On my way to the Magic of Open Source with @apell and @OSBA at the European Digital Summit in Berlin today- let’s talk about IT Excellence with Open Source! #linux #opensource #osaid #digitalsovereignty
#OSAID
@amszmidt@mastodon.social
The way #LLM are voiding classic #copyleft has been experimentally demonstrated years ago, when #GitHub #Copilot was caught distributing well known #AGPLv3 code with a permissive license and wrong author attribution.
Also, technically that incident demonstrated how Copilot's model itself is a derivative work of copylefted works: if a lossy compression of copyrighted material is still subject to authors' #copyright, encoding such compression as arrays of floats that can be executed by virtual processors with a dedicated architecture (so called "inference" engines) does't change its nature of derivative work. Similarly violating copyrights of millions of authors at once doesn't free you of such rights.
That's basically why #OpenAI and friends are so scared by current lack of sustainable business models for their #LLM: they need money to keep Judges away.
But anyway I still have to find a single person that debate with technical competence and in good faith the derivative nature of LLMs from the text corpora compressed in their models.
As for the Hacking License not being a Free Software license, it's debatable after a careful read since the only thing you cannot do with the software is to prevent others from enjoining the same freedom it grants you.
Yet I've never claimed it is Free Software because, sadly, I'm forced to move beyond Free Software by its own limit.
OTOH I'm proud that it's not an #OpenSource license as I'll never submit it to #OSI corrupted #gatekeepers.¹
As for #FLOSS, it's a term designed to confuse free software values with corporate propaganda while marginalizing hackers: its a leaking abstraction designed to fool developers and exploit their naive groupthinking. Having been fooled myself. Never again. 😉
____
¹ The way OSI tried to #openwash the #OSAID, with an over complicated process that doubled #Meta's lobbyists' votes to exclude training data from the requirements confirmed my opinion about them.
@doctormo@floss.social
The way #LLM are voiding classic #copyleft has been experimentally demonstrated years ago, when #GitHub #Copilot was caught distributing well known #AGPLv3 code with a permissive license and wrong author attribution.
Also, technically that incident demonstrated how Copilot's model itself is a derivative work of copylefted works: if a lossy compression of copyrighted material is still subject to authors' #copyright, encoding such compression as arrays of floats that can be executed by virtual processors with a dedicated architecture (so called "inference" engines) does't change its nature of derivative work. Similarly violating copyrights of millions of authors at once doesn't free you of such rights.
That's basically why #OpenAI and friends are so scared by current lack of sustainable business models for their #LLM: they need money to keep Judges away.
But anyway I still have to find a single person that debate with technical competence and in good faith the derivative nature of LLMs from the text corpora compressed in their models.
As for the Hacking License not being a Free Software license, it's debatable after a careful read since the only thing you cannot do with the software is to prevent others from enjoining the same freedom it grants you.
Yet I've never claimed it is Free Software because, sadly, I'm forced to move beyond Free Software by its own limit.
OTOH I'm proud that it's not an #OpenSource license as I'll never submit it to #OSI corrupted #gatekeepers.¹
As for #FLOSS, it's a term designed to confuse free software values with corporate propaganda while marginalizing hackers: its a leaking abstraction designed to fool developers and exploit their naive groupthinking. Having been fooled myself. Never again. 😉
____
¹ The way OSI tried to #openwash the #OSAID, with an over complicated process that doubled #Meta's lobbyists' votes to exclude training data from the requirements confirmed my opinion about them.
@doctormo@floss.social
The #OSI wants to know from people that we hate #AI. Let’s write in to them (Attention, uses a Google Docs form): https://opensource.org/member-survey
« Départ du directeur exécutif de l'Open Source Initiative - les conséquences sur le débat sur l'OSAID » - via la revue de presse @aprilorg
I’ve noticed at least one app installed via F-Droid that asked me to opt-in to on-device “AI”. Which made me curious about whether the F-Droid team has formulated a policy on this.
I scratched this brain itch here, please consider weighing in!
https://forum.f-droid.org/t/f-droid-policy-on-libre-ai/33279
Coming back to the 2025 OSI roadmap ...
" We’ve heard the concerns raised by different stakeholders around the Open Source AI Definition v1.0
...
By the end of 2025, we’ll provide clear, actionable recommendations for the OSI board to develop an updated version of the Open Source AI Definition. You can rest assured that we are working diligently to keep the Open Source relevant."
#StefanoMaffulli, 2025
WTF is this supposed to mean?
Re: a #Mastodon poll, @cwebber wrote:
> “Can't change a poll in progress, that would be election tampering!”
Given that #OpenSource Initiative endorses election tampering, surely it's ok in #FOSS for all purposes — #OSAID related or not.
https://lwn.net/Articles/1014603/
https://codeberg.org/OSI-Concerns/election-results-2025
BTW, I'm amazed #OSI tampered w/ an election when the person w/ the longest political memory in FOSS was a candidate. I am — if nothing else — the most persistent person in #FreeSoftware. ☺
Cc: @richardfontana
🔍 Etikettenschwindel? Viele KI-Modelle werden als „Open Source“ beworben. Aber wann ist eine KI wirklich „open“?
✅ Die @osi hat sich mit dieser Frage beschäftigt und eine Open Source AI Definition (OSAID) entwickelt, die wir unterstützen.
📂 Ein zentraler Punkt: Zugänglichkeit der Trainingsdaten. Denn nur wenn nachvollziehbar ist, wie ein Modell gelernt hat, kann von echter Offenheit die Rede sein.
👉 https://osb-alliance.de/featured/osba-osaid
#OpenSource #OpenSourceAI #OSAID #DigitaleSouveränität #KI #OpenData
I've had concerns for a while about the direction the #OSI has been going with AI for a while. The publishing of the #OSAID weakens the #OpenSourceDefinition.
I'm not a member but the definition does require community participation in terms of the licences we choose and how we advocate for Open Source.
The recent board elections showed a disturbing pattern of obfuscation. We need to be able to trust Organisations that are meant to safeguard what Open source is.
Delighted to see MEP's taking up the fight against #Meta's #Llama #OpenWashing, and referencing the #OSAID as a potential solution.
The #OSI will keep fighting Open washing in Europe and around the world! 💪 🇪🇺
https://www.euractiv.com/section/tech/news/club-of-meps-demand-strict-ai-act-open-source-definition/
@johnmark what do you expect when you ask for action? OSI has pushed out OSAID - which demands full transparency- but is still met with the almost impossible demand for only free data behind AI models. While Red Gat execs now call Llama 4 OSS which is obviously wrong. IMHO we must opposed open washing and embrace #osaid now
Interesting to see #OSI manufacture a problem about people reaching out to others, saying they should rather do this "on our forums where we provide community management and moderation support" https://lwn.net/SubscriberLink/1014603/c791efdd5a4ca47c/
Last I checked (during the #OSAID snafu Sep/Oct 2024), the OSI forums were so tightly moderated that no sane and open discussion could take place anymore. The crackdown was very effective in letting reform-minded community members know their contributions weren't welcome
1/2
The #OSI 's one job is to protect and promote #OpenSource and the Open Source Definition (OSD). Undermining that definition with the weakened #OSAID has alienated much of the Open Source community, myself included.
I had hoped that might be turned around, but this year's reform candidates were likely the last chance to see that happen.
Watching the OSI promoting corporate sponsors was concerning, but this year's blatant election manipulation is the final straw.
They say to never attribute to malice that which can be explained by incompetence.
Unfortunately this year's @osi board elections go way beyond that point.
The current Executive Director is leading #OSI in a direction the community don't want to follow with #OSAID. By changing the rules as you go along in this year's election to exclude candidates who want to reform the OSI you have destroyed any remaining trust.
You shat the bed. It's time to change the sheets.
@ramin_hal9001@fe.disroot.org
I think #RMS was (is?) unable to really abstract away from his roots as an American grow up during #ColdWar.
If he was, he would have been more explicit in looking for a syntesis between #freedom and #communion as foundational model for #FreeSoftware.
I think I got a better ethical model with the #HackingLicense: https://encrypted.tesio.it/documents/HACK.txt
There the two values that dominated 20th century history (and rhetorics) are pillars built on top of a third one, #hacker's #curiosity that provides a dynamic balance among the two (and a third one, candor, intended as intellectual honesty, that is honestly sharing your perspective and knowledge).
So while a think that a second fork from free software is actually needed (the first being #opensource, that as #OSAID made cristal clear, was solely a corporate tactic to exploit #hackers' labour while marginalizing them), I don't think it can be solely based on either freedom (like liberalism) or communion (like communism and socialism) but it need a new set of values to build upon.
Unfortunately, largely due to influence of the US rhetorics since the Marshal Plan, most hackers lack political awareness.
Which is a common issue among most oppressed people these days, but it's something even more dangerous when their skills are used to build weapons (including surveillance and manipulation tools) that furter help oppressors to oppress them and everybody else.
Yet I'm afraid the international socialist movement is completely missing all of these cultural, political and cybernetic dynamics and it's unlikely to evolve beyond 1900, neither in culture nor in methods or tech.
In Italy, for example, all largest unions, #CGIL, #CISL and #UIL, are completely under US BigTech surveillance.
@garbados@friend.camp
I think #RMS was (is?) unable to really abstract away from his roots as an American grow up during #ColdWar.
If he was, he would have been more explicit in looking for a syntesis between #freedom and #communion as foundational model for #FreeSoftware.
I think I got a better ethical model with the #HackingLicense: https://encrypted.tesio.it/documents/HACK.txt
There the two values that dominated 20th century history (and rhetorics) are pillars built on top of a third one, #hacker's #curiosity that provides a dynamic balance among the two (and a third one, candor, intended as intellectual honesty, that is honestly sharing your perspective and knowledge).
So while a think that a second fork from free software is actually needed (the first being #opensource, that as #OSAID made cristal clear, was solely a corporate tactic to exploit #hackers' labour while marginalizing them), I don't think it can be solely based on either freedom (like liberalism) or communion (like communism and socialism) but it need a new set of values to build upon.
Unfortunately, largely due to influence of the US rhetorics since the Marshal Plan, most hackers lack political awareness.
Which is a common issue among most oppressed people these days, but it's something even more dangerous when their skills are used to build weapons (including surveillance and manipulation tools) that furter help oppressors to oppress them and everybody else.
Yet I'm afraid the international socialist movement is completely missing all of these cultural, political and cybernetic dynamics and it's unlikely to evolve beyond 1900, neither in culture nor in methods or tech.
In Italy, for example, all largest unions, #CGIL, #CISL and #UIL, are completely under US BigTech surveillance.
@garbados@friend.camp
Interesting. How does #OpenSource Initiative gain standing to bring a false advertising claim? Is there an article on #OSI's site that explains this enforcement mechanism & plan in detail? (And apologies I missed it until now!)
& is it on the #OSAID roadmap to bring false advertising claims against companies who have products they're calling "open source AI"?
If it is in the roadmap, I feel the community should have been told that during OSAID drafting.
Cc: @richardfontana
:opensource: Interesting platform from @bkuhn & @richardfontana for the upcoming Open Source Initiative board election:
1️⃣ Repeal the Open Source AI Definition (OSAID)
2️⃣ Adopt a process for formal review of previously approved licenses
3️⃣ Remove “code of silence” from Board Member Agreement
4️⃣ Directors should be allowed to use FOSS for Board activities
:opensourceparrot: This would have my vote. What about you?
In any case, WTF is a California-based organisation like @osi doing with a submarine GMT deadline? There's still absolutely no justification for this, except to censor the most important channel of all: elections.
Were I eligible to run I'd have left it to the last minute too, to incorporate all available information like the #FOSDEM revelation that they consider #OSAID 1.0 "too restrictive" and want more corporations for 2.0 which will water #OpenSource #AI down EVEN MORE.
In any process where you are talking about "stakeholders", if any of your stakeholders are not individual human beings, you need to think deeper.
A corporation is never a stakeholder.
The only reason to care about a corporation or other org is because of the real people that it might affect. The employees, perhaps. The customers, sometimes. *Those* are your stakeholders.
If one of your "stakeholders" is a trillion dollar corporation, you really are on the wrong path.
If you're "late to the game" regarding the #OSAID #OpenSource #AI definition from @osi and what is the criticism to it all about, this blog post from @sj, while specifically directed to @mozillaofficial, has a good enough amount of linkdump to be clarifying:
Client Info
Server: https://mastodon.social
Version: 2025.07
Repository: https://github.com/cyevgeniy/lmst