mastodon uno admin 
LATINA: AL VIA IL PROCESSO SATNAM SINGH, IL PADRONE È ACCUSATO DI OMICIDIO VOLONTARIO https://www.radiondadurto.org/2025/04/01/latina-al-via-il-processo-satnam-singh-il-padrone-e-accusato-di-omicidio-volontario/ #AntonelloLovato #Immigrazione #immigrazione #sfruttamento #SatnamSingh #NAZIONALI #Cisterna #omicidio #processo #Diritti #Lavoro #Latina #lavoro #CGIL #FLAI
Ricerche recenti
Opzioni di ricerca
Amministrato da:
#flai
0 post0 partecipanti0 post oggi
La funzione originaria della fotografia.
#Lodi #Torbidi #cgil #flai
https://www.flai.it/primo-piano/focus/il-sindacalista-fotografo-davide-torbidi-racconto-il-fondo-in-cui-nicola-nappo-e-ancora-vivo/
Ha risposto nella discussione
@tante I'm quite confident that free (or licensed) training data and resulting AI models will appear in the future (they're called "vegan models" at https://simonwillison.net/2024/Jan/25/fairly-trained-launches-certification-for-generative-ai-models-t/).
These "free/libre" models will likely lag behind proprietary models (just like Linux lags behind Windows? 
). But FLAI will come. And maybe the OSI will publish a better license then; or else someone else will.
(2/2)
simonwillison.netFairly Trained launches certification for generative AI models that respect creators’ rightsI've been using the term "vegan models" for a while to describe machine learning models that have been trained in a way that avoids using unlicensed, copyrighted data. Fairly Trained …
@tante I kinda disagree with the conclusion.
Yes, this OSI AI license is weak. But that doesn't mean we can't have a better license in the future, for actual "FLAI" (akin to "FLOSS").
I think the article should give some citations for the assertion that "Because the data sets are too big to do actual filtering and ensuring their legality. It’s just too much.", which much of the remaining article builds upon.
(1/2)