AI is breaking free from the data center and moving to the edge—right into your browser. But what does this new power actually look like? This session moves beyond typical demos and invites you to become part of the experiment.

We'll explore the 'how' and 'why' of in-browser AI, focusing on the game-changing duo: WebGPU for raw, cross-platform GPU acceleration and WebAssembly (Wasm) for running high-performance C++/Rust code in JavaScript.

Forget a passive video analysis; we’re building a live, "AI-Powered Scavenger Hunt." The entire audience will be challenged to use their own devices, pointing their cameras to help a browser-based Vision Language Model (VLM) find a list of real-world objects. We'll watch as the model, running entirely on-device, processes dozens of live video feeds simultaneously, demonstrating the future of private, low-latency, and truly interactive web applications.

Key Takeaways:

Understand the roles of WebGPU and WebAssembly in the new AI web stack.

Learn how to load and run sophisticated computer vision models (e.g., ONNX) in a JavaScript-first environment.

Witness a live, large-scale interactive demo that showcases the performance of on-device AI.

Brainstorm real-world applications for decentralized, private, and high-performance browser-based AI.