diff --git a/questions/README.md b/questions/README.md
index 54ab95a..bbd0c58 100644
--- a/questions/README.md
+++ b/questions/README.md
@@ -5934,4 +5934,4 @@ That single rule explains a lot: **`property` is a data descriptor, so an instan
 
 **When you still genuinely need a metaclass (Q120/Q134):** when you must change what a class fundamentally _is_ — customizing `__prepare__` (the namespace used while the class body executes), altering the MRO, controlling `isinstance`/`__call__`, or transforming the class object before `__init_subclass__` even runs. Otherwise prefer the hooks: they **compose across multiple base classes** (metaclasses conflict and force you to write a combined metaclass), and they're far easier to read.
 
-The takeaway: reach for **`__init_subclass__`** (react to subclass creation — registries and validation) and **`__set_name__`** (let a descriptor discover its own attribute name) before ever writing a metaclass; combined with the **data vs non-data descriptor precedence** — the rule that explains why `property` beats an instance attribute while `cached_property` deliberately doesn't — these PEP 487 hooks cover the vast majority of "I thought I needed a metaclass" situations with dramatically less complexity.
+The takeaway: reach for **`__init_subclass__`** (react to subclass creation — registries and validation) and **`__set_name__`** (let a descriptor discover its own attribute name) before ever writing a metaclass; combined with the **data vs non-data descriptor precedence** — the rule that explains why `property` beats an instance attribute while `cached_property` deliberately doesn't - these PEP 487 hooks cover the vast majority of "I thought I needed a metaclass" situations with dramatically less complexity.