[2/2] AI, DevOps, and Kubernetes: Kelsey Hightower on What's Next (Category PlatformEngineering)

Continuing parsing Kelsey interview should mention AI topics and the importance of soft skills

4Artificial intelligence (AI) Hightower is skeptical of hype, but sees fundamental benefits. AI as the new DSLKelsey laughs at 'Prompt Engineering' when people check in on git 500 markdown files with prompts and version them. In fact, we've invented another very unstable programming language. - Indeterministic.: Engineers have fought for predictability all their careers. (tests, idempotence)And now we're introducing AI, which is inherently probabilistic. ("Why guess when you can know for sure?"). - The main benefit of AI: He got vendors and developers to finally write normal APIs and documentation so LLMs could work with them. What we had to do for people (good docks and interfaces)We're doing it for robots now. - Guardrails (Limitations): In the end, it all comes down to creating a rigid framework. (guardrails)To get AI to produce a predictable, “boring” result.

5Development: Man vs. Engineer At the end of the interview, the focus shifts to soft skills and personal growth. **- Team sport.**Kelsey compares working in IT to basketball or football, not athletics. In running, you win or lose alone. In IT, no matter how great an engineer you are, you depend on the team. - Empathy.: It's not just being nice. It’s an understanding that if a developer is afraid to default on Friday, the problem may not be their cowardice, but the unreliability of the platform you’ve built. Professionalism and the Toolbox: Don’t just be a collector of tools. The professional regularly goes through his toolbox. (toolbox)He cleans them and throws them away. Discipline is more important than curiosity.: In a professional environment, you can’t drag Rust or new technology into production just because you’re curious. Choose tools that solve the problem of the business, rather than feeding the ego of the engineer.

#Management #Leadership #PlatformEngineering #Software #SoftwareDevelopment #Architecture #Processes