Elevator Pitch
We are now in an era where AI is an important part of technology. While open source software has now entered the mainstream and has been used by many companies. Proprietary AI and LLM models seem to be going against this direction and raise a lot of questions regarding AI and open source.
Description
In this talk, we will first try to answer the complex question of “what is open source AI?”. Since AI is different from a traditional software application where there are training data involved. it makes it hard to be fully transparent about how an AI or LLM model was trained and this makes a fundamental difference between an open source software - which its source code is available for anyone to inspect. So, an application with its source code being open source but is using a LLM model that is not fully open source would make this fall into a grey area in terms of open source definition.
After that, we will look at how “open” popular LLM models are and compare their performance. We would like to look into whether or not a more “open” model is performing better than a less “open” model. From there we can predict what strategist companies might want to adopt regarding developing LLM models and using LLM models in their applications.
Topics covered: - definition of open source AI - regulations regarding AI models and the challenges - Comparing open source AI and LLM models with proprietary ones - future of AI and open source
Goal By looking at arguments regarding open source and AI and answering questions, audiences would be more aware of how AI is changing the ecosystem of open source software and how to prepare for the changes.
Target Audiences Anyone interested in open source and AI, they maybe those who are actively participating in open source software development or those who rely on them at work.