Schneier joins the criticism of "Open Source" AI:
Hello dear K-9 users!
You probably got a suggested updated for K-9 to "Thunderbird Beta for Testers".
This is just a metadata issue, which displays a wrong name and icon. The update itself is K-9 in version 8.0. You don't get automatically switched over to Thunderbird.
This will probably be fixed in the next index update.
In other news: Thunderbird for Android is available now on #FDroid (as a new separate app) ;)
More people in the Open Source community ought to read this: https://archive.is/paD1W
h/t @Shamar (via LWN)
tl;dr: OSI’s behaviour is even worse than we thought
(And apparently Fakebook was one of the driving forces behind this… and still there are idiots on Fedi who want to connect with them. Bah!)
The #OSAID is deeply flawed exactly because of two #Meta employees (a Llama developer and a lawyer) that were invited to vote about the requirements and excluded training data.
Without training data, you cannot really modify any AI, just fine-tune them. Which is like saying that Windows is #OpenSource because you can tweak the registry.
And without training data you have no hope of being able to effectively study its bias or to identify backdoors planted in the system: https://arxiv.org/abs/2204.06974
#OSI is serving its sponsors, #Google, #Meta, #Microsoft... opening a loophole in the #AIAct.
Ai miei occhi serve anzitutto per evitare che un giudice possa prendere sul serio la OSAID.
Per altri del gruppo serve a evitare che la OSAID prenda il posto della OSD nella pratica: l'OSI è un gatekeeper per le licenze "open source" che si è rivelato pericoloso.
A proposito: hai visto https://opensourcedeclaration.org/index-it-it.html
Community Commitment to Open Source Definition
I remember Google+, and the idee fixe and mad hype around it. (Google was afraid of Facebook.) G+ was a ghost town. But for the first year or so the G+ team reported astronomical engagement numbers. Huhh?
Finally we learned they were counting every G+ notification dropped at the top of Gmail as an "engagement."
Anyway, I wonder how they're measuring this...
https://www.businessinsider.com/google-earnings-q3-2024-new-code-created-by-ai-2024-10
#Debian classify the kind of #AI systems that the #OSI's #OSAID defines as #opensource, as #ToxicCandy: https://salsa.debian.org/deeplearning-team/ml-policy/-/blob/master/ML-Policy.rst
Without training data you cannot ecercise the freedom to study a #ML system and your ability to modify it is severely limited to fine tuning.
Which is like to say that windows is open source because you can tweak the registry.
For these reasons some open source developers are already moving beyond OSI: https://opensourcedeclaration.org
You are welcome to join!
It's not just matter of securing the #OSD, but to update it with a truly open process, with all developers, artists, musicians, data scientists... who contribute their time and valuable skills to open source: https://opensourcedefinition.org/wip/
#Debian classify the kind of #AI systems that the #OSI's #OSAID defines as #opensource, as #ToxicCandy: https://salsa.debian.org/deeplearning-team/ml-policy/-/blob/master/ML-Policy.rst
Without training data you cannot ecercise the freedom to study a #ML system and your ability to modify it is severely limited to fine tuning.
Which is like to say that windows is open source because you can tweak the registry.
For these reasons some open source developers are already moving beyond OSI: https://opensourcedeclaration.org
You are welcome to join!
It's not just matter of securing the #OSD, but to update it with a truly open process, with all developers, artists, musicians, data scientists... who contribute their time and valuable skills to open source: https://opensourcedefinition.org/wip/
Interesting take I almost agree with.
For sure Zuboff is careful to avoid structural questions. For sure she reads the whole thing within an outdated economical framework that cannot describe a cybernetic society.
For sure she focus on surveillance overlooking individuals' manipulation on large scale.
But the statistical manipulation on a large scale is not a "false assumption" but an evident phenomen.
And yes, I agree that we are talking about an emergent feature of Capitalism that, as Marx wrote, is an ever changing moving target.
BUT we should not repeat her error and limit our interpretation of this new form of oppression within a social-economical framework ideated to describe industrial economy in '900.
We need a cybernetic model of society, if we want a chance to conceive a just political theory to fight for.
"Still, it’s unclear whether “Just Go Independent” is a sustainable career path for the number of journalists who we need to have a functioning society. But I do know that relying on the passing interest of billionaires to keep journalism alive is not sustainable. And I know that 250,000 subscribers could fund a lot of independent journalists."
(Original title: The Billionaire Is the Threat, Not the Solution)
https://www.404media.co/the-billionaire-is-the-threat-not-the-solution/
Totally agree.
It's time to leave #OSI behind.
https://osd.fyi
And if you think we need to address #AI and #opensource, let's discuss such change openly in the open https://opensourcedefinition.org/wip/
They posit you can still modify (tune) the distributed models without the training source. You can also modify a binary executable without its source code. Frankly that's unacceptable if we actually care about the human beings using the software.
A key pillar of freedom as it relates to software is reproducibility. The ability to build a tool from scratch, in your own environment, with your own parameters, is absolutely indispensable to both learning how the tool works and changing the tool to better serve your needs, especially if your needs fall on the outskirts of the bell curve.
There's also the issue of auditability. If you can't run the full build process yourself, producing your own results from scratch in a trusted environment to compare with what's distributed, it becomes exponentially harder to verify any claims about how a tool supposedly works.
Without the training data, this all becomes impossible for AI models. The OSI knows this. They're choosing to ignore it for the sake of expediency for the companies paying their bills, who want to claim "open" because it sounds good while actually hiding the (largely stolen and fraudulently or non-consentually acquired) source material of their current models.
Do we want a new definition of "open source" that actively thwarts analysis and tinkering, two fundamental requirements of software that respects human beings today? Reject this nonsense.
#OpenSource #OpenSourceAI #OSI #OpenSourceInitiative #FreeSoftware #AI #GenAI #GenerativeAI