The Adversarial Path to the Personal Assistant - Building Private AI Assistants with Personal Data
AI Engineer via YouTube
Earn a Michigan Engineering AI Certificate — Stay Ahead of the AI Revolution
Learn Excel & Financial Modeling the Way Finance Teams Actually Use Them
Overview
Google, IBM & Meta Certificates — All 10,000+ Courses at 40% Off
One annual plan covers every course and certificate on Coursera. 40% off for a limited time.
Get Full Access
Explore how to build a truly personal AI assistant that knows you as an individual in this 19-minute conference talk from the AI Engineer World's Fair. Discover why current large language models like GPT-4, despite their massive training datasets, cannot provide personalized assistance without significant user input. Learn about an innovative approach that creates a personal, private, and proactive AI assistant by downloading your transaction logs from major consumer platforms like Google, Facebook, Amazon, DoorDash, Strava, and Uber directly to your local device. Understand how implementing a RAG (Retrieval-Augmented Generation) system using this personal data enables immediate utility without requiring extensive user configuration. Gain insights into how a modest amount of powerful personal data can unlock the true potential of personalized AI assistance, moving beyond generic responses to truly individualized support.
Syllabus
The Adversarial Path to the Personal Assistant: Sumit Agarwal
Taught by
AI Engineer