Edge AI
5 pieces on Edge AI.
News & Analysis
Local AI Is a Workload Decision, Not a Capability Race
Three releases—Muse Spark Open, Nemotron 3.5 Super VL, and AI Edge Foresight—show local AI wins only where privacy, latency, or offline requirements justify its costs.
Google AI Edge Foresight Takes Meeting Notes Fully Offline
Google's free experimental macOS app transcribes meetings and audio files entirely on-device with EmbeddingGemma 2, keeping local files and notes out of the cloud.
WebGPU + DuckDB-Wasm Bring Real AI Workloads Into the Browser
James Hall shows how WebGPU, Transformers.js, and DuckDB-Wasm make client-side AI inference practical for production workloads — without sending data to the cloud.
NVIDIA TRTMC: Hugging Face to C++ TensorRT in Two Commands, No ONNX
NVIDIA's TensorRT Model Connect converts supported checkpoints to native C++ inference in two CLI commands, no ONNX export, across 76 model families.
Needle 2: 45M-Parameter Tool-Calling Model in a 14MB Binary
Cactus Compute's Needle 2 runs a full inference session in 28MB of RAM, hits 500 tokens/sec on a Raspberry Pi 5, and needs no GPU.