diff --git a/src/main/resources/assets/voxy/shaders/heriacial_selector/hierarchical_worker.comp b/src/main/resources/assets/voxy/shaders/heriacial_selector/hierarchical_worker.comp
deleted file mode 100644
index 4aba598e..00000000
--- a/src/main/resources/assets/voxy/shaders/heriacial_selector/hierarchical_worker.comp
+++ /dev/null
@@ -1,32 +0,0 @@
-#version 460
-
-layout(local_size_x=8)
-
-
-
-
-//NEW IDEAm use the depth buffer directly to compute the lod level needed to cover it
-
-
-
-
-//Location in world space up to 2x2x2 block size resolution
-#define OctNodeTask uint64_t
-
-
-
-//First 32 bits are the start of child
-// next 16 bits are split into 8 pairs, each pair specifies the type of the subnode (air/empty, partial, full)
-
-
-
-//Tasks are of size uint64_t
-
-
-
-void main() {
-    while (true) {
-        barrier();
-
-    }
-}
\ No newline at end of file
diff --git a/src/main/resources/assets/voxy/shaders/heriacial_selector/persistent_threads.comp b/src/main/resources/assets/voxy/shaders/heriacial_selector/persistent_threads.comp
deleted file mode 100644
index 70d17993..00000000
--- a/src/main/resources/assets/voxy/shaders/heriacial_selector/persistent_threads.comp
+++ /dev/null
@@ -1,37 +0,0 @@
-#version 460
-
-#define WAVE_SIZE 64
-layout(local_size_x=WAVE_SIZE)
-
-#define WorkTask
-
-
-//or make the shape 4x2x4 which has a local size of 32
-
-
-//The work queue is a circular queue with collision detection and abortion
-struct WorkQueueHeader {
-    uint queueSizeBits;
-    uint start;
-    uint end;
-    uint _padd;
-};
-
-
-//Task is a uint32,
-
-//The idea is to use persistent threads + octree culling to recursivly find bottom level sections that satisfy a pixel density requirement
-// given a matrix
-void main() {
-    while (true) {
-
-    }
-}
-
-
-
-//Idea cull/recuse in a manor of 4x4x4 cube (64 bits), have an existance mask per section so that unnessasery computation isnt done on air subsections
-// if a section is fully or partially visible and its aabb does not occupy 1 pixel (or a subset of some specified area/density)
-// then enqueue that section as a job
-// once a node has reached its tail ending, check if its loaded or not, if not, request it to be loaded
-// note that the cpu side can discard sections if they are superceeded by a higher level lod load request etc
diff --git a/src/main/resources/assets/voxy/shaders/heriacial_selector/traversal.comp b/src/main/resources/assets/voxy/shaders/heriacial_selector/traversal.comp
deleted file mode 100644
index efbc0d47..00000000
--- a/src/main/resources/assets/voxy/shaders/heriacial_selector/traversal.comp
+++ /dev/null
@@ -1,24 +0,0 @@
-#version 460
-
-layout(local_size_x=32)
-
-//Works in multiple parts
-// first is a downwards traversal from a base level that finds all the bottom level pixel detail AABBs
-// task queue is firstly filled with large visible AABB's from rastered occlusion
-//
-// each node metadata contains the position in 3d space relative to the toplevel node
-// an offset into the datapool for the child nodes, and union between (a bitmsk of if a child node is full, empty, or mixed (or unloaded))
-//  and a pointer to render metadata for meshlets
-
-
-//The overarching idea is to have meshlets be automatatically selected based on the resulting pixel size/density
-// after 3d projection, we dont want subpixel triangles and we want to be able to automatically account for the
-// perspective warp on the edges of the screen (e.g. high fov == higher density at the center of the screen)
-// from this, the gpu can then (if they are not present) request meshlets be added and thereby automatic lod selection
-// and dynamic building resulting in possibly O(fast) rendering
-void main() {
-    while (true) {
-        barrier();
-
-    }
-}
\ No newline at end of file