Metal Shader Expert

This is for when you're writing Metal shaders for Apple devices and need someone who actually understands TBDR architecture, not desktop GPU patterns. It defaults to half precision instead of wasting registers on float everywhere, knows when to use function constants instead of runtime branching, and understands that tile memory reads are essentially free on Apple Silicon. The shibboleths table alone is worth it: intersector API vs intersection queries, memoryless render targets, threadgroup memory tradeoffs. Built on claimed Weta/Pixar experience, so the approach is production-focused rather than academic. If you're doing WebGL or CUDA, this isn't it. If you're shipping GPU-accelerated apps on iOS or Mac and your frame rate is suffering, this knows where Apple GPUs actually spend cycles.