vue.js, has reached $15,206 in monthly earnings ($182,472 a year) from 231 patrons. The number of patrons has grown consistently since starting his Patreon in March 2016 according to Graphtreon, although earnings have gone up and down over time.
  • While Patreon is one direct approach for generating revenues from users, another one is to offer dual licenses, one free and one commercial. That’s the model of License Zero, which Kyle Mitchell propsosed last year. He explained to me that “License Zero is the answer to a really simple question with no simple answers: how do we make open source business models open to individuals?”
  • License Zero is a permissive license based on the two-clause BSD license, but adds terms requiring commercial users to pay for a commercial license after 90 days, allowing companies to try a project before purchasing it. If other licenses aren’t available for purchase (say, because a maintainer is no longer involved), then the language is no longer enforceable and the software is offered as fully open source. The idea is that other open source users can always use the software for free, but for-profit uses would require a payment.
  • Mitchell believes that this is the right approach for individuals looking to sustain their efforts in open source. “The most important thing is the time budget – a lot of open source companies or people who have an open source project get their money from services,” he said. The problem is that services are exclusive to a company, and take time away from making a project as good as it can be. “When moneymaking time is not time spent on open source, then it competes with open source,” he said.
  • Supporting individuals makes a lot of sense, but often companies want to support the specific projects and ecosystems that underpin their software. Doing so can be next to impossible. There are complicated logistics required in order for companies to fund open source, such as actually having an organization to send money to (and for many, to convince the IRS that the organization is actually a non-profit). Tidelift and Open Collective are two different ways to open up those channels.
  • Fischer and his team wanted to create a platform that would allow open source ecosystems to sustain themselves. “We felt frustrated at some level that while open source has taken over a huge portion of software, a lot of the creators of open source have not been able to capture a lot of the value they are creating,” he explained.Tidelift is designed to offer assurances “around areas like security, licensing, and maintenance of software,” Fischer explained. The idea has its genesis in Red Hat, which commercialized Linux. The idea is that companies are willing to pay for open source when they can receive guarantees around issues like critical vulnerabilities and long-term support. In addition, Tidelift handles the mundane tasks of setting up open source for commercialization such as handling licensing issues.
  • Tidelift initially launched in the JavaScript ecosystem around React, Angular, and Vue.js, but will expand to more communities over time. The company has raised $15 million in venture capital from General Catalyst and Foundry Group, plus former Red Hat chairman and CEO Matthew Szulik.
  • Fischer hopes that the company can change the economics for open source contributors. He wants the community to move from a model of “get by and survive” with a “subsistence level of earnings” and instead, help maintainers of great software “win big and be financially rewarded for that in a significant way.”
  • Where Tidelift is focused on commercialization and software guarantees, Open Collective wants to open source the monetization of open source itself.Open Collective is a platform that provides tools to “collectives” to receive money while also offering mechanisms to allow the members of those collectives to spend their money in a democratic and transparent way.
  • Xavier Damman, founder president of Open Collective, believes that this radical transparency could reshape how the economics of open source are considered by its participants. Damman likens Open Collective to the “View Source” feature of a web browser that allows users to read a website’s code. “Our goal as a platform is to be as transparent as possible,” he said.
  • Open Collective was launched in late 2015, and since then has become home to 647 open source projects. So far, Webpack, the popular JavaScript build tool, has generated the most revenue, currently sitting at $317,188 a year. One major objective of the organization is to encourage more for-profit companies to commit dollars to open source. Open Collective places the logos of major donors on each collective page, giving them visible credit for their commitment to open source.
  • Damman’s ultimate dream is to change the notion of ownership itself. We can move from “Competition to collaboration, but also ownership to commons,” he envisioned.
  • It’s unfortunately very early days for open source sustainability. While Patreon, License Zero, Tidelift, and Open Collective are different approaches to providing the infrastructure for sustainability, ultimately someone has to pay to make all that infrastructure useful. There are only a handful of Patreons that could substitute for an engineer’s day job, and only two collectives by my count on Open Collective that could support even a single maintainer full time. License Zero and Tidelift are too new to know how they will perform yet.
  • The change though can happen through all of us who work on code — every software engineer and product manager. If you work at a for-profit company, take the lead in finding a way to support the code that allows you to do your job so efficiently. The decentralization and volunteer spirit of the open source community needs exactly the same kind of decentralized spirit in every financial contributor. Sustainability is each of our jobs, every day. If we all do our part, we can help to sustain one of the great intellectual movements humanity has ever created, and end the oxymoron of open source sustainability forever.
  •  
    "Open source maintainers are exhausted and rarely paid. A new generation wants to change the economics."
Aurialie Jublin

De la tyrannie des métriques | InternetActu.net - 0 views

  • Pour Muller, nous sommes obsédés par les chiffres. Nous avons une pression irrépressible à mesurer la performance, à en publier les chiffres, et à récompenser les performances depuis ceux-ci, quand bien même l’évidence nous montre que cela ne fonctionne pas si bien. Pour Muller, notre problème n’est pas tant la mesure que ses excès. Trop souvent nous préférons substituer des chiffres, des mesures, à un jugement personnel. Trop souvent le jugement est compris comme personnel, subjectif, orienté par celui qui le produit, alors que les chiffres, en retour, eux, sont supposés fournir une information sûre et objective.
  • Trop souvent, les métriques sont contre-productives, notamment quand elles peinent à mesurer ce qui ne l’est pas, à quantifier ce qui ne peut l’être.
  • Muller montre par de nombreux exemples comment nous nous ingénions à contourner les mesures, à l’image des hôpitaux britanniques qui avaient décidé de pénaliser les services dont les temps d’attente aux urgences étaient supérieurs à 4 heures. La plupart des hôpitaux avaient résolu le problème en faisant attendre les ambulances et leurs patients en dehors de l’hôpital !
  • ...10 more annotations...
  • Pour Muller, l’obsession des métriques repose sur : la croyance qu’il est possible et désirable de remplacer le jugement acquis par l’expérience personnelle et le talent, avec des indicateurs numériques de performance comparative basés sur des données standardisées ; la croyance que rendre ces mesures publiques (c’est-à-dire transparentes) assure que les institutions effectuent leurs buts (c’est-à-dire leur responsabilité) ; la croyance que la meilleure façon de motiver les gens dans les organisations est de les attacher à des récompenses ou des pénalités depuis ces mesures de performance (que les récompenses soient monétaires ou réputationnelles)
  • Il souligne que trop souvent, on mesure le plus ce qui est le plus facile à mesurer, le plus simple. Mais c’est rarement le plus important. En mesurant le plus simple, on en oublie la complexité : ainsi quand on mesure les objectifs d’un employé, on oublie souvent que son travail est plus complexe que cela. On mesure plus facilement les sommes dépensées ou les ressources qu’on injecte dans un projet que les résultats des efforts accomplis. Les organisations mesurent plus ce qu’elles dépensent que ce qu’elles produisent.
  • Bien souvent, la mesure, sous prétexte de standardisation, dégrade l’information, notamment pour rendre les choses comparables au détriment des ambiguïtés et de l’incertitude.
  • Quant aux manières de se jouer des métriques, elles sont là aussi nombreuses : simplifier les objectifs permet souvent de les atteindre au détriment des cas difficiles ; améliorer les chiffres se fait souvent en abaissant les standards… sans parler de la triche, on ne peut plus courante.
  • C’est vraiment avec Frederick Taylor, cet ingénieur américain qui va inventer le management scientifique au début du 20e siècle, que la mesure va s’imposer. Le but même du taylorisme était de remplacer le savoir implicite des ouvriers avec des méthodes de production de masse, développées, planifiées, surveillées et contrôlées par les managers. Le Taylorisme va s’imposer en promouvant l’efficacité par la standardisation et la mesure, d’abord dans l’industrie avant de coloniser avec le siècle, tous les autres secteurs productifs.
  • Les métriques se sont imposées dans les secteurs où la confiance était faible. Les décisions humaines sont devenues trop dangereuses à mesure qu’elles impliquaient une trop grande complexité de facteurs : d’où la prolifération des métriques, des process, des règles… La demande d’un flux constant de rapports, de données, de chiffres… a finalement diminué l’autonomie de ceux qui étaient les plus bas dans les organisations.
  • Dans son livre, Muller évoque nombre de secteurs où les chiffres n’ont pas produit ce qu’on attendait d’eux. Dans le domaine scolaire, en médecine, dans le domaine militaire, dans la police et bien sûr dans les affaires… les chiffres sont souvent mal utilisés et ne parviennent pas à mesurer l’important. Bien souvent, les chiffres déplacent les objectifs, favorisent le court-termisme, découragent la prise de risque, l’innovation, la coopération… coûtent du temps
  • voire sont dommageable pour la productivité elle-même (« Une question qui devrait être posée est de savoir dans quelle mesure la culture des métriques a elle-même contribué à la stagnation économique ? »)…
  • Pour lui, il faut d’abord se poser la question de ce que l’on cherche à mesurer en se souvenant que plus un objet mesuré est influencé par la procédure de mesure, moins il est fiable. Et que ce constat empire quand la mesure repose sur l’activité humaine, plus capable de réagir au fait d’être mesurée, et pire encore quand des systèmes de récompenses ou de punition sont introduits…
  • L’information est-elle utile ? Le fait que quelque chose soit mesurable ne signifie pas qu’il faille le faire (au contraire, bien souvent la facilité à mesurer est inversement proportionnelle à sa signification !). « Posez-vous la question de ce que vous voulez vraiment savoir ! », conseille l’historien.
  •  
    "On n'a pas attendu le numérique pour produire des métriques, mais celui-ci a incontestablement simplifié et amplifié leur production. Quitte à produire un peu n'importe quelles mesures puisque celles-ci semblent désormais accessibles et ajustables en quelques clics. On connaît la rengaine : « ce qui ne peut être mesuré ne peut être géré » (même si la formule de Peter Drucker est en fait plus pertinente que ce que nous en avons retenu). Est-ce pour autant qu'on ne gère que ce que l'on sait chiffrer ?"
Aurialie Jublin

Charter for Building a Data Commons for a Free, Fair, and Sustainable Future - HackMD - 0 views

  • This Charter/Carta provides practical guidance and political orientation for mapping, modelling, managing and sharing data as a Commons. If you follow these guidelines, you will contribute to a Global Data Commons. That is, you will govern your mapping community and manage data differently than people who centralize data control for profit.The Charter does not describe the vision, scope or values of a specific mapping project, but Data Commons principles. It will help you reimagine how you protect the animating spirit of your mapping project and prevents your data from being co-opted or enclosed.The Charter as a whole is the maximum „commons denominator“ of mapping projects that aspire to share data for the common good.Help commonize maps and data! For the people, by the people.
  •  
    "Charter for Building a Data Commons for a Free, Fair, and Sustainable Future"
Aurialie Jublin

The Landlord Wants Facial Recognition in Its Rent-Stabilized Buildings. Why? - The New ... - 0 views

  • The fact that the Atlantic complex already has 24-hour security in its lobbies as well as a clearly functioning camera system has only caused tenants to further question the necessity of facial recognition technology. The initiative is particularly dubious given the population of the buildings. Last year, a study out of M.I.T. and Stanford looked at the accuracy rates of some of the major facial-analysis programs on the market. It found that although the error rates for determining the gender of light-skinned men never surpassed 1 percent, the same programs failed to identify darker-skinned women up to one-third of the time.
  • The fear that marginalized groups will fall under increased surveillance as these technologies progress in the absence of laws to regulate them hardly seems like dystopian hysteria.
  • In November, the City of Detroit announced that it was introducing the use of real-time police cameras at two public-housing towers. The existing program is known as Project Greenlight, and it was designed to deter criminal behavior. But tower residents worried that relatives would be less likely to visit, given the constant stream of data collected by law enforcement.
  •  
    "Last fall, tenants at the Atlantic Plaza Towers, a rent-stabilized apartment complex in Brooklyn, received an alarming letter in the mail. Their landlord was planning to do away with the key-fob system that allowed them entry into their buildings on the theory that lost fobs could wind up in the wrong hands and were now also relatively easy to duplicate. Instead, property managers planned to install facial recognition technology as a means of access. It would feature "an encrypted reference file" that is "only usable in conjunction with the proprietary algorithm software of the system," the letter explained, in a predictably failed effort to mitigate concerns about privacy. As it happened, not every tenant was aware of these particular Orwellian developments. New mailboxes in the buildings required new keys, and to obtain a new key you had to submit to being photographed; some residents had refused to do this and so were not getting their mail."
Aurialie Jublin

Uberland : l'avenir du travail à l'heure des algorithmes | InternetActu.net - 0 views

  • Pour Alex Rosenblat, Uber a provoqué « un changement culturel fondamental sur ce que signifie avoir un emploi ». Uber a tiré un maximum de valeur des conducteurs en surfant sur les limites et ambiguïtés des normes culturelles et juridiques. Pour Adrian Chen, la gestion algorithmique d’Uber est moins une révolution qu’un affinage, à l’image des chauffeurs de taxi eux-mêmes, qui ont longtemps été dans une situation entre autonomie et dépendance, notamment avec nombre d’intermédiaires influant sur leurs possibilités de travail. Uber a simplement rendu cette sous-traitance plus efficace.
  • Adrian Chen pour Intelligencer explique que cette gestion algorithmique de la force de travail suppose une surveillance totale (jusqu’aux tremblements du téléphone pour s’assurer que le conducteur conduit en toute sécurité ! ce qui montre à nouveau, s’il en était besoin, que Uber n’est pas seulement un intermédiaire pour réaliser une transaction). Pourtant, de nombreux chauffeurs semblent satisfaits du travail qu’ils réalisent : ils apprécient la flexibilité et le complément de salaire que leur travail apporte. Même si les récits que Rosenblat a recueillis témoignent principalement d’une résignation épuisée. «Pour certains chauffeurs, ce travail est un véritable boulet, et pour d’autres, c’est un gilet de sauvetage».
  • Dans une interview pour CityLab, la chercheuse rappelle que malgré les efforts d’Uber, le taux de désengagement des chauffeurs est assez fort : 68 % d’entre eux quittent la plateforme dans les 6 mois. Elle pointe néanmoins que les outils d’Uber proposent des services que les collectivités ne savaient pas fournir, notamment en offrant des modalités de transports là où elles n’existaient pas (voir « Plateformes et métropoles » et « Vers une ville numérique ingouvernable »). Uber, derrière sa neutralité de façade, a surtout profité d’un Far West réglementaire pour s’imposer. Partout où elle s’est imposée, la firme a profondément déstabilisé la société (autour de la question du statut des chauffeurs), tout en offrant des avantages populaires (ouvrant de nouvelles dessertes et des prix bas notamment). Toute la question reste de trouver la balance entre les avantages et les inconvénients
  • ...10 more annotations...
  • Dans une tribune pour le New York Times, la chercheuse rappelait la difficulté de cette gestion algorithmique de la relation de travail (voir également « Qui des algorithmes ou des clients seront nos nouveaux patrons »). Une relation profondément asymétrique et qui se crispe notamment dès qu’il y a un problème. Elle souligne combien les informations que délivre Uber à ses chauffeurs comme à leurs passagers sont asymétriques. Un chauffeur de New York évoquait par exemple un échange avec un client qui montrait que celui-ci était facturé 40$ alors que pour le chauffeur, la course n’en coûtait que 28 !
  • Dans les années 70, le sociologue du travail britannique Michael Burawoy, dans son livre Produire le consentement (1979, traduit en français en 2015) constatait que si les travailleurs consentaient à leur propre exploitation, c’était en fait surtout par jeu. Burawoy expliquait que lorsque le travail prenait la forme d’un jeu, la principale source de tension n’était plus dans la relation des travailleurs avec les dirigeants, mais s’étendait aux autres travailleurs, aux machines et contre eux-mêmes. En proposant un sentiment d’autonomie et de contrôle relatifs, la ludification propose aux travailleurs de souscrire aux objectifs de l’entreprise. Lyft propose ainsi des défis à ses conducteurs en échange d’une prime supplémentaire : des défis assez difficiles à décrocher qui reposent sur le nombre de courses à réaliser à certains créneaux horaires.
  • 50 % des conducteurs de Uber cessent d’utiliser l’application au bout de deux mois. 4 % des conducteurs le sont encore après une année. Des rapports récents estiment que le revenu des chauffeurs à l’heure est souvent inférieur au salaire minimum.
  • Sur les forums de chauffeurs ou dans leurs discussions, ceux-ci spéculent sans cesse sur les algorithmes et les moyens de les déjouer, mais sans réellement ne jamais y parvenir. Ils tentent fréquemment de faire des déconnexions en masse de l’application pour faire monter les prix des courses, sans que l’effet n’ait jamais vraiment été assuré. Ces formes de déconnexion volontaires et organisées ressemblent à des microgrèves, à des microprotestations, sans jamais avoir permis de réelles émancipations.
  • Burawoy est à nouveau utile pour comprendre pourquoi. Le fait de « jouer » permet aux travailleurs d’exercer un contrôle limité sur le processus de travail. Mais les victoires qu’ils remportent renforcent leur engagement et leur consentement aux règles du jeu. À l’inverse, leurs échecs dirigent leur insatisfaction contre les obstacles du jeu, et non contre ceux qui en fixent les règles. Apprendre à utiliser intelligemment les paramètres du jeu, de l’algorithme, devient pour eux la seule option imaginable.
  • Après des semaines d’une conduite maniaque pour améliorer sa note… elle a constaté que cette amélioration était sans incidence directe (tant qu’elle restait au-dessus de 4,6/5) : cela ne générait ni réel bonus ni amélioration du niveau de rémunération. « En fait, je perdais de l’argent en essayant de flatter les clients avec des bonbons et de garder ma voiture scrupuleusement propre », simplement parce qu’elle souhaitait rester un pilote bien côté. « Ce qui est à la fois si brillant et affreux dans la ludification de Lyft et Uber, c’est qu’elle repose sur notre seul désir d’être utile, aimé et bon. Les semaines où je suis très bien noté, je suis plus motivé à conduire. Les semaines où je suis mal noté, je suis également plus motivé à conduire. » Une fois qu’on est pris dans la zone du jeu, il n’y a plus d’échappatoire !
  • Heureusement, la réaction à l’ubérisation s’est aussi peu à peu musclée. Pour Les Inrocks, Benoît Fabien revient sur le lancement de la Fédération transnationale des coursiers (FTC), qui a rassemblé fin octobre à Bruxelles des livreurs à domicile de toute l’Europe, à l’initiative du Groupe de recherche pour une stratégie économique alternative, de l’Alter Summit et du réseau ReAct, qui encourage les mobilisations transnationales face aux grandes entreprises. Un moyen de contrer la division et l’individualisation des plateformes de livraison pour rééquilibrer les relations de travail imposées par celles-ci. Les revendications sont nombreuses, mais s’organisent : salaire minimum, transparence des algorithmes, la liberté de se réunir en collectif…
  • les chauffeurs d’Uber s’informent des nouvelles fonctionnalités de la plateforme via ces outils, notamment quand Uber teste sur certains d’entre eux de nouvelles fonctionnalités. Le géographe Mark Graham a montré que plus de 50 % de travailleurs freelances qui utilisent des plateformes comme le Mechanical Turk d’Amazon ou Freelancer, sont en contact avec d’autres travailleurs de plateformes et échangent sur les tarifs, sur les tâches proposées et la qualité de ceux qui proposent du travail. Ces groupes servent souvent à des formes de coordination, comme l’a montré la grève des enseignants de Virginie-Occidentale, où 70 % des 35 000 enseignants de l’État étaient présents et ont permis d’articuler les revendications et d’organiser les protestations. Les groupes permettent également de fabriquer des hashtags, des gifs, des vidéos et des mêmes pour publiciser leurs actions. Si ces outils ne sont pas conçus pour soutenir le militantisme, le relai est assuré par des sites dédiés comme Coworker.org, permettant par exemple aux employés de Starbucks de faire pression sur la direction pour limiter le « clopening », le fait qu’un même employé soit chargé de la fermeture et de l’ouverture d’un magasin.
  • Reste que ces services ne sont pas toujours suffisamment en lien avec les syndicats existants : ils manquent d’assistance juridique, comme le proposent les syndicats, ou de fonds pour être maintenus. Les « startups du travail » ont assurément besoin de la force des syndicats pour se transformer. En Allemagne, IG Metal, l’un des plus importants syndicats, permet aux autoentrepreneurs de le rejoindre. Il a également lancé un site pour comparer les conditions de travail sur les plateformes baptisé Fair Crowd Work.
  • Pour Fredrik Soderqvist de Unionen, un syndicat suédois, les syndicats doivent apprendre également à mieux utiliser les données afin de permettre de créer des services adaptés aux transformations du monde… À l’exemple de Mystro, une application pour conducteurs, qui leur permet de basculer d’un service à l’autre, d’évaluer les demandes, de rejeter celles qui ne sont pas rentables, afin d’aider les conducteurs à prendre de meilleures décisions pour eux-mêmes. Assurément, s’emparer activement de la technologie pour les syndicats et les travailleurs peut leur permettre de contrer l’ubérisation… Ça ne remplace ni la lutte sociale ni la lutte légale, mais la lutte technologique peut permettre de faire des contre-propositions concrètes, de fournir des outils d’aide à la décision des travailleurs algorithmisés pour les aider à en comprendre le fonctionnement et à resymétriser la relation.
  •  
    "Dans Uberland : comment les algorithmes réécrivent les règles du travail (2018, Presse universitaire de Californie, non traduit), la chercheuse Alex Rosenblat (@mawnikr) a synthétisé quatre années de recherche ethnographique avec les conducteurs d'Uber. Un livre où transparaît une vision dystopique du travail où des millions de conducteurs sont gérés par un système technique qui combine à la fois l'autoritarisme du management scientifique à la Frederick Taylor et le leadership cynique d'un Michael Scott (le personnage incarné par Steve Carell dans la série The Office), rapporte Intelligencer."
Aurialie Jublin

It's Time to Break Up Facebook - The New York Times - 0 views

  • Mark’s influence is staggering, far beyond that of anyone else in the private sector or in government. He controls three core communications platforms — Facebook, Instagram and WhatsApp — that billions of people use every day. Facebook’s board works more like an advisory committee than an overseer, because Mark controls around 60 percent of voting shares. Mark alone can decide how to configure Facebook’s algorithms to determine what people see in their News Feeds, what privacy settings they can use and even which messages get delivered. He sets the rules for how to distinguish violent and incendiary speech from the merely offensive, and he can choose to shut down a competitor by acquiring, blocking or copying it.
  • Mark is a good, kind person. But I’m angry that his focus on growth led him to sacrifice security and civility for clicks. I’m disappointed in myself and the early Facebook team for not thinking more about how the News Feed algorithm could change our culture, influence elections and empower nationalist leaders. And I’m worried that Mark has surrounded himself with a team that reinforces his beliefs instead of challenging them.
  • We are a nation with a tradition of reining in monopolies, no matter how well intentioned the leaders of these companies may be. Mark’s power is unprecedented and un-American.It is time to break up Facebook.
  • ...26 more annotations...
  • We already have the tools we need to check the domination of Facebook. We just seem to have forgotten about them.America was built on the idea that power should not be concentrated in any one person, because we are all fallible. That’s why the founders created a system of checks and balances. They didn’t need to foresee the rise of Facebook to understand the threat that gargantuan companies would pose to democracy. Jefferson and Madison were voracious readers of Adam Smith, who believed that monopolies prevent the competition that spurs innovation and leads to economic growth.
  • The Sherman Antitrust Act of 1890 outlawed monopolies. More legislation followed in the 20th century, creating legal and regulatory structures to promote competition and hold the biggest companies accountable. The Department of Justice broke up monopolies like Standard Oil and AT&T.
  • For many people today, it’s hard to imagine government doing much of anything right, let alone breaking up a company like Facebook. This isn’t by coincidence. Starting in the 1970s, a small but dedicated group of economists, lawyers and policymakers sowed the seeds of our cynicism. Over the next 40 years, they financed a network of think tanks, journals, social clubs, academic centers and media outlets to teach an emerging generation that private interests should take precedence over public ones. Their gospel was simple: “Free” markets are dynamic and productive, while government is bureaucratic and ineffective. By the mid-1980s, they had largely managed to relegate energetic antitrust enforcement to the history books.
  • It was this drive to compete that led Mark to acquire, over the years, dozens of other companies, including Instagram and WhatsApp in 2012 and 2014. There was nothing unethical or suspicious, in my view, in these moves.
  • Over a decade later, Facebook has earned the prize of domination. It is worth half a trillion dollars and commands, by my estimate, more than 80 percent of the world’s social networking revenue. It is a powerful monopoly, eclipsing all of its rivals and erasing competition from the social networking category. This explains why, even during the annus horribilis of 2018, Facebook’s earnings per share increased by an astounding 40 percent compared with the year before. (I liquidated my Facebook shares in 2012, and I don’t invest directly in any social media companies.)
  • Facebook’s dominance is not an accident of history. The company’s strategy was to beat every competitor in plain view, and regulators and the government tacitly — and at times explicitly — approved. In one of the government’s few attempts to rein in the company, the F.T.C. in 2011 issued a consent decree that Facebook not share any private information beyond what users already agreed to. Facebook largely ignored the decree. Last month, the day after the company predicted in an earnings call that it would need to pay up to $5 billion as a penalty for its negligence — a slap on the wrist — Facebook’s shares surged 7 percent, adding $30 billion to its value, six times the size of the fine.
  • As markets become more concentrated, the number of new start-up businesses declines. This holds true in other high-tech areas dominated by single companies, like search (controlled by Google) and e-commerce (taken over by Amazon). Meanwhile, there has been plenty of innovation in areas where there is no monopolistic domination, such as in workplace productivity (Slack, Trello, Asana), urban transportation (Lyft, Uber, Lime, Bird) and cryptocurrency exchanges (Ripple, Coinbase, Circle).
  • Facebook’s business model is built on capturing as much of our attention as possible to encourage people to create and share more information about who they are and who they want to be. We pay for Facebook with our data and our attention, and by either measure it doesn’t come cheap.
  • The most problematic aspect of Facebook’s power is Mark’s unilateral control over speech. There is no precedent for his ability to monitor, organize and even censor the conversations of two billion people.Facebook engineers write algorithms that select which users’ comments or experiences end up displayed in the News Feeds of friends and family. These rules are proprietary and so complex that many Facebook employees themselves don’t understand them.
  • Facebook has responded to many of the criticisms of how it manages speech by hiring thousands of contractors to enforce the rules that Mark and senior executives develop. After a few weeks of training, these contractors decide which videos count as hate speech or free speech, which images are erotic and which are simply artistic, and which live streams are too violent to be broadcast. (The Verge reported that some of these moderators, working through a vendor in Arizona, were paid $28,800 a year, got limited breaks and faced significant mental health risks.)
  • As if Facebook’s opaque algorithms weren’t enough, last year we learned that Facebook executives had permanently deleted their own messages from the platform, erasing them from the inboxes of recipients; the justification was corporate security concerns. When I look at my years of Facebook messages with Mark now, it’s just a long stream of my own light-blue comments, clearly written in response to words he had once sent me. (Facebook now offers this as a feature to all users.)
  • Mark used to insist that Facebook was just a “social utility,” a neutral platform for people to communicate what they wished. Now he recognizes that Facebook is both a platform and a publisher and that it is inevitably making decisions about values. The company’s own lawyers have argued in court that Facebook is a publisher and thus entitled to First Amendment protection.
  • Mark knows that this is too much power and is pursuing a twofold strategy to mitigate it. He is pivoting Facebook’s focus toward encouraging more private, encrypted messaging that Facebook’s employees can’t see, let alone control. Second, he is hoping for friendly oversight from regulators and other industry executives.
  • In an op-ed essay in The Washington Post in March, he wrote, “Lawmakers often tell me we have too much power over speech, and I agree.” And he went even further than before, calling for more government regulation — not just on speech, but also on privacy and interoperability, the ability of consumers to seamlessly leave one network and transfer their profiles, friend connections, photos and other data to another.
  • Facebook isn’t afraid of a few more rules. It’s afraid of an antitrust case and of the kind of accountability that real government oversight would bring.
  • Mark may never have a boss, but he needs to have some check on his power. The American government needs to do two things: break up Facebook’s monopoly and regulate the company to make it more accountable to the American people.First, Facebook should be separated into multiple companies. The F.T.C., in conjunction with the Justice Department, should enforce antitrust laws by undoing the Instagram and WhatsApp acquisitions and banning future acquisitions for several years. The F.T.C. should have blocked these mergers, but it’s not too late to act. There is precedent for correcting bad decisions — as recently as 2009, Whole Foods settled antitrust complaints by selling off the Wild Oats brand and stores that it had bought a few years earlier.
  • Still others worry that the breakup of Facebook or other American tech companies could be a national security problem. Because advancements in artificial intelligence require immense amounts of data and computing power, only large companies like Facebook, Google and Amazon can afford these investments, they say. If American companies become smaller, the Chinese will outpace us.While serious, these concerns do not justify inaction. Even after a breakup, Facebook would be a hugely profitable business with billions to invest in new technologies — and a more competitive market would only encourage those investments. If the Chinese did pull ahead, our government could invest in research and development and pursue tactical trade policy, just as it is doing today to hold China’s 5G technology at bay.
  • The cost of breaking up Facebook would be next to zero for the government, and lots of people stand to gain economically. A ban on short-term acquisitions would ensure that competitors, and the investors who take a bet on them, would have the space to flourish. Digital advertisers would suddenly have multiple companies vying for their dollars.
  • But the biggest winners would be the American people. Imagine a competitive market in which they could choose among one network that offered higher privacy standards, another that cost a fee to join but had little advertising and another that would allow users to customize and tweak their feeds as they saw fit. No one knows exactly what Facebook’s competitors would offer to differentiate themselves. That’s exactly the point.
  • Just breaking up Facebook is not enough. We need a new agency, empowered by Congress to regulate tech companies. Its first mandate should be to protect privacy.The Europeans have made headway on privacy with the General Data Protection Regulation, a law that guarantees users a minimal level of protection. A landmark privacy bill in the United States should specify exactly what control Americans have over their digital information, require clearer disclosure to users and provide enough flexibility to the agency to exercise effective oversight over time. The agency should also be charged with guaranteeing basic interoperability across platforms.
  • Finally, the agency should create guidelines for acceptable speech on social media. This idea may seem un-American — we would never stand for a government agency censoring speech. But we already have limits on yelling “fire” in a crowded theater, child pornography, speech intended to provoke violence and false statements to manipulate stock prices. We will have to create similar standards that tech companies can use. These standards should of course be subject to the review of the courts, just as any other limits on speech are. But there is no constitutional right to harass others or live-stream violence.
  • These are difficult challenges. I worry that government regulators will not be able to keep up with the pace of digital innovation. I worry that more competition in social networking might lead to a conservative Facebook and a liberal one, or that newer social networks might be less secure if government regulation is weak. But sticking with the status quo would be worse: If we don’t have public servants shaping these policies, corporations will.
  • Similarly, the Justice Department’s 1970s suit accusing IBM of illegally maintaining its monopoly on personal computer sales ended in a stalemate. But along the way, IBM changed many of its behaviors. It stopped bundling its hardware and software, chose an extremely open design for the operating system in its personal computers and did not exercise undue control over its suppliers. Professor Wu has written that this “policeman at the elbow” led IBM to steer clear “of anything close to anticompetitive conduct, for fear of adding to the case against it.”
  • Finally, an aggressive case against Facebook would persuade other behemoths like Google and Amazon to think twice about stifling competition in their own sectors, out of fear that they could be next. If the government were to use this moment to resurrect an effective competition standard that takes a broader view of the full cost of “free” products, it could affect a whole host of industries.
  • I take responsibility for not sounding the alarm earlier. Don Graham, a former Facebook board member, has accused those who criticize the company now as having “all the courage of the last man leaping on the pile at a football game.” The financial rewards I reaped from working at Facebook radically changed the trajectory of my life, and even after I cashed out, I watched in awe as the company grew. It took the 2016 election fallout and Cambridge Analytica to awaken me to the dangers of Facebook’s monopoly. But anyone suggesting that Facebook is akin to a pinned football player misrepresents its resilience and power.
  • This movement of public servants, scholars and activists deserves our support. Mark Zuckerberg cannot fix Facebook, but our government can.
  •  
    "Since then, Mark's personal reputation and the reputation of Facebook have taken a nose-dive. The company's mistakes - the sloppy privacy practices that dropped tens of millions of users' data into a political consulting firm's lap; the slow response to Russian agents, violent rhetoric and fake news; and the unbounded drive to capture ever more of our time and attention - dominate the headlines. It's been 15 years since I co-founded Facebook at Harvard, and I haven't worked at the company in a decade. But I feel a sense of anger and responsibility."
Aurialie Jublin

« Algorithmes, la bombe à retardement » : un cri d'alarme citoyen - 0 views

  • Pour comprendre la thèse du livre, il faut en lire l’envoi : « À tous ceux que l’on donne perdants ». « On », ce sont les dominants qui, sous couvert de discours méritocratique et exaltant l’égalité des chances, déploient des algorithmes qui traquent les dominés dans un grand nombre de domaines et leur infligent de sévères dommages. Comme à la guerre.
  • Pour nommer ces algorithmes malfaisants, elle parle d’« ADM » pour « Arme de destruction mathématique ». Ces armes sont conçues sans que les soldats ne s’en rendent compte, comme elle l’écrit dans des travaux de recherche en géométrie algébrique et mathématique au MIT.
  • Ainsi du monde du travail, où elle réalise que les recrutements se font de plus en plus sans passer par un jugement humain. Dans un grand nombre de postes où la réputation est en jeu, il y a ceux qui peuvent s’acheter un service d’achats de followers Twitter à 19,95 $ par mois et ceux qui restent bloqués à un nombre confidentiel de fans et qui n’auront jamais de community à manager. À l’autre bout de l’échelle sociale, un nombre croissant de demandeurs d’emplois dans des multinationales comme Home Depot, Lowe’s ou Walgreen voient leurs candidatures bloquées par des algos, pour cause d’antécédents psychiatriques. 
  • ...2 more annotations...
  • L’analyse de Cathy O’Neil sur la puissance de destruction de certains algorithmes mal intentionnés vaut aussi dans le secteur de l’éducation. Aux Etats-Unis, les étudiants les plus pauvres se retrouvent ainsi inondés de pubs pour des facs de seconde catégorie, qui misent tout là-dessus pour remplir leurs amphis, quand les bonnes facs jouent plus de la réputation de leurs profs
  • « Les ADM se nourrissent les unes les autres. Les gens pauvres ont tendance à présenter un mauvais score de crédit et à vivre dans des quartiers à forte criminalité, entourés d’autres individus pauvres. Une fois ces données digérées, le sordide univers des ADM les bombarde de publicités prédatrices pour des prêts hypothécaires de type subprime ou des écoles à but lucratif. Il déploie plus de policiers pour les arrêter et, lorsqu’ils sont condamnés, leur inflige des peines de prisons plus longues… ».
  •  
    "Attention, c'est du brutal. Algorithmes, la bombe à retardement (Novembre 2018, Les Arènes) nous plonge dans les cuisines peu ragoûtantes des fabricants d'algorithmes. L'auteure, Cathy O'Neil, est une mathématicienne américaine de haut vol passée de l'université aux banques puis aux analyses en big data avant de se reconvertir en lanceuse d'alerte. Elle pousse ici un cri d'alarme citoyen : nous ne pouvons pas rester spectateurs d'un monde où nous sommes de plus en plus tributaires d'outils conçus de façon opaque, utilisés à des fins commerciales ou sécuritaires et ayant pour conséquence d'exacerber les inégalités."
Aurialie Jublin

Where Next for #platformcoop? - Danny Spitzberg - Medium - 0 views

  • But in terms of vocational shifts or better business practices, it’s unclear what #platformcoop has produced. The original excitement is giving way to unmet expectations.
  • In Baltimore, a group of returning citizens — men and women who were formerly incarcerated — faced ridiculously unfair barriers to employment. And so, over the past two years, they formed Core Staffing, a staffing agency with 12 members.
  • When I asked about where Core Staffing is investing its energy lately, Joseph said, “we’re trying to figure out how to foster close relationships and community while managing a distributed workforce. What will keep people engaged outside of profit?” Core Staffing currently has 15 members, fewer than virtually any other platform co-operative, but these tensions are already very present.
  • ...8 more annotations...
  • A second question for Joseph revolves around money. Core Staffing is considering a large loan to build its platform. Joseph says, “We’re not going to be profitable for at least a year. I don’t want to us take on a huge amount of debt. But we have unanimous decisions for financial decisions that affect everyone’s equity, and right now, the members are voting ‘yes’ to take it.”
  • We all should obsess less over which organizations may or may not be “real” platform co-operatives, and reframe the conversation to assess how they model cooperativism — such as democratic governance, collective decision-making, and how their capital is made accountable to workers. This approach would cultivate the curiosity and humility necessary to help new participants.
  • Yet their basic questions remained unanswered. Where did “platform cooperativism” come from? How is a “platform” different than a website — don’t most co-ops have an online presence?
  • The more it can succeed with disintermediation, removing links in their supply chains and systems, the more we all learn and benefit. This begins with meeting people where they’re at.
  • While writing this article, I surveyed people about trends and challenges in the co-op movement. I polled groups obsessed with digital tools and online platforms, and yet the responses were far more grounded than I expected. In fact, the two challenges Joseph sees for Core Staffing echo everything: appropriate finance, and engagement at scale.
  • I believe radical technologist Micky Metts said it best: “I have run into many people that have existed in the corporate world but really do not understand co-operative engagement, even if they are caring and loving individuals.” In her mind, removing the fear we’re conditioned with is the first step for co-ops to grow. Data Commons co-founder Noemi Giszpenc takes it a step further by urging “true participation” among all users.
  • After two days of prototyping with Sylvia Morse and Up & Go, a New York platform for home cleaning, I learned that all platforms compete on quality. The members, mainly latinx women, are cleaning professionals. They provide reliable, consistent, five-star service. One user said, “I honestly don’t care if workers own the app.” Ironically, that indifference is a reason why platform co-ops like Up & Go can change the narrative about on-demand labor. It’s a labor of love, and it’s much more than an idea.
  • Robin Hood Co-op, a Finnish activist hedge fund, illustrates the diversity of co-operative models we have in this movement. Ana Fradique, a culture worker and community coordinator with Robin Hood, told me that all bureaucratic structures familiar to co-ops “need to be revized and supported by more flexible, faster technologies and modes of co-operation.” Her perspective comes from “experience with the different gears of operating online networks, where power and decision is faster than what the formal structure allows.” In a listserv, emails rarely achieve the escape velocity required to move from discussion to action.
  •  
    "I wrote this article to examine the idea of "platform cooperativism," where it's headed, and what it needs in order to use technology for economic justice."
Aurialie Jublin

Have We Already Lost the Individual Privacy Battle? - Post - No Jitter - 0 views

  • Individual cookie management and installing special anti-tracking software has never been a common activity, so the majority of users continue to contribute to the collection of personalized data, even if it makes most of us uneasy. We don't trust most Internet sites, but we use them anyway. Pew Research Center wrote about the fate of online trust and summarized that many experts doubt the possibility of progress, saying "people are insured to risk, addicted to convenience, and will not be offered alternatives to online interactions."
Aurialie Jublin

Can Worker Co-ops Make the Tech Sector More Equitable? | The Nation - 0 views

  • Fed up with this heartless model, some tech activists are developing online workplaces that operate as worker-driven communities. Daemo, a pilot program incubated at Stanford University’s Crowd Research Collective, is one such worker-driven crowd-labor platform. Since 2015, Daemo’s developers have been building on MTurk’s interface with a communications system aimed at allowing for more equitable “matching” between work requesters and digital taskers. As a non-hierarchical, nonprofit framework where workers control the operations, Daemo is designed for fairer working conditions, with a minimum wage of $10 an hour, which is a major improvement on MTurk’s precarious labor-outsourcing system.
  • Some former participants in Daemo’s project recently aired sharp criticism of the platform in response to a largely favorable article in Wired. In a collectively authored article on Medium, they argued that, in their practical experience with the platform, decision-making power rests with a “platform team” of researchers and leading developers. Though Daemo has established a Constitution that theoretically is open to amendments and revision based on workers’ input, critics say the day-to-day management remains tightly controlled by researchers.
  • “Whenever they talk about the decentralization, they talk about technical decentralization, like block-chain or decentralized platforms, but most of the time they overlook the governance level, which is more important,” Hashim says. “So it’s about who takes the positions, it’s about who has the right to access information. If you don’t have a well-informed society, you don’t have democracy.”
  • ...4 more annotations...
  • Kristy Milland, an activist with the MTurk advocacy network We Are Dynamo, says she’s given up collaborating with Daemo because “There hasn’t been any deep, involved worker input…. It’s built by academics with the bias they bring to such a platform that they expect will provide them with free data to publish down the road. Just like Amazon built MTurk with their needs in mind, even if many of the roadblocks this caused may have been unintentional.”
  • The “platform cooperativism” concept, as articulated by technologist Trebor Scholz and other academics, is that worker control can be integrated by working with the democratic aspects of the online sphere: entrepreneurial horizontalism and a pluralistic culture of innovation. But with online workspaces proliferating at breakneck speed, it’s a race to see whether these more principled worker-led models will ever be able to compete for market share with the app-based workforce of MTurk. Similarly, small-scale cab-service cooperatives are emerging in the United States, but Uber and Lyft’s mega brands are displacing cabbies by the minute.
  • The problem with crowd labor isn’t that it’s big, or complex; it’s that workers can’t control their means of technological production. According to Joshua Danielson of the Bay Area start-up cooperative Loconomics, Daemo’s model “has the potential to provide an alternative to Amazon Turk,” if the platform combines a good product and good jobs for the producers. The key, he says via e-mail, is “creating a cooperative business model that can be self-sufficient and be able to attract clients. The latter is the more challenging one given the deep pockets of the current players. That said, it’s important to remember that workers are the product, not the platform, and they hold an immense amount of power if they can organize.”
  • The digital frontier offers endless room both for exploitation and for social transformation. But if workers can get ahead of corporations in harnessing the potential of open-source technology, they can disrupt the incumbent Silicon Valley oligarchs from below. So far, technology hasn’t emancipated labor nearly as rapidly as it has liberalized markets. Cooperative thinking can make technological power part of the solution, but only if it’s matched with people power.
  •  
    "The crowdwork sector is dominated by low-paid gigs-can communally run companies make these jobs sustainable?"
Aurialie Jublin

(Un retour sur) Peertube. - Mr. Funk E. Dude - 0 views

  • One of the ways that it distinguishes itself from YouTube and Vimeo however, is in its peer to peer sharing structure. If just one person is watching your video, then it’s pulling it straight from the server. If two or more are watching your video, then it shares bits of the video between them, creating less activity on the server. The idea is that, like all other decentralized social networks, you can run your own instance of the software and the server load is significantly reduced.
  • As with most decentralized social networks each instance of PeerTube has it’s own rules, guidelines, and restrictions. The biggest difference between the instances probably boils down to daily upload file limitations. It’s important when choosing an instance to see what their restrictions are as it may effect your experience.
  • My first problem is with finding anything interesting to watch. Because PeerTube is worldwide you’ll notice right away that a LOT of the videos on the front page are mostly in different languages. French, Spanish, Russian, English, and others. This isn’t necessarily a bad thing in and of itself, but it does slow down the amount of time it takes to find a good video to watch. An option to sort by language would be a big step.
  • ...4 more annotations...
  • My next problem is with theft. There are a LOT of stolen YouTube videos. I’d say more then 3/4 of the videos on most of the PeerTube instances I’ve visited are taken from YouTube. I get how some people might have a grudge against Google and therefor YouTube and think that copying the videos and placing them on PeerTube might hurt YouTube, but in reality it hurts the creators that made them in the first place.
  • So if 3/4 of PeerTube Videos are just stolen YouTube videos, why not just go to YouTube? It’s the same problem any video sharing platform has. Content. If there’s not enough people creating content to watch, no one is going to use the service. YouTube is a HUGE source of content that no new service can really compete directly against. Turning PeerTube into a YouTube dumping ground is just admitting defeat. Until PeerTube finds a way to attract content creators, it’s going to struggle.
  • My last problem is with porn. Now, I’m no prude. I’ve got no issues with porn. It’s great, when in the right context. The problem is when it’s someone slinging their dick like a propeller right next to a video on how to adjust the settings on your hard drive. Fortunately most people put their porn behind content warning which blur the thumbnail and make you click on a warning to see the video, but not everyone does. Most instances will ban people for not using content warnings, but if someone is running their own instance, then the moderators for other instances have to ban the offenders instance from the Federation. It can be a slow process.
  • Until PeerTube fixes these four issues it’s going wallow in obscurity. It’s a good idea, being able to share videos across different social media platforms. Lessening server stress by using P2P. Decentralization. Federation. Content warning. All useful to the users, but it’s the content and how it’s managed by both the users and each instances moderators that will determine if, on a whole, PeerTube can survive and thrive like Mastodon, Pixelfed, or Friendica.
  •  
    "I know I've said a lot of great things about federated social networks and for the most part they are. Nothing is perfect however. With that said, let me introduce, and explain the problem I see, with PeerTube."
Aurialie Jublin

What the hell is a blockchain phone-and do I need one? - MIT Technology Review - 0 views

  • All of a sudden, several crypto-focused handsets are hitting the market, or will soon. The biggest player in the new game is Samsung, which confirmed this month that the Galaxy S10 will include a secure storage system for cryptocurrency private keys. It joins HTC, which for months has been touting the Exodus 1; Sirin Labs, which used proceeds from a huge ICO to build the Finney; and Electroneum, which this week began selling an $80 Android phone that can mine cryptocurrency.
  • In the wildest dreams of enthusiasts, these devices will be a gateway to something called the decentralized web, or “Web 3.0.” In this future version of the internet, blockchains and similar technologies would support decentralized applications—“dapps”—that look and feel like the mobile apps we use today but run on public, peer-to-peer networks instead of the private servers of big tech companies.
  • It’s widely thought that a major impediment to mainstream adoption of cryptocurrency and dapps is that these technologies are too difficult to use for people who are not especially tech savvy. Better user experiences, starting with cryptographic key management, could change that. But getting there is not straightforward, given that key security is paramount: you lose your keys, you lose your assets.
  • ...2 more annotations...
  • This also explains why Ethereum creator Vitalik Buterin seems so excited about one particular feature of HTC’s Exodus 1, called social key recovery. Essentially, users can choose a small group of contacts and give them parts of their keys. If they lose their keys, they can recover them piece by piece from their contacts. Buterin, as usual, is looking far down the road, in this case to a future where people use blockchains to maintain more control over their digital identities and personal data than is generally possible today. Social key recovery is “arguably an early step toward formalized non-state-backed identity,” he tweeted.
  • Indeed, even if these phones take off, the decentralized web will still be mostly a dream. Construction of its foundational infrastructure is in the beginning stages. Perhaps an influx of new users would spawn compelling new applications, which might in turn inspire the development of new infrastructure. But the best the first round of blockchain phones can do is give us a glimpse at a potential future that’s still a long way off.
  •  
    "The first wave of crypto-focused smartphones from big players like Samsung is a small step toward a decentralized web."
Asso Fing

La consommation électrique des logiciels, un enjeu pour les industriels et… l... - 0 views

  • Troisièmement, parce que le législateur pourrait intervenir… « En 2020 entrera en vigueur une législation européenne sur l'écoconception des serveurs et des unités de stockage en ligne, rappelle Rabih Bashroush, professeur en systèmes distribués et génie logiciel à l'Université de Londres-Est, qui a  participé à l'élaboration de ce texte . Il faut un cadre similaire pour la consommation des logiciels. » En France, un député, Fabrice Brun (LR), vient de  demander au gouvernement d'obliger les éditeurs de logiciels à « pratiquer une écriture plus vertueuse, en terme environnemental, du code informatique. »
  • Enfin, parce que jusqu'ici fabricants et gestionnaires ont tenté de réguler la consommation en jouant sur des solutions « matérielles » : utilisation de processeurs ou de cartes mères qui chauffent moins, de mémoires moins énergivores ; déclenchement des sauvegardes la nuit quand les serveurs sont moins occupés… Or, pour certains experts, ces gisements d'économies sont désormais épuisés. « Dans les dix ans à venir, il n'y aura pas de révolution technologique, facilement déployable à grande échelle, qui permettra de réduire la consommation énergétique du numérique : la seule voie possible est celle de la sobriété », affirme Hugues Ferreboeuf, qui dirige les réflexions sur l'empreinte carbone du numérique  au sein du think tank The Shift Project . « Nos data centers comptent d'ores et déjà parmi les plus économes du marché, mais nous devons aller plus loin, s'inquiète François Sterin, directeur industriel chez l'hébergeur Web OVH. Cet enjeu est crucial pour les fournisseurs de cloud, tant du point de vue commercial que sociétal. »
  • Reste donc une seule piste : la consommation induite par les logiciels. La mesurer est assez simple pour les smartphones et les objets connectés, plus compliqué pour les serveurs et PC. « On peut brancher des appareils de mesures - des sortes de wattmètres - sur les smartphones pour récupérer les données de la batterie, explique Thierry Leboucq. La mesure réelle permettra demain d'établir des normes et des labels pour valider l'éco-consommation d'un service numérique. » Pour les objets connectés, qui ne comportent que quelques composants, Alexandre Boé et son collègue Thomas Vantroys, de l'université de Lille, proposent déjà une solution. « Nous avons mis au point un petit appareil de mesure, composé d'un microprocesseur et d'un écran, coûtant une soixantaine d'euros, détaillent-ils. Couplé à des marqueurs introduits dans le logiciel, qui repèrent les fonctions les plus gourmandes, il permet aux fabricants et aux développeurs d'optimiser le fonctionnement et la consommation. »
  • ...6 more annotations...
  • La mesure est plus compliquée sur les serveurs et les PC. « Par exemple, les couches d'abstraction matérielle, ces logiciels intermédiaires entre le système d'exploitation et le matériel informatique, consomment de l'énergie que vous ne voyez pas », met en garde Marc Vautier, qui anime la communauté des Orange Experts « Energie et environnement » au sein du groupe Orange. Dès 2016, Jean-Patrick Gelas, docteur en informatique, avait publié, avec son équipe de l'université Lyon-1, un article exploratoire sur la mesure en temps réel de la consommation, dans la revue « International Journal of Electronic Business Management ». « L'idée était de coupler les mesures de puissance électrique, réalisées par un wattmètre, avec les instructions du logiciel et les tables de correspondance 'fonction réalisée/énergie dépensée' publiées par les constructeurs », explique-t-il.
  • Romain Rouvoy, professeur d'informatique à l'université de Lille, propose une approche basée sur l'estimation. En collaboration avec Orange et le cabinet Davidson Consulting, il a mis au point PowerAPI, un prototype de recherche distribué sous licence open source, disponible sur la plate-forme GitHub.  « Une fois installée sur un serveur, un ordinateur portable ou un robot, notre logiciel va évaluer, à l'aide de modèles mathématiques que nous avons développés, la consommation électrique des applications qui tournent sur ces machines », relate-t-il. Plusieurs industriels de l'automobile et de la robotique se disent intéressés par sa technologie.
  • « Les développeurs ne doivent pas s'empêcher de discuter des cahiers des charges avec les services marketing, qui peuvent, par exemple demander des serveurs disponibles 99,99 % du temps, là où 99,9 % suffiraient et permettraient d'économiser de l'énergie », conseille Erwan Diverrez, architecte logiciels chez Orange.
  • « On peut jouer sur le stockage de données - par exemple les données d'usage qui décrivent la relation de l'utilisateur avec sa machine - et ne les envoyer que lorsque le smartphone est en Wifi ; cela évite d'essayer continuellement de se connecter au réseau mobile », ajoute Sébastien Brault, développeur sur mobile chez Orange.
  • « Pour le Web, il existe des astuces comme la gestion du cache de la mémoire, par exemple, pour ne pas avoir à télécharger plusieurs fois une image qui apparaît sur plusieurs pages d'un site », ajoute Romain Petiot.  « On peut aussi jouer sur les langages de programmation utilisés et les bonnes pratiques de développement associées à ce langage », estime Romain Rouvoy, professeur d'informatique à l'université de Lille.
  • « La moitié de l'empreinte carbone du numérique provient de la production des équipements », résume Hugues Ferreboeuf, coordinateur du rapport « Pour une sobriété numérique » publié par le think tank Shift Project. L'autre moitié se répartit en trois tiers : la consommation des serveurs dans les data centers, les réseaux de télécommunication et les quelque 15 milliards de terminaux et périphériques sur lesquels tournent des logiciels.
  •  
    "Des logiciels tournent sur 15 milliards de terminaux et périphériques à travers le monde. Ce sont eux qui en commandent les besoins en énergie."
  •  
    "Des logiciels tournent sur 15 milliards de terminaux et périphériques à travers le monde. Ce sont eux qui en commandent les besoins en énergie."
1 - 19 of 19
Showing 20 items per page