Update README.md

README.md

@@ -81,7 +81,7 @@ Modal is the computational backbone of this project. Every training run, batch i
 | Full evaluation (421 articles) | 6+ hours (timed out) | 4.4 minutes |
 | Cost per evaluation | ~$25 | ~$0.12 |
 
-This efficiency made the entire V7 experimental cycle possible. Details in [Understanding Why Everything Was Slow](#understanding-why-everything-was-slow) and [The Fine Tuning Journey](#the-fine-tuning-journey-from-timeouts-to-signal). See also: [An Opinionated Argument for Why Modal](https://open.substack.com/pub/basilwong/p/an-opinionated-argument-for-why-modal) for a deeper write-up on why Modal was the right choice for this project.
+This efficiency made the entire V7 experimental cycle possible. Details in [Understanding Why Everything Was Slow](#understanding-why-everything-was-slow) and [The Fine Tuning Journey](#the-fine-tuning-journey-from-timeouts-to-signal). See also my blog: [An Opinionated Argument for Why Modal](https://open.substack.com/pub/basilwong/p/an-opinionated-argument-for-why-modal) for a deeper write-up on why Modal was the right choice for this project.
 
 ---