diff --git a/NOTES.md b/NOTES.md
index c3d9eb6..0b8bc7a 100644
--- a/NOTES.md
+++ b/NOTES.md
@@ -9,6 +9,9 @@ RAM64
 RAM512
 : I used the premise of the previous note and decided to index the RAM64 moduls by the LSB instead of the customary MSB. It still works :metal:
 
+PC
+: The Program Counter was tricky. I ended up using a Or8Way (instead of a `Or3Way`, which we don't have) along with `Mux8Way16` to pick the input for the register. Since the Mux has duplicate inputs at this point (supports 8, but we only have 3 special, and 1 neutral case) - this can be optimized by switching to a `Mux4Way16`, along with some other changes.
+
 Fill.asm
 : Figured out that my RAM16K implementation was wrong while working on this. The rough pseudocode would be:
     ```c
diff --git a/README.md b/README.md
index 26faa22..f3719b9 100644
--- a/README.md
+++ b/README.md
@@ -52,9 +52,21 @@ Make sure you read through the [Hardware Simulator Tutorial][s] to understand th
 
 [s]: https://b1391bd6-da3d-477d-8c01-38cdf774495a.filesusr.com/ugd/44046b_bfd91435260748439493a60a8044ade6.pdf
 
-## [Project 3: Machine Language Programming](https://www.nand2tetris.org/project03)
+## [Project 4: Machine Language Programming](https://www.nand2tetris.org/project03)
 
 Counting number of instructions by `wc -l $file.hack`
 
 - [x] Mult (18)
 - [x] Fill (98)
+
+## [Project 5: Computer Architecture](https://www.nand2tetris.org/project05)
+
+- [ ] `Memory.hdl`
+- [ ] `CPU.hdl`
+- [ ] `Computer.hdl`
+
+Computer chip tests:
+
+- [ ] `Add.hack`
+- [ ] `Max.hack`
+- [ ] `Rect.hack`