Why open source is the heart of human progress | Matt Hicks | TEDxSingapore Salon

Key Concepts:

• Open Source
• Community Governance
• Collaboration
• AI Models (Language, Coding, Time Series)
• Apache License
• Granite Model
• Contributors vs. Users
• Instruct Lab
• Synthetic Data Generation
• YAML File Format
• Toil (Repetitive, Mundane Tasks)
• Collective Intelligence
• Open Data, Open Model Licensing, Open Weights
• "First they ignore you, then they laugh at you, then they fight you, then you win."

1. The Role of Communities and Governance in Open Source:

• Open source projects without communities risk evolving in an unbalanced direction. Diverse perspectives from different users and contributors are essential.
• Governance is critical to determine what gets included in a project and how needs are addressed over time.
• Red Hat closely examines the governance of open-source projects as an indicator of their long-term viability.

2. Open Source as a Game Changer: The Linux Example and the Granite Model:

• Linux provided accessible learning opportunities, enabling the speaker to transition from hardware to software.
• The availability of AI models under the Apache license (e.g., IBM's Granite model) allows for commercial use and free distribution, fostering innovation.
• "If you give it away for free, you'll see incredible innovation built on it." - This reflects the philosophy behind open-sourcing AI models.

3. Balancing Usage and Contribution in Open Source Communities:

• Durable open-source projects require a balance between users and contributors.
• Contributing to open-source projects provides a sense of pride and helps drive the technology forward.
• The challenge in AI is the high cost of building foundation models (hundreds of millions of dollars).

4. Instruct Lab: Enabling Domain Experts to Contribute to AI:

• Red Hat's Instruct Lab project aims to enable individuals with domain expertise to contribute to AI models without deep AI knowledge or high training costs.
• Instruct Lab uses synthetic data generation to overcome the need for large datasets.
• Contributors can submit their knowledge in a simple YAML file format, and the project handles the heavy lifting of training the AI model.
• Example: Training an AI model on knock-knock jokes from an eight-year-old to demonstrate accessibility.

5. AI as a Tool to Reduce Toil and Amplify Human Ingenuity:

• AI can be used to automate repetitive tasks ("toil") in software development, allowing communities to focus on their passions.
• "AI can be a tool that lets us do the work that we're passionate about."
• By reducing toil, AI can amplify human ingenuity and make volunteer contributions more impactful.

6. Shaping AI as a Tool:

• Users should be able to shape AI tools for their specific tasks, rather than being limited to a "magic chat box window."
• Open data, open model licensing, and open weights are essential for empowering users to customize AI.
• The goal is to shift the perspective from being a user of AI to shaping it as a tool.

7. The Ethos of Open Source: Patience, Persistence, and Confidence:

• The speaker learned at Red Hat that the path of open source involves being ignored, laughed at, fought, and then winning.
• "First they ignore you, then they laugh at you, then they fight you, then you win."
• Patience, persistence, and confidence are essential for navigating this cycle.

8. Generational Perspectives on Open Source and AI:

• Younger engineers view AI as their generation's Linux, with the same potential for shaping technology.
• The ethos, passion, and belief in open source are expected to remain consistent across generations.

9. Fears and Concerns about the Future of Open Source:

• The speaker's main fear is that companies will successfully argue against open source, leading to fragmentation and limiting innovation.
• Concerns include arguments about security, collaboration challenges, and geopolitical issues.
• There are fears that AI will be controlled by a few corporations working with governments, rather than being open to individual contributions.

10. Conclusion:

The conversation emphasizes the importance of community, governance, and collaboration in open source, particularly in the context of AI. Projects like Instruct Lab aim to democratize AI development by enabling domain experts to contribute without requiring extensive technical knowledge. The vision is to use AI as a tool to reduce toil, amplify human ingenuity, and empower individuals to shape technology for their specific needs. While there are concerns about the future of open source, the speaker remains optimistic about the potential for innovation and the passion of the next generation.