

Shipping massive film rolls is a logistical problem. And you can’t infinitely copy film rolls either. 70mm in theatres is an outdated format that’s not scallable for the future. Although I can see how 70mm improves picture quality even after digitizing to 4K: colors, contrast and dynamic range is much better.
LLM watermarking is economically desireble. Why would it be more profitable to train worse LLMs on LLM outputs? I’m curious for any argument.
Also, what has deep-fakes anything to do with LLMs? This is not related at all.
A certificate for “real” content is not feasible. It’s much easier to just prevent LLMs to train on LLM output.