Well said. I think this kind of "no algorithm = fair algorithm" is a bit like "free market = fair market". no way!

But at least is the type of feed where on top can be built additional filters defined by the user? maybe? I still have to see and try the feed customisation.

muchas gracias, un artículo en español explicando investigación: www.elsaltodiario.com/glovo/glovo-... pero la investigación de la autoridad ha añadido elementos muy importantes que aún no han sido relatados. (+ atribuciones correctas: Gaetano Priori análisis técnico; I operativa y estrategia)

Right! the only explanation I could come up with was that it was such a multidisciplinary effort that it became difficult to make it convincing. I want to believe at this point that the aim was to prove a point -to replicate it. it would be a fail if we had to spend 5 years on each and every action

scusa forse ho usato la parola sbagliata. non "impugnare" nel senso di contestare, ma nel senso di replicare. prender e riutilizzare. declinare nel loro stato. non sarebbe una guerra, ma un lavoro di concerto. potrebbe funzionare?

Forse hai risposta: la sentenza del Garante riguarda GDPR e statuto lavoratori [quindi, solo italiano], ma il diritto del lavoro in EU non è poi così diverso, quindi, sarebbe possibile per altre DPA impugnare questo procedimento e replicare l'azione verso Glovo, se presente sul loro territorio?

watch a video of our talk at the CCC, where we explain both the methodology and the reverse engineering techniques in detail: reversing.works/posts/2023/1... (-6/end)

The Italian DPA investigation covers many aspects, but if you want an overview of our contribution, you can read it at algorithmwatch.org/en/glovo-tra... or (5/6)

Meanwhile, we should also acknowledge the success of the Italian Data Protection Authority, which once again proved to be up to the challenge like no other in the EU, and the Digital Freedom Fund, which supported our efforts (as tracking.exposed) (4/6)

We hope that trade unions, workers' councils and other NGOs fighting for digital rights will take this as an example and support a more adversarial approach when it comes to exposing abuses. (3/6)

Two of the main findings, tracking outside working hours and sending private data to third parties, were the result of our technical analysis, also documented here: www.etui.org/publications... (2/6)

Ciao Christian! quale reazione radicale immagini? che rivedano le loro logiche (forse cambieranno infrastruttura del tutto, visto che sono fusi a delivery hero? ec.europa.eu/commission/p...) o che abbandonino il mercato italiano? non so cosa aspettarmi.

promoting secure, anonymous communication methods. If you’d like to reach out securely, you can find me at faffa-securesubmission.eu, where even I can’t know who’s contacting me—unless, of course, they choose to share.

Otherwise DMs are open!

A truly open LLM system needs to document all this overridden information for future peer review, and to update the accuracy of the information requires the same power that censorship and "AI alignment" have used.

Is there any work I'm missing that tries to guarantee such accountability?

This is true for factoids in our brains, paper hard copies, search engine indexes, LLM vector spaces, and historical video transcripts.

I made 400 days!! and then abandoned. It was just a game and not an effective way to learn a language.
I basically lost against gamification :D

Last week we were submitted to Hacker News and this gave us a good temporary visibility, converted in receiving many new political figures and new faces. Please consider sharing Don't Spy EU until we are not sure has been seen by the interested subjects. The time is running out, thank you!

Of course, for a moment I was going to forget to include my current Prime Minister, who is also responsible for this dumpster fire that the AIAct could become. Btw AI misgender and misread most of what it claim to do dontspy.eu/x/74

As EDRi summarizes in this slide, the situation doesn't look good for civil society needs

The troubling consequence is that AI companies are presenting themselves as "the only ones who can certify accurate information," while at the same time enabling the worst kind of automated garbage.

Read between the lines: blogs.microsoft.com/on-the-issue...