The OSI lacks competence to define Open Source AI | Hacker News Reader

The OSI lacks competence to define Open Source AI

samjohnston.org

I'd title this as "some guy's hot take: The OSI lacks competence to define Open Source AI"

That’s redundant. It’s obvious this is an opinion piece. And they’re not wrong either.

They may not be wrong with their primary point, but they neglected to provide an argument to support it. Instead of bothering to argue, they just provide a litany of freeze peach dog whistles. Anyone who doubts that ideology's value would rather have an actual argument laid out.

Perhaps, but that wasn’t the criticism laid out by the GP. In fact they didn’t really have an argument at all, they only tried to discredit the author as “some guy” and his opinion as a “hot take”. Essentially a pseudo-ad hominem to sidestep the argument entirely.

Sorry I should have said "some guy with no expertise writes blog post to claim he has more expertise than literal experts".

The "co-design" process is literally the "Do Your Own Research" of technical standards (https://jods.mitpress.mit.edu/pub/costanza-chock/release/4):

"We believe that everyone is an expert based on their own lived experience, and that we all have unique and brilliant contributions to bring to a design process."

Not that they're practicing what they preach anyway - there's only one voice that matters and that's the OSI's Executive Director (https://discuss.opensource.org/t/proposal-to-achieve-consens...):

"We see the role of the designer as a facilitator rather than an expert."

Your 2nd link does not lead to a page where I could find the sentence "We see the role of the designer as a facilitator rather than an expert."

Ah yes, because bringing in experts who have allegiances to the very same organizations you wish to legislate has worked amazingly so far. And again, you’re utterly failing to provide an actual argument that properly references anything within the article. Instead you continue to sidestep the issue by attacking his credibility and not his points. Perhaps that’s because you don’t have a proper refutation that the OSI are woefully incompetent and are not suited to be in charge of the future of OS AI?

Not only does he lack the expertise, he's building an AI using llama and marketing it as "Open Source"

The guy has an interest in the definition failing. If it doesn't he won't be able to market his AI as Open Source.

/me waves

One of the sections ("Blind Leading the Blind") is about the blog post author asking an AI whether the OSI board had expertise in AI , and him being shocked by the answer that they do not.

In all fairness, asking the same AI the same question right now produces:

-----

Based on the search results provided, there is no explicit mention of any Open Source Initiative (OSI) team or board members claiming expertise in Artificial Intelligence. However, there are a few relevant points to note:

1. The OSI has been actively involved in initiatives related to AI, particularly in developing an Open Source AI Definition (OSAID)[1][3]. This suggests that the organization is engaging with AI-related topics, even if individual expertise is not specified.

2. The OSI board includes members with diverse backgrounds, and they are working with external experts on AI-related projects. For example, Sayeed Choudhury, an OSI board member, is leading the Open Forum for AI (OFAI) initiative at Carnegie Mellon University[1].

3. The OSI has collaborated with various stakeholders, including AI experts from organizations like Mozilla Foundation, Google, Microsoft, Hugging Face, and others, in the process of developing the Open Source AI Definition[3]. This indicates that while OSI may not have internal AI experts, they are actively engaging with AI professionals from other organizations.

4. The OSI has organized events, webinars, and podcasts featuring experts discussing AI-related topics[3]. This suggests that the organization is committed to understanding and addressing AI issues, even if they may not have in-house AI expertise.

While the OSI is clearly involved in AI-related initiatives, the search results do not provide specific information about individual OSI team or board members claiming expertise in Artificial Intelligence. The organization appears to be leveraging external expertise and collaborations to address AI-related challenges in the context of open source software.

Citations: [1] https://opensource.org/blog/the-open-source-initiative-joins... [2] https://opensource.com/article/23/3/questions-osi-board-cand... [3] https://opensource.org/deepdive [4] https://www.osler.com/en/insights/updates/osi-releases-first... [5] https://www.europarl.europa.eu/RegData/etudes/STUD/2021/6629... [6] https://www.sfi.ie/research-news/stories/ai/ [7] https://www.linkedin.com/posts/osidigital_the-osi-digital-te... [8] https://www.youtube.com/watch?v=_l_BREt8Afk

-----

I, for one, have no issue with the board of an organization getting external expertise for topics that they believe such external input is needed.

Normally yes, but said board is taking a very liberal view of "experts"; the "co-design" process literally says "everyone is an expert" (https://jods.mitpress.mit.edu/pub/costanza-chock/release/4):

"We believe that everyone is an expert based on their own lived experience, and that we all have unique and brilliant contributions to bring to a design process."

The result is a proposed standard that does not do its one job: protect the four freedoms that define Open Source (and Free Software in general).

Reply on news.ycombinator.com