Craig McLuckie on Culture as a Team's Operating System in the AI Era

Introduction to AI and Open Source Challenges

• The decisions made about AI adoption, architecture trade-offs, and team collaboration will have a long-lasting impact on systems, and making the right choices is challenging due to the rapidly changing landscape 10s.
• Craig McCluckie is the founder and CEO of Stacklog, and he has a background in enterprise technology, having worked at Google where he contributed to the development of Google Compute Engine and Kubernetes 2m6s.
• McCluckie also participated in the bootstrapping and early leadership of the Cloud Native Computing Foundation, which aims to bring together a community to support Kubernetes and drive innovation in the space 2m6s.
• Open-source communities are special because they allow people to come together and work on technology they care about, driven by a desire to create something useful and see it being used, rather than for financial gain 4m42s.
• The open-source community is a pure expression of engineering joy, where people can bring their creativity and skills to work on something they love and care about, with the potential for broad use and impact 4m42s.

Craig McCluckie's Background and Open Source Philosophy

• The advent of generative AI is changing the open-source community, making it challenging due to the echo of what is happening in mainstream organizations, and affecting the collaboration and community dynamics 6m15s.
• The open-source community provides a great environment for young engineers to work on complex projects, access rich mentorship, and let their creative juices flow, outside of traditional organizational structures 6m15s.

Impact of Generative AI on Open Source Communities

• The introduction of coding AI has become a significant problem for open source communities, as it overwhelms them with low-quality, AI-generated code submissions that are not consistent with the community's work, creating tension among maintainers who have to review these submissions 10s.
• Open source communities traditionally welcome new engineers and provide them with opportunities to learn and grow by assigning them "good first issues" to work on, but the influx of AI-generated code has disrupted this process, making it difficult for new engineers to contribute meaningfully 42s.
• The increasing availability of cheap and good code generated by AI tools is raising questions about the relevance of traditional, structured systems used in open source communities, such as the mpm ecosystem, and whether they will be able to adapt to a new world where code becomes an intermediate language 2m6s.

AI Adoption and Organizational Culture

• The adoption of generative AI is having a significant impact on organization culture, and its effects vary depending on the organization's maturity, including operational maturity, cultural maturity, and a new dimension called AI maturity 4m6s.
• Organizations with high AI maturity are better equipped to integrate AI tools into their day-to-day life cycle, making developers more productive, but this also presents a huge challenge as AI tools become an integral part of the development process 6m10s.
• Engineers are having to adapt to a new role, similar to a manager, where they review and provide direction on code produced by AI tools, rather than writing the code themselves, which can lead to emotional fatigue 10s.

Challenges of AI Code Generation in Development

• The use of AI code generation tools can lead to a significant increase in code production, with some organizations seeing a 300% increase in code and a 400% increase in bugs, highlighting the need for structure and maturity in the development process 10s.
• The introduction of AI code generation tools has led to a shift away from the traditional goal of making pull requests smaller and more granular, with some engineers now producing large pull requests, such as 3,000-line requests, due to the ability of these tools to produce a lot of code quickly 10s.
• Engineers may be producing larger pull requests because they are used to working on a task for a certain amount of time and now have tools that can produce more code in that time, and also due to urgency and incentive structures that prioritize demonstrating fluency in the use of these tools 2m6s.
• The use of AI code generation tools can create tension and friction within teams, particularly if there is not enough structure and maturity in the development process, and if the team is not aware of how to use these tools effectively, leading to discontent and trouble 10s.

Productivity Metrics and Cultural Misalignment

• The measurement of productivity by lines of code is not an effective metric, as it does not take into account the value and quality of the code, and can lead to counterintuitive incentive structures that prioritize quantity over quality, as noted by the Bill Gates quote 2m6s.
• Culture is considered the operating system of a team, defining how the team processes information and establishing a set of commonalities, with the best teams having high levels of diversity and a common core that serves as a set of principles to revert to 10s.

Building and Tuning a Deliberate Culture

• A deliberate culture is not a one-size-fits-all story, but rather needs to be tuned to what the team is trying to accomplish and balanced for the set of constraints the team has, such as available resources 2m6s.
• To design a deliberate culture, an exercise can be done where the team's core ideals are distilled out through interviews, and then identifying the things that people care about and are objectively important to the mission, such as operating with precision and care in a security domain 2m6s.
• The identified core elements are then distilled down to a handful of deliberate culture anchors, which become the foundation for talking about the culture constantly, including during interviews and promotion processes 2m6s.
• The culture anchors serve as a description of how the team assesses people as a good addition to the team, and people can learn the culture, but it should reflect the nature of the team, making it an anchor for describing who gets into the team 2m6s.
• The promotion process should tie back to the specifics of the culture, with organizations having levels such as suite one, suite two, and staff suite, and making sure there are effective ties to the culture 2m6s.
• Companies like Amazon have created a culture that can span different domains, but it is still important to tune the culture to the team's specific activities and constraints 2m6s.

Maintaining and Evolving Organizational Culture

• To reinforce culture, decisions should be related to the cultural markers or anchors described for the organization, and reassessing is crucial as hypocrisy can kill the culture, which means leaders must make hard choices and be consistent with the culture 10s.
• The culture needs to evolve and adapt to changing business imperatives, and leaders must deliberately update the culture and communicate the changes to the team, ensuring that the culture remains aligned with the organization's mission 2m6s.
• Culture is the total operating system of an organization, and it cannot be reduced to a procedural blueprint, but rather it is a dynamic and evolving system that requires awareness and deliberate effort to maintain, with some companies like Amazon having successfully created a self-replicating culture 4m30s.

Implicit vs. Explicit Culture and Leadership Role

• There is a distinction between implicit and explicit culture, with implicit culture being what the team is actually doing, and explicit culture being what the organization says its culture is, and it is essential to ensure that these two are compatible 6m20s.
• A CEO's job is grounded in three key things: setting the culture of the organization, setting the strategy, and finding leaders who can execute the strategy within the parameters of the culture, with culture being the most important and least well understood aspect 8m40s.
• Leaders like Jeff Bezos have successfully created cultures that are self-replicating and can scale, but what works for one organization may not work for another, and it is essential to find a culture that aligns with the organization's values and mission 10m0s.

Culture as a Foundation for Organizational Success

• The success of an organization is heavily dependent on its culture, which can take many forms and does not have to be uniform, as long as it is a good culture that solves the problem at hand, and this understanding needs to be embraced by the organization's leadership 10s.

Changing Career Paths and Roles for Engineers

• The career journey for engineers is undergoing significant changes due to the impact of AI, which is providing access to a vast amount of human knowledge and changing the way engineers work, with tasks such as writing unit tests or fixing bugs being displaced 2m6s.
• The traditional path to becoming a great engineer, which involves writing code, understanding systems, and exploring the boundaries between teams, is being disrupted, and it is unclear how to guide individuals from the start of their career to becoming experts in distributed systems design 2m6s.
• The role of engineers is evolving, with value creation happening at a different layer, and code becoming more like an intermediate language, requiring new skills such as assessment and risk mitigation, which may not be supported by the current education system 2m6s.
• The future of engineering and the skills required to succeed in this field are uncertain, and it is essential to think about how to support teams and individuals in this new landscape, although there is no clear map for where it is heading 2m6s.

Leadership in the Age of AI and New Tools

• As a team lead, it is essential to recognize that not being an expert in understanding new systems, such as copilot, cursor, cloud code, and other tools, can hinder the ability to serve the team, and the temptation to rely on these tools to do the work can be constant 10s.
• New leads often face the challenge of balancing their own expertise with the need to support their team, rather than displacing and alienating them by doing the work themselves, which requires a delicate balancing act to maintain 42s.
• The role of a team lead is to drive the productivity of other people, not to fall into doing more work themselves, and to support their team in producing the work product, rather than treating AI teammates as a priority over human teammates 1m6s.
• It is crucial to maintain the culture of the team by emphasizing that responsibilities remain the same, despite the use of new tools, and that everyone must own everything they do, regardless of whether it was produced by a coding agent or someone else 2m6s.

Adapting to New Systems and Mindsets

• Letting go of assumptions and existing rules is necessary to take advantage of the opportunities presented by new tools and systems, such as stochastic systems, which require a different approach and mindset 4m30s.
• Team leads must challenge themselves to think differently and consider new ways of working, rather than relying on traditional operating cadences and assumptions, to produce something in a more efficient way 5m20s.
• To generate value, it is necessary to learn what works and what doesn't through experimentation, as people cannot simply assume that a system will work by designing it and stitching together a few components 10s.

Developing and Experimenting with Agentic Systems

• Agentic systems do not function like simple designs, and their development requires hands-on experience and a willingness to learn from mistakes, rather than just assuming that a design will work 42s.
• The development of agentic systems can be compared to operating a lawnmower, which requires constant effort to achieve the desired results, and over time, expertise and competence can be built up to improve the system 1m6s.

Conclusion and Contact Information

• For those who want to continue the conversation, Craig McLuckie can be found on LinkedIn under his name, where he is likely to be the top result in a search 2m6s.