Recent searches
Search options
Administered by:
In the #OpenSource Initiative Board Elections, many are confused about the 2 different “constituencies”. It's like Congressional districts: I'm in the “Affiliate” district & @richardfontana is in the “Member” district.
Each has unique challenges. My challenge is getting attention of very busy #FOSS org leaders to vote. If you have contacts at these orgs, please contact them. Ask them to rank me first & to *not* rank the incumbent, Carlo Piana. It's time for a change.
I offer my time for BBB chats, phone calls, email exchanges, or public discussion here with any Affiliate organization of the #OpenSource Initiative.
While @richardfontana & I have our Shared Platform…
https://codeberg.org/OSI-Reform-Platform/platform#readme
… we purposely kept it narrow so that we remain flexible to serve needs of our constituents that are unrelated to those urgent issues in our platform.
If you are involved w/ any #OSI Affiliate…
https://opensource.org/affiliates
… I'd be glad to talk with you & hear your concerns.
@bkuhn @richardfontana i was going to try to meet up with y'all at FOSDEM and ask how i can help here
i am very concerned about OSI in its current state, and while i have mixed opinions on OSAID (i think it helped to stop AI companies from calling their proprietary model weights "open source"), i don't think it has done anything to meaningfully advance software freedom in regard to model weights/model reproduction
@ariadne @bkuhn @richardfontana one problem is that in many cases basically nobody can rerun the training because of costs.
@alwayscurious @bkuhn @richardfontana yes, but that's not an excuse to call something that lacks the necessary data to reproduce it "open source"
even if it costs $$$$$$$$$$ to reproduce, if it is theoretically possible that is still a significant improvement over the current status quo
@alwayscurious @bkuhn @richardfontana this is basically the position a lot of us took when we challenged stefano on his efforts. this is why there is a "good, better, best" model in the OSAID now.
stefano was solving for a different problem than freedom, arguably: the OSAID efforts were an attempt to block Meta from calling LLaMA "open source" when it clearly has terms that are not compatible with software freedom. what i would like to see is an effort that is targeted at blocking openwashing while also protecting software freedom more robustly than that "good, better, best" compromise we arrived at.
@ariadne @alwayscurious @richardfontana
If goal of OSAID was merely to block others from using the term #OpenSource to refer to their so-called “AI” products under moniker “open source”, that is was known to be unachievable.
“Open source” was likely already a generic term before OSI existed, & they never even tried to trademark the term.
Anyone can call anything they want by that moniker & OSI can't do a thing about it.
Same is true with the OSD: lots of stuff called “open source” isn't.
@bkuhn @alwayscurious @richardfontana it is not legally possible to block improper use of the term, but it is possible to make their use of it indefensible from a PR perspective.
i believe the goal of OSAID was the latter, giving people concrete talking points about why the proprietary models are proprietary and not “open” as advertised.
the original OSD is also useful as such a criterion.
@ariadne
And *that* is a key reason why OSAID shouldn't be a Definition. The only leverage #OSI *ever* had to convince the world #OpenSource means the license met the OSD was goodwill & organizational credibility that they represented a community consensus. Regardless of whether you like the OSAID or not, there is strong disagreement in the community due to lack of wall clock time to investigate the issue fully.
OSI undermined its own credibility by manufacturing consent.
@bkuhn @richardfontana yep fully agreed on this :)
@bkuhn @ariadne @alwayscurious @richardfontana Bradley, you're in the wrong realm. It's false advertising.
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
@bkuhn @pchestek @richardfontana i am not aware of OSAID being anything other than a PR campaign 
(and a very loose “definition” i guess)
As a general policy matter in tech, we should be extremely skeptical when Big Tech tells us: “you can't learn or verify what this does. Just trust us. You'd need to be really wealthy to figure it out, anyway!”
Who is it that wants everything to have LLM-based generative models inside? I don't think the average software consumer wanted it.
The ballyhoo is akin to that with BlockChain. Turns out: Blockchain is an interesting solution to some specific problems, but not revolutionary.
@ariadne wrote:
> I was going to try to meet up with
> y'all at FOSDEM
Sadly, I got COVID-19 at #FOSDEM and missed the last day (and my own keynote!).
I noticed you're replying in this thread where I was asking to chat with Affiliate Reps who had questions (since they're my constituents). If you are, I'd like to chat offline in a more in depth conversation.
Also happy to answer your questions here in public if that suits you better, I'm just very longwinded so this format is tough for me!
@bkuhn no, i’m on the member side sadly. but i have friends who are in decision making positions at several of the affiliates. so i guess my question is, would you mind if i direct them to your post?