The Llama large-scale language model claims to be open source, but OpenInfra COO Mark Collier says it's not at all.
Translated from 'Open Source' Has a Definition, Let's Get Serious About Defending It , by Mark Collier.
Artificial intelligence may be a rare transformational era in technology, and it is underestimated. The reality of a multi-trillion dollar market may surprise us all. So it’s no surprise to see the world’s largest companies and governments developing strategies, products, and workflows to capture this part of the market.
But in an overlapping Venn diagram of misleading AI regulations, broken definitions of open source, and all manner of hubris, the future of the global economy and innovation is at risk.
Regulations can shape it, but hubris cannot cure it, so let’s focus on the third variable that the classification demands our immediate attention: the intentional or negligent collateral damage caused to the global open source community by participants in the AI melee.
Llama 3, announced last week, is just the latest in a series of obscene attempts to target the term - we should all remember that it has a definition , and we all work hard to create and defend it. While Llama 3 LLM (and its predecessors) is impressive and deserves credit for achieving benchmarks that advance practical artificial intelligence, one inaccuracy has been reported repeatedly, and it requires empathic correction: Llama 3’s The Meta license is not open source .
To be fair, there is ambiguity here, and the Open Source Initiative is leading an open, community-driven process to detail how the open source definition applies in the world of artificial intelligence. But even without this process, we already know that Meta's custom license violates multiple principles of the current definition of open source and any eventual work by the OSI community on artificial intelligence by restricting the ability to use and create derivative works.
This element of the open source definition—unimpeded permission to create and exploit derivative works—is at the heart of why open source has become a powerful driver of innovation around the world. Any license that restricts this is, by definition, not open source. The following excerpt from the Llama 3 license illustrates this:
b. Redistribution and Use.
i. If you distribute or make available the Llama Materials (or any derivative works thereof), or a product or service that uses any of them, including another AI model, you shall (A) provide a copy of this Agreement with any such Llama Materials; and (B) prominently display “Built with Meta Llama 3” on a related website, user interface, blogpost, about page, or product documentation. If you use the Llama Materials to create, train, fine tune, or otherwise improve an AI model, which is distributed or made available, you shall also include “Llama 3” at the beginning of any such AI model name.
We can be highly confident that the forthcoming definition of open source AI is unlikely to include model assets licensed under such commercially restrictive terms as those set by Meta and others, as they fundamentally limit access and therefore freedom.
Although the media often incorrectly refers to Meta's Llama series of models as " open source ," thus causing confusion, a recent article has been modified to clarify the situation.
Built on open source
It’s important that we recognize the efforts the open source community is already making to support advances in AI workloads.
Nvidia relies on Kata containers and Kubernetes to seamlessly migrate existing AI/ML workloads to confidential environments while combining LLM with GPU-accelerated computing. A few weeks ago, the OpenStack community announced Caracal , the project's 29th release with features including enhanced support for AI workloads, including live GPU migration.
For companies like StackHPC , they already use Dawn, the UK’s largest supercomputer, to support AI workloads. The company posted on LinkedIn, "StackHPC is focused on delivering HPC and AI for private clouds, and we look forward to bringing the features and benefits of OpenStack Caracal Edition to our community and customers."
Open source plays a crucial role in the evolution of AI and is already bearing fruit. Open source faces threats — from malicious relicensing and shady marketing jargon — and the global community of developers and users who have come to rely on open source’s founding principles are understandably nervous and a little angry.
Our global community plays a vital role in educating new audiences about open source as they are exposed to the term for the first time. That is, it unlocks immeasurable creativity and market value, and it has a very specific definition that must be defended if that value is to be preserved for the AI revolution .
RustDesk suspends domestic services due to rampant fraud Apple releases M4 chip Taobao (taobao.com) restarts web version optimization work High school students create their own open source programming language as a coming-of-age gift - Netizens' critical comments: Relying on the defense Yunfeng resigned from Alibaba, and plans to produce in the future The destination for independent game programmers on the Windows platform . Visual Studio Code 1.89 releases Java 17. It is the most commonly used Java LTS version. Windows 10 has a market share of 70%, and Windows 11 continues to decline. Open Source Daily | Google supports Hongmeng to take over; open source Rabbit R1; Docker supports Android phones; Microsoft’s anxiety and ambitions; Haier Electric has shut down the open platformThis article was first published on Yunyunzhongsheng ( https://yylives.cc/ ), everyone is welcome to visit.