Taylor Swift a encore trouvé un moyen d'innover, et de manière remarquable.

Entre les États-Unis et l'Inde

I shortened my sentences. I used keywords that Google could identify easily. I wrote in a way that allowed Google to understand our content, following a set of nebulous rules laid out by the overlords. The rules were never expressly stated, but trial and error from millions of people that came before our blog seemed to point to a common path. And that path was “Juicy Info Nuggets”.

In September of 2023 Google started rolling out what it called its “Helpful Content Update“. The stated goal of this update to search results was to reduce the amount of spam and AI-generated crap that littered the internet. The actual result? The destruction of small publishers, blogs, and the Rise of Reddit.

And where does that leave bloggers, small publishers, and the like? Well, after years of training online publishers to write in a fashion that Google web crawlers could easily identify, in easily digestible nuggets of information, Google will be taking those nuggets and giving them away for free. Because hey, when you’re a huge monopoly you can do whatever you want.

Les entreprises ITSM-NG avec XWiki et
ITSM-NG, en lient avec Opensource Experts.

Un audit d'accessibilité peut être réalisé avec l'outil Ara

Basé sur la dernière version du référentiel général d’amélioration de l’accessibilité (RGAA 4.1), avec Ara vous pouvez :
– Faire un état des lieux (audit partiel de 25 ou 50 critères)
– Faire un audit complet, dit de conformité (106 critères)
– Générer votre rapport d’audit et votre déclaration d’accessibilité

A good example comes from looking towards the impact of incorporating AI into search engines such as Google. A single generative AI query could use 4 to 5 times more energy than a regular search engine query. Others found that average energy consumption per search could be 6.9–8.9 Wh, compared to 0.3 Wh for a standard Google search. This gives us an enormous range of 4-30 times larger. Whichever end of the scale the figures land, it’s a significant increase.

Does the result quality increases at the same rate?

The report is available on the green web foundation

Working with SemVer leads to two observations:

• only two version numbers are sufficient: there's no real difference between patch and minor.
• deprecate than remove cycle: using a major version only leads to the removal of deprecation warnings and their associated feature. It means bumping for 1 to 2 should works, if the version 1 does not throws warnings anymore.

how to judge the semver approach? Either describe the thing actually behaves (descriptivist) or describe about the right and wrong ways to use a particular tool (prescriptivist).

Another instance of this pattern playing out I’ve noticed are log levels. You can get very philosophical about the difference between error, warn and info. But what helps is looking at what they do:

• error pages the operator immediately.
• warn pages if it repeats frequently.
• info is what you see in the prog logs when you actively look at them.
• And debug is what your developers see when they enable extra logging.

A list of things to avoid

C'est toujours compliqué de détecter les messages générés par IA, mais il y a ici une tendance: les messages générés augmentent.

Trois startups proposent trois solutions::

1. Une horloge atomique miniaturisée de la taille d'un boîtier ATX horizontal par Infleqtion et son système Tiqker. Cette boîte vient de décrocher le jackpot avec 11 millions de dollars du Département de la Défense américain. Matthew Kinsella, leur CEO, l’explique simplement : “Si vous connaissez votre point de départ, vos changements de vitesse et de direction, et que vous disposez d’une mesure du temps ultra-précise, vous savez exactement où vous êtes. Et ça, c’est impossible à pirater !”
2. SandboxAQ utilise des capteurs quantiques pour détecter les variations du champ magnétique terrestre, comme les oiseaux migrateurs.
3. Rendre les capteurs quantiques plus sûrs via https://q-ctrl.com/

Le financement immoral des compteurs Linky au détriment des consommateurs.

Why research needs free open source softwares.

It shows two things:

• Science must not use commercial software
• Science is underpaid if it has to use pirated software

(via https://nicolas-delsaux.hd.free.fr/Shaarli/shaare/eptdLQ)

Are hCaptcha not accessible to screen readers?

Découverte d'alternatives libres et d'outils de protection contre le pistage facile à installer.