Update blog

Author: cmdr2

content/blog/2025-11-05-1762335811.md

@@ -11,7 +11,7 @@ tags:
 This post concludes (for now) my [ongoing deep-dive into ML compilers](https://cmdr2.github.io/notes/2025/10/1760088945/), while researching for [sdkit v3](https://cmdr2.github.io/notes/2025/10/1760085894/). I've linked (at the end) to some of the papers that I read related to graph execution on GPUs.
 
 Some final takeaways:
-1. ML compilers might break CUDA's moat (and fix AMD's ROCm support)
+1. ML compilers might break CUDA's moat (and fix AMD's ROCm support).
 2. A single compiler is unlikely to fit every scenario.
 3. The scheduler needs to be grounded in truth.
 4. Simulators might be worth exploring more.