Show HN: I built a AI powered ad blocker that runs entirely in browser (santhoshaditya.netlify.app)

A developer has created an AI-powered ad blocker that operates entirely within the browser, responding to a take-home assignment challenge. This innovative extension identifies and blocks ads—specifically text and images—by utilizing advanced machine learning techniques. Key to its functionality is the use of the CLIP model from OpenAI for image classification and a BERT-based model for text content. The solution intercepts network requests and performs ad detection without relying on static domain blocklists, thus adapting to the dynamic nature of ad-loading scripts on web pages. This project is significant for the AI/ML community as it pushes the boundaries of browser capabilities and machine learning deployment in real-time contexts. By leveraging technologies such as transformers.js with ONNX Runtime, the developer successfully minimized latency issues associated with backend processing, achieving an efficient client-side model execution. While the extension is still in the testing phase, the approach showcases the potential for more adaptive, intelligent ad-blocking solutions, challenging traditional methods and paving the way for further research and development in this area.