Cosmonic CTO Outlines Four Ways to Run AI Inference on Wasm
Start here
Cosmonic CTO Bailey Hayes described four approaches to running AI inference on WebAssembly with wasmCloud: wasi-nn, wasi-webgpu, compiling models to Wasm, and custom interfaces and providers. A demo image-analyzer component that calls the Ollama API for image classification is only 292 kilobytes, far smaller than a traditional container. Hayes said wasi-webgpu has no off-the-shelf solution yet and suits more experienced practitioners. For web, edge or local-first use, she recommends a quantized, more compressed model, and says Ollama works well for local and edge applications but requires some form of workload lock before running on the GPU.
Generated from reports · updated 19 h ago
Report timeline
Follow the reports to see the event from different sides.
- Hacker News · AIRunning distributed ML and AI workloads with wasmCloud
Cosmonic CTO Bailey Hayes outlines four ways to run AI inference on WebAssembly with wasmCloud: wasi-nn, wasi-webgpu, compiling models to Wasm, and custom interfaces and providers. A demo image analyzer component uses the Ollama API for image classification and is only 292 kilobytes, far smaller than a traditional container.
Heat for this event
- Current, comparable range
- 6
- Peak, comparable range
- 10November 19 09:00
- Change in 24 hours
- –
The trend compares the same participants over hours observed continuously, so its range can be narrower than the current heat count. Hover or tap the chart for hourly heat; the arrow keys move through it.