$ cat post/musing-on-maglevs,-m1-gpus,-and-the-ai-infrastructure-boom.md
Musing on Maglevs, M1 GPUs, and the AI Infrastructure Boom
August 7, 2023. Another day in the life of a platform engineer. I’ve been thinking a lot about the tech landscape lately—specifically, the AI infrastructure explosion that feels like it’s just begun after the ChatGPT craze. It’s wild to see how quickly things are moving.
This week was particularly eventful. On Monday, while I was sipping my morning coffee, I saw an article on Hacker News about magnetic levitation at room temperature. The idea of being able to levitate objects without supercooling is fascinating, but it’s also a stark reminder of the vast and unpredictable frontiers in technology.
Later that day, we were working on optimizing our AI infrastructure for a new project. We’re using WebAssembly (Wasm) more and more on the server side, which feels like a natural progression from JavaScript running everywhere. But every time I try to squeeze another optimization out of Wasm, it’s like trying to fit a square peg into a round hole. The performance gains are there, but sometimes you just have to live with the limitations.
On Wednesday, we had an internal meeting where one of our engineers presented some DORA metrics for our team. Our lead time for changes is still too high, and our deployment frequency could use improvement. We’re always pushing for better, but it’s easy to feel like we’re running in place sometimes. One argument we had was about FinOps—how much should we focus on cloud cost optimization? It’s a balancing act between innovation and keeping the lights on.
On Thursday, there was some excitement around OpenTF announcing a fork of Terraform. I’ve been using Terraform for years now, and it’s amazing to see how much it has grown. But the thought that it might be forked is both thrilling and a bit scary. Change can be good, but stability is also critical.
Friday morning brought a reminder of the real-world impact of tech with a story about ISPs policing online speech. It’s one thing to build tools and platforms; it’s another to grapple with their social implications. I think more developers should consider the broader impacts of what they’re working on—like whether their work is contributing positively to society.
Then, on Friday afternoon, I was reviewing some code for a new project. We were using an M1 GPU driver that just came out. The first version was promising but had some issues with performance and stability. Working through the bugs and optimizations felt like a game of Tetris—trying to fit all these pieces together while keeping the system running smoothly.
Reflecting on it, I realized how much the tech landscape is shifting. AI infrastructure feels like a new frontier, full of both promise and peril. It’s exhilarating but also daunting. We’re constantly learning and adapting, navigating through the latest trends and tools, from magnetic levitation to room temperature superconductors, and everything in between.
In the end, it’s about doing our best with what we have today, while keeping an eye on tomorrow. Who knows what new technologies will come out next week? Or even next year. But for now, I’m just glad to be part of this journey.
That’s my take for August 7, 2023. Hope it gives you a taste of the day-to-day reality in tech today.