GenAI to Code Faster is Data Center Thinking
Automation, automation, automation. We hear about it all the time. Investors know it can improve margins. Executives know it can save time. Managers know it will mean less mistakes. It’s even the A in the DevOps CAMS model (Culture, Automation, Metrics, Sharing).
It’s true, automation is all of those things. But they are besides the point. The real value in automation is that it enables innovation and value creation. How?
Data center thinking
I was working with a company that had made the transition from the data center to cloud. Now instead of going to the data center to plug in cables, they were crafting networks, load balancer rules, hosts, security patches, etc. with a mix of scripting and console tweaking. Each environment was still a bespoke artisanal creation and it took about 2 weeks. Then they invested in real automation, and afterwards the same outcome took how long? 20 minutes.
The development team was furious! The operations folks were now talking about unlocking potential, iterating, and more, and instead of the development team being able to complain that Ops were the blockers on each project, now they were the bottleneck. To the dev team, the process of building in the cloud was just like the data center. The only thing that had changed was that they had to wait less time for the Ops part to be built.
But that’s not what had happened. Automation didn’t just save time, it had unlocked the ability to iterate quickly to improve the product. The Second Way of DevOps is about shortening and amplifying feedback loops. That is what automation really provided: the ability to build something, change something, see if it worked, tear it down, and quickly build a pristine environment for the next experiment. Over. And over. The ability to rapidly experiment and learn. This rapid learning ability unlocks innovation and value creation because we are no longer bound by the lengthy time constraints of bespoke artisanal creation.
Talk to a developer about the new generative AI coding tools. The thing that gets them excited is not being able to generate boilerplate code faster (data center thinking). Repeatedly generating boilerplate violates Don’t Repeat Yourself (DRY) principles. The thing that is exciting is that they can experiment. It takes much less time to quickly try out a new idea. They can tinker faster. This allows them to learn more quickly, find the best answers, and get better results.
Using automation to do things more quickly is great. Using automation as a helper to make the humans more empowered to deliver more value, more quickly, is where the real differentiation comes to light.