DevOps & History of AI: John Willis

In this episode, I sit down with John Willis, an accomplished author and a serial entrepreneur known for his contributions to the DevOps community. The conversation spans a wide range of topics, including John’s professional journey, how systems thinking and complexity have influenced his evolution, and the history of AI.

We discuss the history of DevOps, the importance of theory-based practice and how philosophy is integral to advancing science and technology. Tune in for a deep dive into the intersection of technology and human behavior.

Members of the Cyb3rSyn Community can watch/discuss the podcast episode on the www.cyb3rsynlabs.com portal or the mobile app (iOS and Android). The video is also on YouTube here 👇🏾

Beyond his credentials and a long list of accomplishments, John is a personal inspiration for me when it comes to lifelong learning and epistemic humility. As John put it, we are ‘insatiable learners’ who must remain humble at every turn.

Here are the key insights and my reflections from our conversation, spanning from 1980s mainframes to the future of Quantum Computing.

John identifies first as an "ops person," cutting his teeth on massive infrastructure outages where mistakes cost millions. One of his most striking observations is that the "complexity bar" hasn't actually changed.

While we may view Amazon or Google as the peak of complexity today, John argues that the systems Exxon ran back in the 1980s were equally complex relative to their domain. While one can argue that the technology alone is complicated, the human factors definitely made it complex.

John described a "convergence" around 2008–2009. Before this, operations meant "first-generation infrastructure" - tedious, error-prone scripts. Then came a flood of new theory: Mark Burgess’s work on idempotency, and John Allspaw & Richard Cook introducing the community to complexity theory via papers like "Why Complex Systems Fail". It wasn't just about new tools; it was realizing, "Of course, that's the way we're supposed to do this".

We discussed the tension between "best practices" (copying what works) and "theory based practice" (understanding the principles and the why). John shared a story from his time at GE Capital in the 90s. Long before he read Douglas McGregor, he took a "manager modeling" class taught by a female president in the organization - a rarity at the time. She taught them principles of empowerment (Theory Y) without ever using the academic labels. Having said this, John explicitly acknowledged that Jack Welch was a terrible leader.

John views his learning journey as an hourglass:

While some might call John a sycophant of Deming, he points how he has moved on and learned widely from other authors and legends in the field.

For those who treat the DevOps Handbook as a bible, John revealed that the first draft was apparently terrible. They gave it to about 100 people, and the feedback was brutal because it lacked stories. This failure led them to pivot and include case studies from Nordstrom and Target, grounding the theory in reality.

We revisited the legendary "Wall of Confusion" slide by Andrew Clay Shafer, depicting developers and ops separated by a brick wall. The goal of DevOps wasn't just to break down the silo between dev and ops or automation; it was desiloing the organization.

John also shared a regret: the early movement was a "wave" of practical solutions (like "10 deploys a day") that excited people but lacked documented theory and the deep conversations that underpinned it. He wishes they had better cataloged those early "open space" conversations about burnout and culture, which would have made us a smarter community today.

John was also pragmatic in acknowledging that if they had over-indexed on theory, principles and philosophy back in the day, DevOps might not have become as well-known it is today.

We then shifted to John’s recent book, Rebels of Reason, which covers the history of AI. He aimed to write for two personas: Ben Rockwood (the deep expert) and his mother-in-law (the layperson).

While the book is about technology, he mentioned how he remains fascinated about the human stories around it.

One vivid story involved a "Real-life Good Will Hunting" - a boy named Walter Pitts. He is remembered for having spent three days in a library, at the age of 12, reading Principia Mathematica and sending a letter to Bertrand Russell pointing out what he considered serious problems with the first half of the first volume. Russell was appreciative and invited him to study at Cambridge University at age 12. The offer was not taken up; however, Pitts did decide to become a logician and went on to make profound contributions to neural networks.

Finally, John gave a preview of his next project on the history of Quantum Computing, tentatively titled Rebels of Uncertainty. This brings us back to humility.

If you enjoyed the conversation, I highly recommend checking out John’s books:

That’s it for this week. Stay tuned for more multidisciplinary insights next week.