Ethereum co-founder Vitalik Buterin has endorsed TiTok AI, a groundbreaking method for compressing images for blockchain storage.
Unlike the popular social media platform TikTok, TiTok AI focuses on reducing image sizes to make them more feasible for on-chain storage.
TikTok AI was developed by researchers at ByteDance and the Technical University of Munich.
This new technique compresses images into 32 small data pieces (bits) without losing quality, making it highly efficient for blockchain applications.
Buterin praised TiTok AI on the decentralized social media platform Farcaster, noting, “320 bits is basically a hash. Small enough to go on chain for every user.”
The new compression method has significant implications for storing profile pictures (PFPs) and non-fungible tokens (NFTs) on the blockchain.
It uses advanced artificial intelligence and machine learning to tokenize images, breaking down traditional 2D grid constraints into a more flexible 1D format. This results in faster processing speeds and better image quality.
TikTok AI capitalizes on region redundancy, identifying and utilizing repetitive information within different parts of an image to reduce data size.
According to its research paper, TiTok can compress a 256×256 pixel image into just 32 discrete tokens, making it 410 times faster than some existing methods while maintaining competitive generation quality.