Autonomie numérique et technologique

Code et idées pour un internet distribué

Linkothèque. Carnet de bord


Depuis son implantation en France en 2012, Airbnb s’est imposée comme un acteur central de l’hébergement touristique. En une décennie, la plateforme – avec d’autres comme Booking ou Abritel – a profondément redessiné les pratiques de voyage. L’étude retrace cette expansion rapide : d’une présence ponctuelle dans les métropoles, les stations de ski et du littoral en 2013, elle s’est diffusée dès 2019 dans la quasi-totalité du territoire, épousant l’attractivité des grands sites patrimoniaux, des parcs naturels ou des littoraux, et connaissant une forte accélération post-Covid. Festivals, événements culturels et manifestations sportives créent des pics de réservations, révélant l’intégration des plateformes dans l’économie événementielle locale. L’analyse souligne aussi les tensions : rôle des résidences secondaires, concurrence ou complémentarité avec l’hôtellerie, effets sur le logement, mais aussi maintien de commerces de proximité. Côté usagers, un Français sur deux a déjà réservé via Airbnb ; les profils d’utilisateurs sont plus jeunes, plus diplômés, plus urbains que la moyenne, mais reflètent l’ensemble de la société. Les séjours sont courts (5 jours en moyenne), souvent en couple ou en famille, et motivés par des arbitrages budgétaires. Si les plateformes offrent souplesse et diversité, elles n’effacent pas toutes les inégalités d’accès aux vacances : classes modestes, jeunes femmes et ruraux restent les plus empêchés.

Nombre moyen de services et commerces de proximité dans les communes de 1.000 à 3.000 habitants en fonction du nombre de nuitées réservées sur Airbnb dans la comnune

Dans les bourgs n’affichant aucune réservation via Airbnb, autrement dit les moins touristiques, le nombre moyen de commerces et services de proximité recensés n’atteint que 3,5. Dès lors qu’un flux minimal de visiteurs se manifeste sur la plateforme, cet indicateur progresse nettement pour dépasser 5 commerces et services de base en moyenne. La tendance se confirme avec l’intensité touristique : dans les communes où le volume de nuitées franchit les 10 000 en 2024, la moyenne grimpe à 7 commerces ou services essentiels. Dans les localités les plus attractives, où les nuitées dépassent 20 000 par an, elle franchit même le seuil symbolique des 10.

changeme picks up where commercial scanners leave off. It focuses on detecting default and backdoor credentials and not necessarily common credentials. It’s default mode is to scan HTTP default credentials, but has support for other credentials.

changeme is designed to be simple to add new credentials without having to write any code or modules. changeme keeps credential data separate from code. All credentials are stored in yaml files so they can be both easily read by humans and processed by changeme. Credential files can be created by using the ./changeme.py --mkcred tool and answering a few questions.

Captura de pantalla de archivomedialabmadrid.org

El programa medialabmadrid ocurrió entre el 2002 y 2006 bajo la dirección de Karin Ohlenschläger y Luis Rico. Se concibe como un laboratorio abierto al diálogo entre las más diversas prácticas artísticas, científicas y tecnológicas. La estructura modular y abierta de este programa se articula en torno a cinco áreas de actividades interconectadas: Arte, Ciencia, Tecnología, Sociedad y Sostenibilidad (ACTS+S)

Captura de pantalla de spatialagency.net.

Spatial Agency is a project that presents a new way of looking at how buildings and space can be produced. Moving away from architecture’s traditional focus on the look and making of buildings, Spatial Agency proposes a much more expansive field of opportunities in which architects and non-architects can operate. It suggests other ways of doing architecture.

Captura de pantalla de feralatlas.supdigital.org.

Feral Atlas invites you to navigate the land-, sea-, and airscapes of the Anthropocene. We trust that as you move through the site—pausing to look, read, watch, reflect, and perhaps occasionally scratch your head—you will slowly find your bearings, both in relation to the site’s structure and the foundational concerns and concepts to which it gives form. Feral Atlas has been designed to reward exploration. Following seemingly unlikely connections and thinking with a variety of media forms can help you to grasp key underlying ideas, ideas that are specifically elaborated in the written texts to be found in the “drawers” located at the bottom of every page.

Naturopathie, magnétisme, psychologie énergétique… Depuis la crise sanitaire, les « médecines douces » prolifèrent. Un phénomène dangereux qui mène trop souvent à des dérives thérapeutiques, voire sectaires. Margot Brunet, journaliste spécialisée en sciences et santé, a enquêté sur ces charlatans dans son livre « Naturopathie. L’imposture scientifique » (éd. Les Échappés, parution le 9 octobre).

C’est tout le problème de cette nébuleuse, savoir comment les nommer. Si on évoque les « médecines douces », on leur donne un vernis scientifique, puisqu’on les fait entrer dans un cadre médical. On pourrait dire « pratiques non conventionnelles en santé », mais il faut reconnaître que c’est un peu long et, surtout, il s’agit davantage de bien-être que de santé. Quant à savoir ce qu’on met dedans, là aussi, c’est compliqué : on ne peut pas vraiment faire de liste de pratiques, puisque n’importe qui peut créer sa spécialité et choisir l’intitulé qu’il veut. Dans le livre, j’ai donc décidé de les définir comme toutes les pratiques non encadrées et non éprouvées scientifiquement qui se targuent d’avoir des effets sur la santé.

Pour beaucoup, se tourner vers les pseudo-thérapies est un acte de revendication, une opposition au système politico-sanitaire actuel. Les gens qui consultent ce genre de praticiens n’y vont pas particulièrement parce qu’ils sont déçus du système de santé, ils y vont parce qu’ils croient à un autre système de pensée.

D’autant que certaines pratiques sont désormais remboursées par les mutuelles, ce qui est un cercle vertueux pour ce marché : les gens paient plus cher leurs mutuelles pour avoir accès à ces médecines douces, et les pseudo-thérapeutes reçoivent plus de clients… qui viennent d’autant plus volontiers que les consultations sont remboursées par les mutuelles.

The easier approach is to cluster the markers on the client-side, due to leaflet plugins like Leaflet.markercluster.

…client-side clustering has its limitations. In order to ake it work, you need to load all the markers at once! This is not a problem if you have a small number of markers, but when you have thousands or millions of markers, it can quickly become a performance bottleneck.

Backend clustering addresses this problem by processing the data on the server. Only clusters for the current viewport and zoom level are sent to the client. This minimizes bandwidth usage and reduces client-side processing. In my solution, I utilized Supercluster, a high-performance clustering library suitable for Leaflet maps.

When I tried to render around 40k Geojson objects in Leaflet’s canvas mode, it took around 30 seconds to render the map. When I used Leaflet’s GeoJSON, instead of React Leaflet GeoJSONs, initial rendering took a few seconds. The response from the library maintainer was that React Leaflet is an additional abstraction and it’s expected that rendering is less performant. My first advice would be avoid leaning solely on this library.

While Leaflet’s canvas mode is an option, rendering over 100k objects stretches its limits. Fortunately, there are solutions that perform very well with large amounts of data, such as WebGL rendering engines and Vector Tiles. Leaflet doesn’t support WebGL or Vector Tiles out of box, but its plugins bridge the gap.

Leaflet.markercluster seems like a go-to solution, yet its performance falters beyond 100k markers. For a leap in performance, consider supercluster.

The Sora app introduces a particularly potent feature: the ability for users to grant permission for their likeness to be used by friends in video creations. With just a few spoken words and head movements, the AI can capture a person’s digital twin, ready to be inserted into any scenario imaginable.

This technology erodes the foundational trust we once had in video evidence. Previously, seeing was believing. Now, every video clip must be viewed with a healthy dose of scepticism. The potential for misuse is immense, from personal harassment and character assassination to large-scale disinformation campaigns.

While there are discussions around invisible watermarks and detection software, we are in a race against our own creations. As a society, we are largely unprepared for a world in which we cannot trust our own eyes. This isn’t a distant, futuristic problem; it is here now. The immediate need for digital literacy education and robust verification standards has never been more urgent.

Barlow fue granjero, letrista de The Grateful Dead, pionero de los derechos digitales, un conector natural, un acaparador de relaciones y capital social que contaba sus amistades por miles. Fue cofundador de The Electronic Frontier Foundation (EFF) y Freedom of the Press Foundation (FPF), y moldeador del tono y los argumentos con los que hemos defendido Internet en los últimos 30 años. Su Declaración de independencia del Ciberespacio publicada en 1996 fue capaz no solo de poner en la agenda la necesidad de combatir los intentos gubernamentales de controlar Internet, sino también de dotar a la reivindicación de una lírica, una fuerza y una imaginación fascinantes. Cómo resistirse a la utopía libertaria que proponía Barlow.

“Gobiernos del Mundo Industrial, vosotros, cansados gigantes de carne y acero, vengo del Ciberespacio, el nuevo hogar de la Mente. En nombre del futuro, os pido en el pasado que nos dejéis en paz. No sois bienvenidos entre nosotros. No ejercéis ninguna soberanía sobre el lugar donde nos reunimos”. Era la Declaración de Independencia del Ciberespacio de John Perry Barlow. Falleció en 2018, y para entonces muchos habitantes del ciberespacio sospechábamos que su idea, más que ingenua o utópica, era siniestramente similar al discurso tecnolibertario. Silicon Valley, recordemos, nació de la contracultura californiana y sus contradicciones.

el espíritu de los tiempos es favorable a las restricciones parciales a ciertas plataformas, y no solo en regímenes como el chino, donde sus ciudadanos están acostumbrados a las VPN. Los motivos, aunque legítimos y normalmente relacionados con la protección de la soberanía nacional, la lucha contra la desinformación o la defensa de los menores y los derechos de autor, desatan dudas sobre si se acabarán cercenando derechos mayores, porque la ejecución es compleja y la vigilancia de los vigilantes, aún más. Europa lleva tiempo levantando muros regulatorios. EE UU (que, por cierto, podría desenchufarnos de internet en un segundo) tendrá una versión de TikTok diferente. En España nos reímos del pajaporte, pero en Francia y Reino Unido se requiere una verificación de edad obligatoria para el acceso a las páginas adultas. En este último lugar la comprobación se aplica también a redes sociales, foros y otras páginas donde los menores pueden encontrar contenido nocivo, llegando a situaciones surrealistas como una posible restricción a la Wikipedia. Nadie nos prometió que internet fuera a ser una república utópica e independiente del mundo real, pero tampoco que fuera a convertirse en el caso opuesto.

I have a bold hypothesis.

I think AI companies might be intentionally configuring their smart models to not solve problems quickly, particularly the coding requests, because if they did, their revenue would be severely impacted by quick and short solutions.

If the model over engineers, then I need to talk to it a lot to fix the problem, particularly with setups that use APIs, so they get way more tokens burnt. SO more revenue for them (or I need a more unlimited subscription).

Yea, I know, sounds insane. But think about your experience with code generation — wasn’t it significantly more efficient before the models became smart/thinking/whatever? How come these significantly better models produce significantly more over engineering, where the only benefit is burning tokens?

To backup NextCloud database, first you need to find out the database name by running the following command on your NextCloud server.

sudo mysql -u root -e "show databases"

As you can see, my NextCloud database name is nextcloud. Once you have identified the NextCloud database name, use mysqldump utility to make a backup like below. Replace red text with your own NextCloud database name.

sudo mysqldump -u root nextcloud-database-name > ~/nextcloud.sql

o back up NextCloud config folder, first cd into Nextcloud installation directory. (Depending on your setup, your Nextcloud installation directory may be different. For example, If you set up Nextcloud with LAMP stack, the directory might be /var/www/nextcloud/. If you set up Nextcloud with LEMP stack, the data directory might be /usr/share/nginx/nextcloud/).

cd /var/www/nextcloud/

Then use tar to back up the config folder.

sudo tar -cpzvf ~/nextcloud-config.tar.gz config/

Run the post-upgrade routine.

sudo -u www-data php7.4 /var/www/nextcloud/occ upgrade

Regenerate initramfs for the New Kernel

To fix the issue, you need to regenerate the initramfs for the new kernel version. Run the following command in the terminal:

   sudo update-initramfs -u -k <version>

Replace <version> with the actual kernel version string for the kernel that you were unable to boot into. For example, it might look something like 4.15.0-36-generic.

You can find the kernel version by running uname -r if needed.

Update GRUB

Once the initramfs has been successfully generated, update the GRUB bootloader by running:

   sudo update-grub

This command ensures that GRUB recognizes the updated kernel and its corresponding initramfs.

Captura de pantalla de citybik.es.

Some years ago we wanted to create an android app for our local bike sharing system and found out there was no open data available for us to do so. Other systems around the world had exactly the same problem.

Instead of creating yet-another-app, we created Citybikes, a project that provides bike sharing data for apps, research and projects to use.

To date Citybikes supports more than 400 cities and the Citybikes API is the most widely used dataset for building bike sharing transportation projects.

Captura de pantalla de once.com

Something happened to business software.

You used to pay for it once, install it, and run it. Whether on someone’s computer, or a server for everyone, it felt like you owned it. And you did.

Today, most software is a service. Not owned, but rented. Buying it enters you into a perpetual landlord–tenant agreement. Every month you pay for essentially the same thing you had last month. And if you stop paying, the software stops working. Boom, you’re evicted.

El cineasta presenta su particular mirada a la ciudad en una obra que funciona como una película sinfónica a la que Guille Galván (Vetusta Morla) pone banda sonora

Se titula Madrid, Ext., se lee “Madrid exterior”, pero las letras de “ext” también podrían ser las tres primeras de “extinción”. La última película de Juan Cavestany, que se estrena este viernes, es una mirada a Madrid, su mirada y su Madrid. El comienzo es una declaración de intenciones: el Museo de Ciencias Naturales. Quizá no es lo primero que a cualquiera le viene a la cabeza si piensa en esa ciudad, pero es uno de los primeros recuerdos de este madrileño de 58 años. Y, entonces, cambia la percepción: ¿cuántos niños madrileños han quedado ―como Cavestany― impactados por ese elefante disecado del museo? Es probable que esa imagen esté impregnada en las neuronas de muchos vecinos de la capital. Y así, ya se entra en el documental: el espectador ya es cómplice, reconoce esas imágenes y las hace suyas, ya es una mirada compartida. Además, al director le gusta la idea de museo como quien clasifica, ordena y muestra sus piezas, exactamente lo que él pretende en el filme, que sea un repositorio, un archivo de la flora y fauna de la ciudad, mucha de ella en peligro de extinción.

El metro, túneles (porque el Madrid exterior también es interior), obras, cines, quioscos, videoclubs (ojo a las expectativas de quien llamó a uno de ellos Videoclub Siglo XXI y a la historia del videoclub Star), ultramarinos, bares con barra de zinc y baldosas de terrazo, autobuses de la EMT, la Castellana, zapaterías, panaderías, portales más grandes que muchos pisos, porteros, andamios, grúas, toldos verdes, mucho ladrillo visto, cementerios, peluquerías, la bolera de Chamartín, estudios fotográficos, la procesión de Jesús el Pobre, ropa tendida, hospitales, piscinas, barcas, el estanque del Retiro, el de la Casa de Campo, salas de baile, comercios con entidad propia: Fajas Ruiz, Muelles Ros, el bar Muñiz…

OUTLINE:
(00:00) – Introduction
(00:58) – Sponsors, Comments, and Reflections
(08:48) – Programming – early days
(26:13) – JavaScript
(36:32) – Google Chrome and DOJ
(44:19) – Ruby programming language
(51:30) – Beautiful code
(1:09:31) – Metaprogramming
(1:12:52) – Dynamic typing
(1:20:10) – Scaling
(1:33:03) – Future of programming
(1:50:34) – Future of AI
(1:56:29) – Vibe coding
(2:05:01) – Rails manifesto: Principles of a great programming language
(2:29:27) – Why managers are useless
(2:38:48) – Small teams
(2:44:55) – Jeff Bezos
(3:00:13) – Why meetings are toxic
(3:07:58) – Case against retirement
(3:15:15) – Hard work
(3:20:53) – Why we left the cloud
(3:24:04) – AWS
(3:33:22) – Owning your own servers
(3:39:35) – Elon Musk
(3:49:17) – Apple
(4:01:03) – Tim Sweeney
(4:12:37) – Fatherhood
(4:38:19) – Racing
(5:05:23) – Cars
(5:10:41) – Programming setup
(5:25:51) – Programming language for beginners
(5:39:09) – Open source
(5:48:01) – WordPress drama
(5:59:18) – Money and happiness
(6:08:11) – Hope

This is a great 100% free Tool I developed after uploading this video, it will allow you to choose an LLM and see which GPUs could run it… : https://aifusion.company/gpu-llm/

Min Hardware requirements (up to 16b q4 models) (eg. Llama3.1 – 8b)
RTX 3060 12GB VRAM : https://amzn.to/3M0HvsL
Intel i5 or AMD Ryzen 5
Intel i5 : https://amzn.to/3WGZtp3
Ryzen 5 : https://amzn.to/46IigoC
36GB RAM
1TB SSD : https://amzn.to/4cBebEd

Recommended Hardware requirements (up to 70b q8 models) (eg. Llama3.1 – 70b)
RTX 4090 24GB VRAM : https://amzn.to/3AjIHow
Intel i9 or AMD Ryzen 9
Intel i9 : https://amzn.to/3YCeLxW
AMD Ryzen 9 : https://amzn.to/3YIaUiT
48GB RAM
2TB SSD : https://amzn.to/3YFQ83A

Professional Hardware requirements (up to 405b and more) (eg. Llama3.1 – 405b)
Stack of A100 GPUs or A6000 GPUs
https://amzn.to/3yojZ5T
Enterprise grade CPUs
https://amzn.to/3YDgByw
https://amzn.to/4dEbfY2

El efecto bouba/kiki fue descubierto por el psicólogo Wolfgang Köhler en 1929.[1]​ En sus experimentos llevados a cabo en la isla de Tenerife, donde el español es la lengua primaria, Köhler mostró formas similares a las de la imagen de la derecha a una serie de sujetos, y encontró una fuerte preferencia a asociar la forma puntiaguda con el nombre «takete» y redondeada con el nombre «baluba» («maluma» en la versión de 1947).

The « Extinct Media Museum Otemachi » is a privately owned museum that houses an array of obsolete devices spanning from the 1930s to the present day. It showcases items like the « home video camera, » « APS camera, » « digital camera, » « typewriter, » « personal computer, » « PDA, » and « recording media » from various eras.

This museum stands out for its unique approach, where many of the devices on display are still operational. Visitors are encouraged to handle the items and examine them closely. However, it’s advisable to seek assistance from the staff to safely remove and handle exhibits from the shelves.

An access control list (ACL) is a more advanced approach to security. It can implement the user/group/anonymous user approach with the basic rwx attributes but typically implementations do much more. In the case of Linux, the POSIX access control list (ACL) is usually supported.

Mandatory Access Control (MAC) is a policy-based approach that provides even more control over how security is implemented and controlled. It extends past the basic access controls of DAC and ACL to allowing an administrator to have fine grained control over what changes users can make. With DAC, a user simply needs write access to be able to change the attributes of a file or directory. The ability to create a file requires write access to the directory a file will be created in.

MAC systems normally provide the ability to specify access attributes as well as the ability to see and modify those attributes. The ability to create and delete files or directories can provide fine grain control. In addition, files and directories can have attributes that can be matched against rules that can control where and how data can be used.

MAC systems usually extend their control beyond the file system. This allows network interfaces, ports and other logical and physical devices to be monitored. This approach can even extend to services such as a system’s firewall. An application can be limited to the ports and interfaces they are allowed to use as well as the files, directories and other resources such as applications they have access to. SELinux is the MAC normally associated with Linux but there are others as well.

I don’t use chroot, but the default setup for modern versions of FPM already compartmentalizes everything adequately for example, the private /tmp directory. I agree with others that chroot is an outdated way of doing things.

Also, I use SELinux…yet another way of achieving many of the same goals of chrooting. I’d highly recommend setting up SELinux if you are not already using it. If you’re concerned enough about security that you’d even think of chrooting php-fmp, you probably want to set up SELinux and have it on « Enforcing » (it’s useless on « Permissive » mode, that’s really only suitable for the configuration phase of test servers.) Not only will it provide security with PHP, but you get a whole bunch of other security benefits of it.

I have done some pretty sophisticated things with a web server under SELinux, requiring me to manually change a number of policies, and while I have had a few prolonged sessions of frustration, maybe 3-4 hours at a time of banging my head against the wall trying to get the permissions set up properly, it is totally worth it. It’s all up-front work, and once you learn how to do it it’s very easy.

In context of fresh Chinese smartphones’ telemetry leaks – OnePlus, Xiaomi, Oppo and Realme, the data privacy concerns are rising again. Wearable devices know their user better than anyone else: where he/she lives, eating, sleeping; which sport, TV show, store they like and which politics they support. Even more, they can predict what user wants to eat on lunch with big chances to success, simple ’cause they know what he/she bought last week. We don’t need to even discuss a value of health, finances or private photos – leaking a small piece of related information can make a big hole in a ship called « Comfort Life ». No matter who you are – an ordinary worker, bank owner, green activist or journalist specialising on corruption. Some people will surely disagree here with the slogan « I have nothing to hide » slogan – folks, no point in discussing it anymore ’cause a billion places on Internet still host the finsihed battles; this Wikipedia page should uncover the main points.

Enter the Open-SmartWatch from [pauls_3d_things]. This ESP32 powered watch packs a gorgeous circular 240×240 TFT display, DS323M RTC, BMA400 three-axis accelerometer, and a 450 mAh battery inside of a 3D printed enclosure that can be produced on your average desktop machine. WiFi and Bluetooth connectivity are a given with the ESP32, but there’s also an enhanced edition of the PCB that adds another 4 MB of RAM, a micro SD slot, and a Quectel L96 GPS receiver.

An Open Source Smartwatch For Your Favorite Devices. Low Cost, High Fidelity.

The PineTime is a free and open source smartwatch capable of running custom-built open operating systems. Some of the notable features include a heart rate monitor, a week-long battery, and a capacitive touch IPS display that is legible in direct sunlight. It is a fully community driven side-project which anyone can contribute to, allowing you to keep control of your device.

We envision the PineTime as a companion for not only your PinePhone but also for your favorite devices — any phone, tablet, or even PC. In its current state the PineTime ships with a community firmware called Infinitime, which works with Linux computers and phones, Windows machines, and Android devices. IOS support is currently being looked into for Infinitime.

A fully custom open source Smartwatch with all the features you are used in normal smartwatches

Disclaimer I haven’t worked on this in ages. I never got to get all Parts working together. I would not suggest building it on your own at the moment. If you want to contribute feel free to do so.

Project Airframe is an open-source smartwatch designed to give developers and hobbyists full access to both hardware and software. With state-of-the-art sensors, wireless charging, GPS, and more, it’s a versatile tool for exploring wearable technology. Below, you’ll find everything you need to get started, from hardware specifications to software instructions.

Le Cunff defiende todo lo contrario al consenso popular, que dice que para conseguir el éxito es necesario centrarse en un objetivo grandioso y dirigirse hacia él sin distracciones. Ese camino lineal, dice, favorece la presión social, el burnout, la competición y sorpresas como la que se encontró ella: un destino que la hacía infeliz. Por eso es mejor ir fijándonos en aquello que nos ilumina la mirada y probarlo realizando pruebas controladas y poco arriesgadas.

Para conseguirlo, propone realizar pactos con nosotros mismos bajo el siguiente esquema: “Haré x durante y tiempo”, donde x es algo que nos provoque interés e y una temporada corta, de hasta tres meses. Es decir, menos “a ver si voy al gimnasio” y más “iré a clases de baile dos veces a la semana durante el mes de septiembre”. Y cuando ese tiempo termine, como pequeños científicos, evaluaremos la situación.

Una meta demasiado clara nos impide ver las posibilidades que se van desplegando ante nosotros. Y, ya que vamos a tratarnos como sujetos de prueba, tiene más sentido ir construyendo el camino poco a poco, disfrutándolo, que empecinarse en alcanzar una promesa final que quizá ni siquiera exista.

  1. atime — access time = last time file opened
  2. mtime — modified time = last time file contents was modified
  3. ctime — changed time = last time file inode was modified

Numeric arguments can be specified as:

+n for greater than n,
-n for less than n,
n for exactly n.

Note that « less than » means « strictly less than », so -mtime -14 means « last modified at the current time of day, 13 days ago or less » and -mtime +14 means « last modified at the current time of day, 15 days ago or more ».