diff --git a/week1/README.md b/week1/README.md
index 55546ccab60259ef138dc18fc8c51aa7c18c9960..73f1fc4f18b5afb514da8425734821f37ccc526c 100644
--- a/week1/README.md
+++ b/week1/README.md
@@ -2,8 +2,11 @@
 
 This week I wanted to start with a "normal" approach that felt pythonic with a particle class and a bunch of for loops / if statements, but still using numpy arrays. Then rewrite it with everything possible vectorized in numpy and see the difference in solve times for each.
 
-Below is a 5000 particle simulation with 2000 time steps that took 0.66 s to solve for (and incidentally much longer to write to a .mp4 file).
-![](img/animation_fast.mp4)
+Below is a 5000 particle simulation with 1000 time steps that took 0.18 s to solve for (and incidentally much longer to write to a .mp4 or .gif file).
+
+![](img/animation_fast.gif)
+
+https://gitlab.cba.mit.edu/davepreiss/nmm/-/blob/a6f0474d8cdb6bec89cc871ceecbd14fa6b71472/week1/maxwells_demon_fast.py
 
 [](maxwells_demon_fast.py)
 
@@ -16,4 +19,4 @@ gpu is good at multithreading
 optimized for throughput over latency
 might be slower on gpu for small numbers of particles
 
-jax and numba will both 
\ No newline at end of file
+jax and numba will both