Title: Reflections on the Looming Cloud Cost Tsunami

---

October 24, 2022. The day I woke up to another barrage of Slack messages from my Ops team, each one a reminder that our cloud bill was spiraling out of control. It felt like the tsunami warnings had come true, and we were already knee-deep in water.

A Cloud Bill Soaring Out of Control

The past few months have been a whirlwind of tech trends—AI/ML infrastructure growth post-ChatGPT, platform engineering normalization, and the ongoing FinOps revolution. But amidst all this excitement, one thing has consistently kept me up at night: our cloud costs.

I’ve spent countless hours arguing with my team about which services to use, how to optimize, and where we can cut back. It’s a never-ending game of optimization chess, trying to balance performance requirements with financial constraints. This isn’t just a problem for us; it’s an industry-wide challenge that feels like a permanent state of crisis.

The Great Pyramid Scans and Developer Experience

On the lighter side, I stumbled across an intriguing story about someone 3D scanning the interior of the Great Pyramid at Giza. It’s fascinating to think about how modern technology can help us explore ancient wonders in ways that were unimaginable just a few decades ago. This kind of innovation is what keeps me excited about tech.

But even as I was inspired by these grand projects, my mind kept drifting back to the mundane task of optimizing our cloud infrastructure. The thought of managing costs and resources while ensuring we stay ahead of cutting-edge technologies can be overwhelming. It’s like trying to ride a wave on a surfboard that keeps getting bigger.

WebAssembly’s Promise and Reality

WebAssembly (Wasm) is another area where I’ve been investing time. The promise of running compiled code on the web without losing performance—now, there’s something that sounds revolutionary. But as with most technologies, it comes down to how well it fits into our existing workflows.

We’ve experimented a bit, but Wasm hasn’t quite become a core part of our tech stack yet. The learning curve is steep, and it requires significant changes in our development processes. However, there are hints that this could be the future for server-side execution. For now, we’re keeping an eye on it, waiting to see where the technology will go.

GitHub Copilot’s Mixed Feelings

The GitHub Copi­lot investigation hit a nerve with me. I’ve been using Copilot in my personal projects and found some scenarios where it’s incredibly useful—like quickly generating boilerplate code or suggesting syntax improvements. But there are also moments when it feels like the tool is trying too hard to be helpful, leading to code that doesn’t quite fit our style.

This duality reminds me of the broader conversation around AI in software development. It’s a double-edged sword; on one hand, it can significantly boost developer productivity and creativity. On the other, there are concerns about over-reliance and potential issues with maintainability. As platform engineers, we’re constantly navigating this balance.

The Great Debate: LLMs vs. Traditional AI

The LLM (Large Language Model) hype has been heating up since ChatGPT’s debut. There’s a lot of buzz around these models, but the practical applications are still evolving. Our team has had internal debates about whether to invest more time and resources into LLMs or stick with traditional machine learning approaches. The verdict is still out on which path will yield better results in our specific use cases.

Reflecting on My Journey

As I sit here, trying to make sense of all this, it’s clear that the tech landscape continues to evolve at breakneck speed. We’re not just dealing with new tools and technologies; we’re also grappling with the implications of these advancements on our workflows and budgets.

In many ways, it feels like we’re living in a perpetual state of optimization—a never-ending cycle where every decision can either save us money or lead us into deeper complexity. But that’s what keeps things interesting, right? The constant challenge to do more with less, while still delivering high-quality products and services.

For now, I’ll keep plugging away at optimizing our cloud infrastructure, learning from these debates, and exploring new technologies like WebAssembly. After all, it’s in the daily grind where true progress happens.

---

That’s a wrap for today. Back to the trenches!