GenAI is everywhere, but for most app developers, building with AI seems daunting. We’re constrained by LLM APIs, lack of control, and rising costs. What if we could democratize AI development for the web without diving into Python data science?
Enter Wasm, WebGPU, and WebAI - key concepts enabling the use of open AI models with WebAssembly and WebGPU. This allows developers to create sophisticated web applications while maintaining control over AI infrastructure and costs, all in JS.
This talk demonstrates how using open-source models, instead of proprietary cloud AI services, enables greater customization, data privacy, and cost management. We’ll explore how WASM and WebGPU can efficiently run these models in-browser, reducing server-side processing. Through demos, we’ll discuss the optimal architectural setup to architect this specific paradigm of GenAI apps.
We’ll also cover strategies for optimizing web performance with open models, addressing trade-offs and implementation considerations for full-stack apps, as well as introducing the cutting edge webNN API to enable web apps and frameworks to take advantage of the best available hardware and software optimisations for various devices.
This is our opportunity to elevate AI and web development. Let’s reclaim control of our GenAI web apps using WASM and open models, making AI accessible to all web developers!
