Platform dev musings from someone who's built too many things, broken even more, and has opinions about the state of our industry that you probably won't like.
Recent thoughts on platform development, tooling, and the occasional rant about why everything is broken.
Six months, three outages, and countless sleepless nights later, here's what actually happened when we moved 200+ services to our new platform.
I've been using AI coding tools for 18 months now. Here's the uncomfortable truth about productivity gains, technical debt, and what we're actually optimizing for.
Hot take: Most companies that think they need K8s probably just need better deployment scripts and a load balancer.
I'm Ray Timmons, and I've been building platform infrastructure for over a decade, currently serving as Head of Platform Development at Podsphere, where I spend my days making sure our podcasting platform doesn't fall over when millions of people decide they want to listen to true crime podcasts simultaneously.
I've got strong opinions about software architecture, a healthy skepticism toward anything that promises to "revolutionize" development, and an unhealthy relationship with monitoring dashboards. I believe in boring technology, well-tested code, and the radical notion that software should actually work reliably.
When I'm not debugging distributed systems or explaining why we can't just "add AI to it," I write about platform engineering, share war stories from the trenches, and occasionally rant about the state of our industry on YouTube.
The software that keeps me productive, sane, and occasionally both. These aren't sponsorships—just tools that have earned their place in my workflow.
An AI-powered code editor that actually understands context. I was skeptical at first (because I'm skeptical of everything), but after six months of use, it's become indispensable for rapid prototyping and refactoring legacy code. The autocomplete is eerily good, and it's saved me countless hours of writing boilerplate.
My go-to for code review, architecture discussions, and rubber duck debugging when my actual rubber duck is being unhelpful. Claude's reasoning about complex systems is surprisingly nuanced, and it's particularly good at spotting edge cases I might have missed. Worth every penny of the subscription.
Research tool that cuts through the noise of modern web search. When I need to quickly understand a new technology, compare tools, or find authoritative sources on platform engineering topics, Perplexity delivers cited, relevant results without the SEO spam that's infected Google.
Voice synthesis for video content and demos. The quality is genuinely impressive, and it's particularly useful for creating consistent narration for tutorials or generating placeholder audio for platform demos. The British accent models are spot-on, which matters more than you might think.
Video editing that feels like document editing. I can edit videos by editing the transcript, which is exactly the kind of workflow innovation that makes sense. Perfect for cleaning up technical talks and removing the inevitable "ums" and "ahs" that plague my speaking.
Documentation and knowledge management with AI assistance. I use it for technical specifications, architecture decisions, and project planning. The AI features help with writing clarity and structure, which is particularly useful when documenting complex platform decisions for future engineers (including future me).