Zero-Copy GPU Inference from WebAssembly on Apple Silicon
agambrahma
58 points
21 comments
April 18, 2026
Related Discussions
Found 5 related stories in 70.1ms across 4,930 title embeddings via pgvector HNSW
- Launch HN: RunAnywhere (YC W26) – Faster AI Inference on Apple Silicon sanchitmonga22 · 199 pts · March 10, 2026 · 61% similar
- Reverse Engineering Apple's GPU Energy Model on the M4 Max ricebunny · 16 pts · March 15, 2026 · 57% similar
- Apple Silicon and Virtual Machines: Beating the 2 VM Limit (2023) krackers · 178 pts · April 11, 2026 · 53% similar
- Apple approves driver that lets Nvidia eGPUs work with Arm Macs naves · 403 pts · April 04, 2026 · 51% similar
- Go on Embedded Systems and WebAssembly uticus · 142 pts · April 03, 2026 · 50% similar
Discussion Highlights (4 comments)
wmf
This works in wasmtime not browsers.
trueno
> on Apple Silicon, a WebAssembly module's linear memory can be shared directly with the GPU: no copies, no serialization, no intermediate buffers enhance > no copies, no serialization, no intermediate buffers would it kill people to write their own stuff why are we doing this. out of all the things people immediately cede to AI they cede their human ability to communicate and convey/share ideas. this timeline is bonkers.
saagarjha
I'm curious what this offers over just building the host side code to be native?
pjmlp
Goodbye WebAssembly "security". Also, these folks should be amazed by 8 and 16 bit games development, or games consoles in general.