Colin Eberhardt: AI Governance, Agentic Coding, and the Future of Open Source

In this episode of Risk-First: Stars of Software, Rob Moffat talks with Colin Eberhardt, CTO of Scott Logic, long-time FINOS contributor, and one of the principal authors of the AI Governance Framework.

Colin has spent years helping financial institutions adopt new technologies safely—without slowing innovation to a crawl. Which makes him exactly the right person to talk to about the biggest technological shift the software industry has seen in decades: AI.

The conversation explores what AI governance actually looks like in practice, why banks struggled to work out whose problem AI even was, and how large organisations can adopt powerful new tools without accidentally causing chaos.

Along the way, Rob and Colin dive into:

• Why AI governance isn’t about bureaucracy, but about helping organisations understand risks they didn’t even know they had
• How non-deterministic systems break many traditional software engineering techniques
• Why testing and feedback loops may become the most important tools in AI-driven development
• The rise of agentic coding loops that can autonomously iterate until tests pass
• How AI could radically change legacy system migration, software delivery, and developer productivity
• Whether AI will flood the world with open-source projects… or quietly make open source less necessary

• Scott Logic
  https://www.scottlogic.com
  UK-based software consultancy focused on complex platforms, trading systems, and large-scale engineering challenges.
• FINOS AI Governance Framework
  https://github.com/finos/ai-governance-framework
  Open-source framework describing risks and mitigations when adopting generative AI in financial services.

• AI Augmented Coding Weekly — Colin’s newsletter
  https://newsletter.scottlogic.com
  Commentary and analysis on how AI is changing software engineering practices.
• The AI Daily Brief podcast
  https://podcasts.apple.com/us/podcast/the-ai-daily-brief/id1669813433
  Regular updates on AI developments, industry trends, and major model releases.

• Claude Code / Anthropic tools
  https://www.anthropic.com
  AI coding agents and autonomous development workflows.
• Next.js
  https://nextjs.org
  Popular React framework used as an example of modern web infrastructure and AI-assisted cloning.
• Ladybird browser project
  https://ladybird.dev
  Experimental open-source browser engine referenced during discussion of AI-assisted codebase recreation.