GitHub’s Post

For the last few months, I've been looking into open source AI — the confusion around defining 'open source' AI, the potential risks or issues with restrictive licenses, the performance differences across open LLMs, and so on. I think this is a fascinating and (murky world) with many ripple effects for developers. Happy to share an article on LeadDev today that looks into open source AI and tries to make some sense for engineering leaders on how to approach it! There are some great gems here on how to evaluate and safely use open-source AI. Thanks to everyone who took the time to share commentary with me for this story! This one features, in order of appearance: - Michael Lieberman, the cofounder and CTO of software supply chain security company Kusari. - Nathan Lambert, a machine learning scientist at the @Allen Institute For AI (Ai2) - Marcus Edel, machine learning lead at Collabora - Paul Harrison, senior security engineering lead at Mattermost - Roman Shaposhnik, cofounder of Ainekko and the The IP Paperclip Project Ltd. - Stefano Maffulli, executive director of Open Source Initiative (OSI) - Leonard Tang, cofounder of Haize Labs The US government is taking a “cautious yet optimistic” approach. What do you think, should you be careful with 'open source' AI? Is it worth it to use it? There's definitely more to be said about the state of AI, and I'm excited to follow this space. #opensource #AI #openAI #opensourceAI #OSI #llm #genai

It was clear throughout this conference that the work to create a standard Open Source AI Definition that upholds the fundamental freedoms of Open Source is vital for addressing some of the key challenges in AI and ML development and democratization.

Open Source Initiative Hits the Road to Define Open Source AI... The OSI is planning a series of workshops to help it refine its working definition of open source AI, with support from Amazon, Cisco, Google and the Sloan Foundation.

This is not only semantic, Open source AI is NOT open source code as detailed in the below article but a mix of honest openness usually brought by academic researchers (who need to disclose their work to get academic credits btw) and some marketing tricks (in my view at least) from our well known com specialists. All open initiatives are still good to take and recognise though 👇 https://lnkd.in/eJTam3Uh

JOIN THE DISCUSSION ON OPEN SOURCE AI The Open Source Initiative (OSI) is on a mission to define “Open Source AI.” It wants to be able to provide a framework that can be used to determine if an AI system is open source or not, as there is not currently a framework for doing so, the OSI explained. OSI is driving a multi-stakeholder process to define an “Open Source AI” and you can be part of the conversation. https://lnkd.in/gFQqzyt6

🤔 The age-old battle between open source and proprietary software has hit the AI space, stirring up debates on what "open source" truly means for AI. 🤖💡 🤗 Paul S. describes for TechCrunch the current state of standardising open-source for #GenAI. Read more at - https://lnkd.in/er4ipkJJ - What does 'open source AI' mean, anyway? | TechCrunch 🤯 Meta’s Mark Zuckerberg has been in the spotlight for his “open source AI” claims. However, many argue that Meta’s Llama models aren't truly open source. This debate is at the heart of what the Open Source Initiative (#OSI), led by @Stefano Maffulli, is trying to address through conferences, workshops, and reports. 🤫 The OSI has been defining "open source" for over 25 years, but applying these definitions to AI is tricky. Joseph Jacks, an open-source evangelist, insists there’s no such thing as open-source AI because neural network weights (NNWs) aren’t comparable to software code. 🧠🔍 😎 Despite the challenges, #OSI is working on the Open Source AI Definition. This draft includes principles like granting freedoms to use, study, modify, and share AI systems. A key issue is whether an AI can be “open source” without its training dataset being available. Maffulli emphasizes knowing where the data came from and how it was processed is more critical. 😱 The OSI aims to make the Open Source AI Definition robust and adaptable, expecting a "stable version" by October. This definition will help navigate the complexities of AI openness, ensuring clarity and transparency for the future. 🌐📜 📌 We, at knowdroids.ai are paying a lot of attention to the open source development. Some of our droids are prepared based on open source #LLMs and components. We believe this is the best solution for some of our clients. Not always the biggest and most expensive #LLMs are needed as brain for our solutions. 🚀 #AI #OpenSource #TechInnovation #OSI #ArtificialIntelligence

What does ‘open source AI’ mean, anyway? It should grant freedoms to use the system for any purpose without seeking permission; to study how the system works and inspect its components; and to modify and share the system for any purpose.

What does ‘open source AI’ mean, anyway? It should grant freedoms to use the system for any purpose without seeking permission; to study how the system works and inspect its components; and to modify and share the system for any purpose.

In TechCrunch, Paul Sawers argues that the term ‘open-source AI’ lacks a clear definition. Stefano Maffulli, executive director of the Open Source Initiative (OSI), is trying to develop a clear definition of the new term. #OpenSourceAI and #OpenSourceSoftware have fundamental differences. “The intentions might be similar, but they are not like-for-like comparable, and this disparity is what the OSI is trying to capture in its definition.” As the tech industry continues to deepen its understanding of AI capabilities, these definitions will inevitably evolve and adapt with the progression of time. Read more about the OSI’s process ⬇️

The struggle between open source and proprietary software is well understood. But the tensions permeating software circles for decades have shuffled into the artificial intelligence space, in part because no one can agree on what “open source” really means in the context of AI. https://lnkd.in/dU-NNUAR