@ed@opensource.org
@ed@opensource.org avatar

ed

@ed@opensource.org

Executive Director - Open Source Initiative https://social.opensource.org/@osi

This profile is from a federated server and may be incomplete. Browse more on the original instance.

osi, to opensource
@osi@opensource.org avatar

Join @ed to hear the latest news and comment on the drafts of the Definition, the 10th public town hall meeting, today in ~2hours (17:00 CEST)

https://opensource.org/events/open-source-ai-definition-town-hall-may-31

ed,
@ed@opensource.org avatar

@osi we're live now https://osi.gl.rna1.blindsidenetworks.com/rooms/osi-grp-8m9-yi6/join

ed, to random
@ed@opensource.org avatar

Goodbye Pittsburgh, I discovered a beautiful city at #pyconUS

ramsey, to random
@ramsey@phpc.social avatar

@osi I sent an email to the license-discuss list (I’m a member of the list), but I can’t tell whether the list actually got it. It doesn’t show up in the archives (nothing from May is in the archives, but maybe no mail has been sent this month?): https://lists.opensource.org/pipermail/license-discuss_lists.opensource.org/

Who should I contact to find out if the message was actually received? Maybe there’s moderation happening, and no one has approved it yet?

ed,
@ed@opensource.org avatar

@ramsey @osi this is weird. can you please email the list owner and list admin?

ed, to opensource
@ed@opensource.org avatar

Reminder: Open Source AI Definition workshop at PyCon US https://opensource.org/events/open-source-ai-definition-workshop-at-pycon-us #OpenSource #PyConUS

ed, to opensource
@ed@opensource.org avatar

The goal is to present a stable version of the definition in October at the All Things Open event in Raleigh, North Carolina. https://opensource.org/blog/the-open-source-ai-definition-gets-closer-to-reality-with-a-global-workshop-series

ed, to opensource
@ed@opensource.org avatar

The Open Source Initiative (OSI) invites the Python community to participate in the global conversation to find the definition of Open Source AI. https://opensource.org/events/open-source-ai-definition-workshop-at-pycon-us #OpenSource #PyConUS

ed, to random
@ed@opensource.org avatar

If you run a business producing Open Source products or your company’s revenue depends on Open Source in any way, we want to hear your insights https://opensource.org/blog/practical-open-source-2024-submit-your-proposal

ed, to opensource
@ed@opensource.org avatar

“The work programme shall also include the actions to be undertaken to ensure effective participation of relevant stakeholders, such as small and medium enterprises and civil society organizations, including specifically the open source community where relevant” https://opensource.org/blog/cra-standards-request-draft-published #OpenSource

luis_in_brief, to random
@luis_in_brief@social.coop avatar

I’ll be there.
https://mastodon.social/@avilarenata/112370958652498664

ed,
@ed@opensource.org avatar

@luis_in_brief I'm waiting to hear back from the conference organizers: if my talk is accepted, I should be there too

KathyReid, to random
@KathyReid@aus.social avatar

My Google Pixel 4a 5G died this afternoon and it won't turn on - I am trying all the rebooting / forced restarting options, but nothing is working so far.

The key lesson I am learning is how dependent I am on everything on my phone - my music is on my phone, audio books are on my phone, is on my phone, entertainment in the form of games are on my phone.

I knew I was dependent, but not just how dependent I was.

ed,
@ed@opensource.org avatar

@KathyReid that's exactly what happened to mine too. Extremely annoying and time consuming to rebuild the banks 2fa stupid systems they started using

informapirata, to aitech Italian
@informapirata@mastodon.uno avatar

È già finita l’era dell’intelligenza artificiale open source?

Le aziende vogliono lavorare su modelli addestrati da altri per sviluppare soluzioni chiuse, un po' perché non vogliono che gli oligopolisti si impadroniscano dei propri dati privati, un po' per ragioni di profitto. Ma nessuna azienda ha addestrato i propri modelli su dati prodotti internamente.

(grazie a @noccioletta per la segnalazione)

@aitech

https://www.internazionale.it/notizie/alberto-puliafito/2024/03/31/intelligenza-artificiale-open-source

ed,
@ed@opensource.org avatar

@informapirata grazie della segnalazione. Mi sembra un'opinione basata su un precontratto sul significato di Open Source applicato alle IA che non è allineato con le bozze della definizione. Conosci il giornalista? Ci vorrei parlare
@aitech

ed,
@ed@opensource.org avatar

@informapirata non ci ho parlato , ha citato un commento da un'intervista a Pascale di Euro News, ripresa da Yahoo news. Lo cerco su LinkedIn, grazie @aitech

ed, to opensource
@ed@opensource.org avatar

Seeking document reviewers for Pythia and OpenCV #OpenSource #ai
https://discuss.opensource.org/t/seeking-document-reviewers-for-pythia-and-opencv/279

ed, to opensource
@ed@opensource.org avatar

We are entering the last 2 months of drafting: this is the time to raise concerns and shape the final text of the Open Source AI Definition. https://opensource.org/blog/open-source-ai-definition-weekly-update-mar-25 #OpenSource

ed, to opensource
@ed@opensource.org avatar

A very good summary of the amount of work that @osi has taken on https://www.technologyreview.com/2024/03/25/1090111/tech-industry-open-source-ai-definition-problem/
#OpenSource #ai

ed, to opensource
@ed@opensource.org avatar

The town hall meeting with the latest updates on the #OpenSource #AI is starting now: join it live https://osi.gl.rna1.blindsidenetworks.com/rooms/osi-grp-8m9-yi6/join or watch the recordings on discuss.opensource.org

ed, to opensource
@ed@opensource.org avatar

This is the kind of value that greedy companies take from #OpenSource and want you forget/dismiss as you read their tragicomical announcements of a license change

RE: https://fosstodon.org/

simon, to random
@simon@simonwillison.net avatar

This looks like a big deal: https://huggingface.co/blog/Pclanglais/common-corpus - a 500 billion word training corpus of entirely public domain data designed for training LLMs, gathered with a "lengthy process of copyright duration verification"

I wrote a few more notes here: https://simonwillison.net/2024/Mar/20/releasing-common-corpus/

ed,
@ed@opensource.org avatar

@ramsey @simon I think we should evaluate a different approach: don't expand the scope of copyright, allow text&data mining for all but write rules that force trained models to be public, if trained on copyrighted material. That's more appealing to me than giving more power and technical advantage to the giant corporations who will enter into agreements with giant data aggregators.

ed,
@ed@opensource.org avatar

@ramsey @simon the gist is that the model weights must be publicly available, in exchange for freedom to data mining. My fear is that large corporations will license data anyway, artists will continue to starve as they do now, despite DRMs, DMCA and similar draconian laws, only platforms would thrive in a copyright expansion.

ramsey, to random
@ramsey@phpc.social avatar

@ed Just read the OSI email about draft 0.6 of the Open Source AI definition, and noted this statement: “working groups analyzed four machine learning systems and established a list of default required components to make modifications to an AI system, specifically indicating that the original training datasets are not a required componen.”

What is the rationale for why the original training datasets are not required for an AI system to be considered open source?

ed,
@ed@opensource.org avatar

@ramsey there is a large thread on the forum on the data topic :) there are arguments for both inclusion and exclusion of training datasets from the default requirements. Ultimately the votes of the working groups didn't signal a strong preference to include them.

ed,
@ed@opensource.org avatar

@Crell you're making an assumption that's yet to be proven correct in the US, let alone proven in other legislations. Please follow and comment on the forum thread, this issue is too complex to deal with 200characters on a fast and furious social medium.

ed, to random
@ed@opensource.org avatar

A little known fact is that the replacement of GNU Coreutils are not using the copyleft licenses GPL and LGPL but they're released with non-reciprocal licenses.

I think it's unfortunate, losing copyleft code is a bad thing for society. But I don't see how this tide can turn quickly.

https://mastodon.social/@sovtechfund/112036923555408749

ed, to opensource
@ed@opensource.org avatar

Central question: Should the definition of Open Source AI take a gradient approach (such as the case with RAIL licence), judging the openness of the components of a model, rather than the whole of it? https://opensource.org/blog/open-source-ai-definition-weekly-update #OpenSource

ed, to ai
@ed@opensource.org avatar

Deep Dive #AI town hall meeting is about to start: join it to ask questions and hear the latest on the Open Source AI Definition process
https://connect.rna1.blindsidenetworks.com/invite/to?c=7wDx-PWaaPEaEvOEwckt7UE6-kSwEf-d6BNeS3zf7QY&m=77qwg7ojv5p8espbeommqnwl7mlqu8ytqi19z3s8&t=1708674442419&u=osi

ed, to opensource
@ed@opensource.org avatar

Gemma is not !

ed,
@ed@opensource.org avatar

@daridrea indeed it is. And given the US policy landscape, it may shine light on a possible path for some sort of "Open" to even exist in AI... It's hard for people not in the weeds of DC to realize how grim the situation is.

  • All
  • Subscribed
  • Moderated
  • Favorites
  • •
  • JUstTest
  • mdbf
  • everett
  • osvaldo12
  • magazineikmin
  • thenastyranch
  • rosin
  • normalnudes
  • Youngstown
  • Durango
  • slotface
  • ngwrru68w68
  • kavyap
  • DreamBathrooms
  • tester
  • InstantRegret
  • ethstaker
  • GTA5RPClips
  • tacticalgear
  • Leos
  • anitta
  • modclub
  • khanakhh
  • cubers
  • cisconetworking
  • megavids
  • provamag3
  • lostlight
  • All magazines
    •