diff --git a/src/content/posts/WIP-rugbrodsarbejde.md b/src/content/posts/WIP-rugbrodsarbejde.md
index d381bfa..826f814 100644
--- a/src/content/posts/WIP-rugbrodsarbejde.md
+++ b/src/content/posts/WIP-rugbrodsarbejde.md
@@ -1,6 +1,6 @@
 
 ---
-slug: rugbrodsarbejdeIt takes many small steps
+slug: rugbrodsarbejde 
 title: "Rugbrødsarbejde"
 tags: [ai]
 pubDate: 2026-01-01
@@ -9,15 +9,25 @@ draft: true
 
 # ...
 
-Lots of people areasking themselves these days if they should stop coding by hand more or less entirely. More are probably asking themselves what they should generate, and what they shouldn't. I'm in the second category.
+Lots of people are asking themselves these days if they should stop coding by hand more or less entirely. More are probably asking themselves when they can lean in, and when they shouldn't. I'm in the second category.
 
 For context, I largely agree with Antirez' recent (and excellent) [blog post](https://antirez.com/news/158) on the matter.
 I also believe that good software engineering practice, like thoughtful test coverage, is more important than ever.
-But that's a post for another day.
+And I don't really vibe code personally, if you follow the definition that vibe coding means not actually reading the code. 
+I don't know if I ever will.
 
-# What to generate?
 
-How do you actually decide what to generate? Now I'll explain the title of this post, which should make very little sense to you unless you're danish. "Rugbrød" means rye-bread, and "arbejde" means work. The term is about boring routine-work essentially. And if you're in doubt whether to code-gen/automate something then I've found it to be a useful heruristic. "Is it rugbrødsarbejde?" if yes, then go ahead and generate it. You should be in a good position to review and assess it, and use the time-saving on something more important. Is it not? Well, then you probably benefit from getting your hands dirty. And at some point, that kind of work might just end up being rugbrødsarbejde.
+But all of that is a post for another day.
+
+# When to lean in?
+
+How do you actually decide when to lean in and generate larger chunks of code? 
+
+First let me explain the title of this post, which should make very little sense to you unless you're danish. "Rugbrød" means rye-bread, and "arbejde" means work. 
+The term is essentially about boring routine-work that you know perfectly how to do. 
+And if you're in doubt whether to code-gen/automate something then I've found it to be a useful heruristic. 
+"Is it rugbrødsarbejde?" if yes, then go ahead and generate it. You should be in a good position to review and assess it, and use the time-saving on something more important. Is it not? Well, then you probably benefit from getting your hands dirty. And at some point, that kind of work might just end up being rugbrødsarbejde.
+On the other hand, you might find work that used to be rugbrødsarbejde to be foreign to you. And that you might need to get your hands dirty again.
 
 - TODO: add study about learning through writing
 - TODO: add something about simplicity
@@ -25,5 +35,5 @@ How do you actually decide what to generate? Now I'll explain the title of this
 # Hard problems
 
 There's a lot to be said about using agents for experimental work. Throwing them at bugs, or iterating on something challenging that you don't find the time to.
-I think agents can be great for learning, but it's challenging. They easily become a crutch.
+I think agents can be great for learning, but it requires discipline. They easily become a crutch that you depend on fully.
 If you decide to let an agent loose on non-rugsbrødsarbejde over night, you probably want to understand the solution it came up with, might even want to see if you can recreate it from scratch.