How I designed the most efficient deepfake detector in the world with $100
summary
Claims a practical 100% detector for images from ThisPersonDoesNotExist by exploiting the site’s global cache and indexing every served face. Uses a $100 Jetson Nano to download ~72k images/day, stores FaceNet vectors in Elasticsearch kNN, and matches suspicious profile pictures to a 15–16 million image corpus to return a visual match and an exact timestamp of generation. Works in real runs but only for images generated while the downloader is active.
AI-written summary · about
chapters (5)
related talks
46mHow I built the world’s most efficient deepfake detector with $100
35mHow I Built the World's Most Powerful Deepfake Detector for €100
56mDeep Fakes 2.0 - How Neural Networks are Changing our World
51mPlaying Offense and Defense with Deepfakes
58mPushing Deepfakes to the Limit-Fake video calls with AI
01h 04mFace Verification using Few-shot Deep Learning
32mHeart to Heart and Eye to Eye with Deep Fakes
30mAI Gone Rogue: Exterminating Deep Fakes Before They Cause Menace