Despite a lot of advancements and optimizations on all technological fronts, image compression techniques have not changed a lot for more than 30 years (in the case of JPEG). Or at least their progress is very slow.
https://techcrunch.com/2020/05/07/deep-render/
Deep Render's founders claim that they are completely reinventing compression for images and videos and optimise more than 50% in size with better quality!
They claim that they achieve this by optimising the colour, hue and saturation details of an image to what human eye can see and grasp and getting rid of all other details. And they do this optimisation using AI & ML.
This is insanely blows my mind. Just imagine more than 50% of worlds storage capacity will be freed up overnight or it will take half the time to send good images of space from light-years away!
There are still a lot of unanswered questions that puzzles me. Some of them are, after optimisation do they then compress into JPEG or a completely new file format? If they are completely new format how quickly do current rendering devices and browsers adapt to this new file format? How do transparencies look? Is this lossy technique good enough for medical sciences or space?
Till I get those answers, I will be keenly watching Deep Render's progress.
I wish Deep Render's founders a very good luck for this important cause.