Apple's new PICO codec makes your photos tiny, but will it actually launch?
Apple just dropped PICO, a fancy neural codec that claims to shrink images like magic without losing quality. It’s significantly faster than existing AI solutions, yet ironically, there’s no word if it will ever actually leave the lab.
The team at Apple has unveiled PICO (Perceptual Image Codec), an AI-driven system that compresses images by prioritizing what the human eye actually notices. Instead of following rigid, hand-written rules like legacy formats, this system is trained to identify and discard visual data deemed irrelevant to our biological perception, resulting in massive file size reductions.
Benchmarked against industry heavyweights like AV1, VVC, and JPEG AI, PICO purportedly slashes bitrate requirements by two to three times while maintaining comparable visual fidelity. When pitted against other specialized neural codecs, it still manages a 20-40% efficiency lead, though Apple conspicuously omitted JPEG XL from these comparisons.
Speed is the real party trick here. On an iPhone 17 Pro Max, the system processes a 12-megapixel image in under 400ms combined for encoding and decoding. This is a significant leap, considering most current ML-based codecs are tethered to power-hungry server-side hardware like Nvidia V100 GPUs.
Currently, this remains a research project rather than a consumer feature. The lack of a roadmap for iOS or iCloud integration suggests this might just be another piece of impressive tech destined to sit in a GitHub repository until the next major developer conference. It is truly peak corporate behavior to invent a revolutionary way to save space while leaving users struggling with base-tier storage limits for another year.
Source: Apple Open Source
Comments
This is where the magic happens: AI reads your discussion and rewrites the article based on the most interesting comments. Each strong comment adds points to the meter below. Once the meter is full, the article updates live — no page reload needed.