We recently hosted an Ask-Me-Anything (AMA) session featuring Mark Burgess, creator of CFEngine and Promise Theory, focused on applying AI to reduce toil in production systems and improve reliability engineering practices.
Mark started by defining "toil" clearly: it's repetitive, necessary work that does not build knowledge or value over time. Reducing toil, he explained, is crucial for maintaining both system stability and engineer satisfaction.
One of the core ideas discussed was how AI should serve as a complementary tool, supporting human workflows rather than replacing human judgment. Mark stressed that AI should handle tasks like heavy data processing, anomaly detection, and repetitive pattern recognition. Engineers, meanwhile, should retain control of intent, decision-making, and strategic oversight.
A significant part of the discussion addressed the difference between raw data storage (e.g., data lakes, wikis) and actionable knowledge. Mark described traditional data repositories as "data graveyards" due to their lack of active retrieval mechanisms. He emphasized that genuine knowledge requires context, experience, and personal familiarity with systems, areas where human involvement remains essential.
Additionally, Mark introduced his "Semantic Space-Time" model, a lightweight graph-based approach to representing knowledge. This model allows engineers to efficiently analyze relationships between data points, quickly filter noise, and identify critical insights, streamlining troubleshooting and incident response.
Audience questions explored practical ways AI can augment, rather than replace, human tasks. Mark reinforced the necessity of clearly defined, human-driven intent when integrating AI, and highlighted the importance of maintaining personalized knowledge bases and iterative improvement through human oversight.
For further exploration of Mark Burgess’s personal projects, including SSTorytime, Promise Theory and more, visit his website and GitHub repository.
You can try our APX product to see Mark's takeaways in practice at http://tryapx.com/. Test how AI tools can begin eliminating toil today.