Benvenuto nel Poliverso | Display

mastodon - Collegamento all'originale

petersuber

1 anno fa • •

petersuber
1 anno fa • •

#Meta says that its #Llama #AI tools are "the best #OpenSource models of their class, period."

From @kylelwiggers: "There’s only one problem: the Llama…models aren’t really “open source”… Open source implies that devs can use the models how they choose…But…Meta has imposed certain licensing restrictions…Llama models can’t be used to train other models. And app developers with over 700M monthly users must request a special license from Meta."
techcrunch.com/2024/04/20/this…

#LLMs #Licensing

TechCrunch is part of the Yahoo family of brands

^{techcrunch.com}

#opensource #ai #meta #licensing #llama #LLMs @Kyle Wiggers ✔

Questa voce è stata modificata (1 anno fa)

in reply to petersuber

mastodon - Collegamento all'originale

petersuber

in reply to petersuber • 8 mesi fa • •

Update. Here's Mark #Zuckerberg defending #Meta's approach to #OpenSource #AI.
about.fb.com/news/2024/07/open…

Open Source AI Is the Path Forward

Mark Zuckerberg outlines why he believes open source AI is good for developers, Meta and the world.

^{Mark Zuckerberg, Founder and CEO (Meta)}

#opensource #ai #meta #zuckerberg #llama

in reply to petersuber

mastodon - Collegamento all'originale

THIS ACCOUNT HAS MOVED

in reply to petersuber • 8 mesi fa • •

are they releasing the data sets? Because it's all bullshit if they are "open sourcing" only the code and a trained network weight set.

See lawfaremedia.org/article/why-t…

Why the Data Ocean Is Being Sectioned Off

Bigger is better approaches in AI create an inexhaustible appetite for users’ data, leading to a rise in user data expropriation, sectioning off of the internet, and “data feudalism.”

^Default

in reply to petersuber

mastodon - Collegamento all'originale

petersuber

in reply to petersuber • 7 mesi fa • •

Update. Another #Zuckerberg defense of the #Meta approach to #OpenSource #AI, this time co-authored by Daniel #Ek, the CEO of #Spotify. Zuck and Ek use the case for OS as an argument against #EU regulations of AI.
archive.is/IybwU

#opensource #spotify #ai #eu #meta #zuckerberg #ek

Tech Cyborg reshared this.

in reply to petersuber

mastodon - Collegamento all'originale

petersuber

in reply to petersuber • 7 mesi fa • •

Update. The Open Source Initiative (@osi, #OSI) is trying to define what counts as #OpenSource #AI.
simonwillison.net/2024/Aug/27/…

"There is one very notable absence from the definition: while it requires the code and weights be released under an OSI-approved #license, the #TrainingData itself is exempt from that requirement."

Debate over “open source AI” term brings new push to formalize definition

Benj Edwards reports on the [latest draft](https://opensource.org/deepdive/drafts/open-source-ai-definition-draft-v-0-0-9) (v0.0.9) of a definition for "Open Source AI" from the [Open Source Initiative](https://opensource.org/).

^{simonwillison.net}

#opensource #ai #osi #license #TrainingData @Open Source Initiative

Tech Cyborg reshared this.

in reply to petersuber

mastodon - Collegamento all'originale

Jacob Something

in reply to petersuber • 7 mesi fa • •

That's interesting, in particular as the actual text says:

"Preferred form to make modifications to machine-learning systems: ... Data information: Sufficiently detailed information about the data used to train the system ... Data information shall be made available with licenses that comply with the Open Source Definition. ... if used, this would include the training methodologies and techniques, the training data sets used, ..."

This at least recommends sharing the data?

@Open Source Initiative

in reply to petersuber

mastodon - Collegamento all'originale

petersuber

in reply to petersuber • 7 mesi fa • •

Update. More on #OpenWashing and the complexity of defining #OpenSource #AI.
hackernoon.com/is-that-llm-act…

"The Open Source AI Definition (#OSAID) is still open for public review and feedback. If you’d like to participate in shaping the future of Open Source AI, you can submit comments."

Is that LLM Actually "Open Source"? We need to talk Open-Washing in AI Governance

In this blog, we dive deep into the complexities of AI openness, focusing on how Open Source principles apply—or fail to apply—to Large Language Models (LLMs).

^{Sal Kimmich (hackernoon.com)}

#opensource #ai #Openwashing #osaid

reshared this

in reply to petersuber

mastodon - Collegamento all'originale

petersuber

in reply to petersuber • 6 mesi fa • •

Update. More controversies around the still-evolving Open Source Initiative (@osi, #OSI) definition of #OpenSource #AI.
theregister.com/2024/09/14/opi…

Begun, the open source AI wars have

This is going to be ugly. Really ugly

^{Steven J. Vaughan-Nichols (The Register)}

#opensource #ai #osi @Open Source Initiative

Questa voce è stata modificata (6 mesi fa)

Tech Cyborg reshared this.

in reply to petersuber

mastodon - Collegamento all'originale

petersuber

in reply to petersuber • 6 mesi fa • •

Update. New study: "Our extensive experiments reveal that #OpenAccess high-performance #LLMs can be adeptly reverse-aligned to output harmful content, even in the absence of manually curated malicious datasets. Our research acts as a whistleblower for the community, emphasizing the need to pay more attention to safety of open-accessing LLMs."
aclanthology.org/2024.findings…

On the Vulnerability of Safety Alignment in Open-Access LLMs

Jingwei Yi, Rui Ye, Qisi Chen, Bin Zhu, Siheng Chen, Defu Lian, Guangzhong Sun, Xing Xie, Fangzhao Wu. Findings of the Association for Computational Linguistics ACL 2024. 2024.

^{ACL Anthology}

#openaccess #LLMs

in reply to petersuber

mastodon - Collegamento all'originale

petersuber

in reply to petersuber • 6 mesi fa • •

Update. If #OpenSource #LLMs are vulnerable to hacks that generate harmful content (this thread, prev post), they also "bring several advantages to cybersecurity systems" that can reduce those risks.
venturebeat.com/security/how-o…

#opensource #LLMs

Tech Cyborg reshared this.

in reply to petersuber

mastodon - Collegamento all'originale

petersuber

in reply to petersuber • 6 mesi fa • •

Update. HELIOS Open (@heliosopen) comments on v. 0.0.9 of the Open Source Initiative (#osi, @osi) definition of #OpenSource #AI.

"If the definition doesn’t start by emphasizing the openness of training data out of the gate, [we] worry it will not get added in later."

#opensource #ai #osi #TrainingData @Open Source Initiative @HELIOS Open

Questa voce è stata modificata (6 mesi fa)

reshared this

in reply to petersuber

mastodon - Collegamento all'originale

William Gunn

in reply to petersuber • 6 mesi fa • •

@osi Open Source has a lot of potential for good, but we have to be real careful with how we think about weights. We need to bring in some ideas from bio like gain-of-function research and dual-use tech, because just thinking in terms of software licensing leaves out the ability to make important distinctions.

For example, in bio research, #openscience means publishing #openaccess and providing #opendata, but it doesn't mean sending virus samples with dangerous mutations to anyone

#openscience #opendata #openaccess @Open Source Initiative

in reply to petersuber

mastodon - Collegamento all'originale

petersuber

in reply to petersuber • 6 mesi fa • •

Update. "#Meta has been criticised for calling its #AI models #OpenSource by the group that has spearheaded open-source technology in the software world for the past 25 years. The social media company is 'confusing' users and 'polluting' the term open-source by using it to describe its #Llama family of #LLMs, said Stefano Maffulli, head of the Open Source Initiative [#OSI, @osi]."
archive.is/N5CFG

#opensource #ai #osi #meta #llama #LLMs @Open Source Initiative

in reply to petersuber

mastodon - Collegamento all'originale

petersuber

in reply to petersuber • 6 mesi fa (Ricevuto 5 mesi fa) • •

Update. "The source code for #Winamp [IP owned by #Llama] has been taken offline…This comes as no surprise, as there have been signs. You see, when the source code first appeared on GitHub, there were numerous issues with it. Take, for instance, the fact that forking was not allowed, distribution of modified versions was not allowed, and only official maintainers were allowed to distribute the source code for Winamp."
news.itsfoss.com/winamp-disast…

Winamp's Brief Experiment on Opening Their Source Ends in Disaster

It was a short-lived dream for Winamp in the open source world, even if it was imperfect.

^{Sourav Rudra (It's FOSS News)}

#winamp #llama

in reply to petersuber

mastodon - Collegamento all'originale

petersuber

in reply to petersuber • 5 mesi fa • •

Update. "While the Open Source Initiative (OSI, @osi) is diligently working on defining the term “#OpenSource #AI,” our work [at the @linuxfoundation] focuses on a narrower scope, extending from the Model Openness Framework we’ve developed in LF AI & Data. These definitions represent a natural evolution of our ongoing efforts and are aligned with the broader goals of openness, transparency, and collaboration that underpin the open source community."
lfaidata.foundation/blog/2024/…

Embracing the Future of AI with Open Source and Open Science Models – LFAI & Data

^{lfaidata.foundation}

#opensource #ai @Open Source Initiative @The Linux Foundation

Questa voce è stata modificata (5 mesi fa)

in reply to petersuber

mastodon - Collegamento all'originale

petersuber

in reply to petersuber • 5 mesi fa • •

Update. "The Open Source Initiative (OSI, @osi)…today released version 1.0 of its #OpenSource #AI Definition (OSAID)." Good coverage of the controversies and dissents.
techcrunch.com/2024/10/28/we-f…

The definition itself
opensource.org/ai/open-source-…

We finally have an 'official' definition for open source AI | TechCrunch

The OSI, the self-appointed arbiter of all things open source, has released its first definition of 'open source' AI.

^{Kyle Wiggers (TechCrunch)}

#opensource #ai @Open Source Initiative

in reply to petersuber

mastodon - Collegamento all'originale

petersuber

in reply to petersuber • 5 mesi fa • •

Update. HELIOS Open (@heliosopen) asked its advisory committee to comment on the new Open Source Initiative (#osi, @osi) definition of #OpenSource #AI.
heliosopen.org/news/defining-o…

Defining Open Source AI: Current Conversations within the Academic Community

Defining Open Source AI: Current Conversations within the Academic Community The Open Source Initiative (OSI) , a California public benefit corporation and not-for-profit community of technology experts, recently published the Open Source AI D…

^{Caitlin Carter (Higher Education Leadership Initiative for Open Scholarship)}

#opensource #ai #osi @Open Source Initiative @HELIOS Open

in reply to petersuber

mastodon - Collegamento all'originale

petersuber

in reply to petersuber • 5 mesi fa • •

Update. "If you believe Mark Zuckerberg, #Meta's #AI large language model (#LLM) Llama 3 is #OpenSource. It's not. The Open Source Initiative (#OSI, @osi) spells it out in the Open Source Definition, and Llama 3's license – with clauses on litigation and branding – flunks it on several grounds. Meta, unfortunately, is far from unique in wanting to claim that some of its software and models are open source. Indeed, the concept has its own name: #OpenWashing."
theregister.com/2024/10/25/opi…

The open secret of open washing – why companies pretend to be open source

Allowing pretenders to co-opt the term is bad for everyone

^{Steven J. Vaughan-Nichols (The Register)}

#opensource #ai #osi #meta #Openwashing #LLM @Open Source Initiative

in reply to petersuber

mastodon - Collegamento all'originale

petersuber

in reply to petersuber • 4 mesi fa • •

Update. "Maximally ‘open’ #AI allows some forms of oversight and experimentation on top of existing models. However, we find that openness alone does not perturb the concentration of power in AI. Just as many traditional #opensource software projects were co-opted in various ways by large technology companies, we show how rhetoric around ‘open’ AI is frequently wielded in ways that exacerbate rather than reduce concentration of power in the AI sector."
nature.com/articles/s41586-024…

Why ‘open’ AI systems are actually closed, and why this matters - Nature

A review of the literature on artificial intelligence systems to examine openness reveals that open AI systems are actually closed, as they are highly dependent on the resources of a few large corporate actors.

^Nature

#opensource #ai #Openwashing

in reply to petersuber

mastodon - Collegamento all'originale

petersuber

in reply to petersuber • 3 mesi fa • •

Update. @sj argues that the #MozillaFoundation contradicted its own earlier positions when it endorsed the Open Source Initiative (#osi, @osi) definition of #OpenSource #AI.
samjohnston.org/2024/12/18/a-f…

A Forgotten Manifesto: Mozilla Betrays Its Own Values on Open Source AI - Sam Johnston

The Mozilla Foundation (MoFo), "a global nonprofit dedicated to keeping the Internet a public resource that is open and accessible to all," proudly proclaims

^{Sam Johnston}

#opensource #ai #osi #mozillafoundation @Open Source Initiative @Sam Johnston

in reply to petersuber

mastodon - Collegamento all'originale

petersuber

in reply to petersuber • 3 mesi fa • •

Update. "The #OpenScholar team has released not only the code for the language model but also the entire retrieval pipeline, a specialized 8-billion-parameter model fine-tuned for scientific tasks, and a datastore of [#OpenAccess] scientific papers. 'To our knowledge, this is the first open release of a complete pipeline for a scientific assistant LM —from data to training recipes to model checkpoints,' the researchers wrote in their blog post announcing the system."
venturebeat.com/ai/openscholar…

#LLM #OpenSource

OpenScholar: The open-source A.I. that’s outperforming GPT-4o in scientific research

OpenScholar, an innovative AI system by Allen Institute for AI and University of Washington, revolutionizes scientific research by processing 45 million papers instantly, offering researchers citation-backed answers and challenging proprietary AI sys…

^{Michael Nuñez (VentureBeat)}

#opensource #openaccess #LLM #openscholar

Questa voce è stata modificata (3 mesi fa)

in reply to petersuber

mastodon - Collegamento all'originale

petersuber

in reply to petersuber • 2 mesi fa • •

Update. "While #OpenAI has open sourced models in the past, the company has generally favored a proprietary, closed source development approach. “[I personally think we need to] figure out a different open source strategy,” #SamAltman said…In a follow-up reply, Kevin Weil, OpenAI’s chief product officer, said that OpenAI is considering open sourcing older models that aren’t state-of-the-art anymore."
techcrunch.com/2025/01/31/sam-…

#AI #LLMs #OpenSource

Sam Altman: OpenAI has been on the 'wrong side of history' concerning open source | TechCrunch

In a Reddit AMA, OpenAI CEO Sam Altman said that he believes OpenAI has been 'on the wrong side of history' concerning its open source approach.

^{Kyle Wiggers (TechCrunch)}

#opensource #ai #openai #LLMs #samaltman

in reply to petersuber

mastodon - Collegamento all'originale

petersuber

in reply to petersuber • 2 mesi fa • •

Update. "Researchers at #HuggingFace are trying to replicate [#DeepSeek] from scratch in what they’re calling a pursuit of “open knowledge”…[They seek] to build a duplicate of R1 and #OpenSource all of its components, including the data used to train it…Technically, R1 is “open” in that the model is permissively licensed…However, R1 isn’t “open source” by the widely accepted definition because some of the tools used to build it are shrouded in mystery."
techcrunch.com/2025/01/28/hugg…

Hugging Face researchers are trying to build a more open version of DeepSeek's AI 'reasoning' model | TechCrunch

A group of Hugging Face engineers, including the company's head of research, are spearheading an effort to replicate DeepSeek's R1 model.

^{Kyle Wiggers (TechCrunch)}

#opensource #ai #LLMs #huggingface #DeepSeek

Questa voce è stata modificata (2 mesi fa)

in reply to petersuber

mastodon - Collegamento all'originale

petersuber

in reply to petersuber • 2 mesi fa • •

Update. "On Tuesday, #HuggingFace researchers released an #OpenSource #AI research agent called "Open Deep Research," created by an in-house team as a challenge 24 hours after the launch of OpenAI's Deep Research feature."
arstechnica.com/ai/2025/02/aft…

Hugging Face clones OpenAI’s Deep Research in 24 hours

Open source “Deep Research” project proves that agent frameworks boost AI model capability.

^{Benj Edwards (Ars Technica)}

#opensource #ai #LLMs #huggingface

in reply to petersuber

mastodon - Collegamento all'originale

petersuber

in reply to petersuber • 1 mese fa • •

Update. From Julien Sobrier: "We need a common understanding of what an open model means [for #AI and #LLMs]. We want to watch out for any #OpenWashing, as we saw it with free vs #OpenScience software."
artificialintelligence-news.co…

Endor Labs: AI transparency vs ‘open-washing’

As the AI industry focuses on transparency and security, debates around the true meaning of “openness” are intensifying.

^{Ryan Daws (AI News)}

#ai #openscience #Openwashing #LLMs

in reply to petersuber

mastodon - Collegamento all'originale

petersuber

in reply to petersuber • 1 mese fa • •

"Bruce Perens, who wrote the original #OpenSource Definition and parted ways with #OSI [@osi] in 2020, denounced the idea of the OSAID [Open Source #AI Definition] last year. He believes AI is incompatible with the open software movement because 'its output is inherently plagiarism…The Open Source AI Definition requires less of AI than the original Open Source Definition requires of any other form of software,' said Perens…'My contention is that it isn't Open Source and is Openwashing.'"

#opensource #ai #osi @Open Source Initiative

Questa voce è stata modificata (1 mese fa)

reshared this

in reply to petersuber

mastodon - Collegamento all'originale

petersuber

in reply to petersuber • 1 mese fa • •

Useful table showing in what respects major #AI / #LLM tools are open and in what respects they are not.
osai-index.eu/the-index?type=t…

From the European #OpenSource AI Index.
osai-index.eu/

The Index

A community-driven public resource on open-source generative AI systems in the European Union.

^{European Open Source AI Index}

#opensource #ai #LLM

reshared this

in reply to petersuber

mastodon - Collegamento all'originale

Henrique Santos

in reply to petersuber • 6 mesi fa • •

From my perspective, a machine learning model can be considered open source if anyone can "compile" it from scratch using source code and source data. I don't think I can train my Llama model from scratch.

⇧