@quinta I'm not convinced about the "Learning" part.
Machines don't learn shit. As they obvious don't have any kind of understanding of they're doing, or any kind of "thinking" process. They just produce statistically plausible-sounding bullshit. Which is… well, mostly (by far) still bullshit.
It's more like "Planet-burning machines for profit" rather than Machine learning or learning algos. Or anything related to "learning" or to "Intelligence".
Framamia (🎶 here we go agaiiin 🎶), Framasoft met les pieds dans l'#IA
#FramamIA est un site-outil qui permet de partager définitions, enjeux, veille et questionnements autour de l'Intelligence Artificielle.
🤔 Présentation sur le blog : https://framablog.org/?p=36102
🤩 Découvrir : https://framamia.org
Si nous avons la liberté d'expérimenter autour de Framamia (et de Lokas #spoileralert), c'est grâce àr vos dons. Contribuez à cet élan de solidarité en allant sur https://soutenir.framasoft.org
Not a traditional license, but this project has explicit political goals: https://code.tesio.it/p/self-hosting/doh.cgi/file?name=POLITICS.txt&ci=tip
It doesn't need to enforce them: anyone using the software structurally contribute to the goal unless you run it on #BigTech servers (but that would be a pointless waste of money).
The #HackingLicense it adopts has a few political goals too, but more subtle.
@Shamar sì sì, problema politico immenso. Immenso.
E immensamente sottovalutato dall'opinione pubblica.
Come quello dell'infiltrazione palese, evidente, oserei dire trasparente, della lobby BigTech dentro TUTTI i partiti del Parlamento.
Ovviamente è diverso quando a infiltrare e condizionare la politica di un paese è un tuo "alleato" o un paese "non amico". Ma il problema resta comunque grave
💥 Celebrating a Decade of Devuan: The Fight for Freedom Continues! 💥
A decade ago, we embarked on an audacious journey, resolved to “not go gentle into that good night” and fiercely defend our freedom of choice. Today, the reality is stark: the Linux landscape has been increasingly commercialized, compromising its core values and spirit.
6am, I just checked my email and found this gem:
« I remember Elena visited my school in Los Angeles over 5 years ago, and presented her documentary, The Illusionists. The viewing inspired me to take a gender studies class, and ultimately reframed how I consume media and see the beauty industry… »
No need to address me in the 3rd person dear student, I don’t have a "team". Just a person hoping to make a small difference. I’m sitting in bed with a happy tear streaming down my face 🥲
"We would like to add our voices, particularly in strong support of the requirement for data, which we consider to be the ‘source’ of AI systems."
Nop, I'll add an example in the following days... :-)
As for the CGI+Crontab version, actually it's not that hard to hack something together with minimal changes to existing code and few new source files, but I'm facing a couple of issues that suggests a proper fork:
1. the CGI ends up being larger than 6Mb because of curl & openssl that, as far as I can tell, are not needed because all the CGI has to do is to save the proper queue item on the filesystem
2. no command handle the global queue alone
The second issue is quite easy to solve with minimal changes and could help everybody.
The first however requires a lot of careful reorganization of the code.
On the other hand, I don't like the idea of a fork diverging too much, first because you are doing a great work here so much I think #snac deserves much more visibility.
And yet, technically speaking, a fork would make much sense to keep both code bases clean and focused to their use case.
Did you see this PR?
https://codeberg.org/grunfink/snac2/pulls/229
On top of that I'm also trying to hack a CGI+Crontab version of snac, but it's not ready yet.
Completeness vs Openness: Data Dependencies and Open Source AI
https://samjohnston.org/2024/11/24/openness-vs-completeness-data-dependencies-and-open-source-ai/
#OpenSource #ArtificialIntelligence #Open #AI #OpenAI #OpenData