My Response to C Heller’s “Generative Video: a human’s thoughts…”
I was writing a reply to Christina Heller’s wonderful, candid, authentic reflections on Google’s video/speech generation tool Google VEO 3 and it turned into a minor essay (ie LinkedIn said I overran it’s limit) so I’m posting it here.
When folks said “why do people who love VR hate reality” we countered — “in fact it gives you a greater appreciation for it.”
When a cinematographer prompts a shot they will request something different than a person who does not understand scene composition, focal lengths, circles of confusion, etc.
When a set designer prompts a shot they will pay attention to the environmental storytelling
And when an actor prompts a shot, maybe the inner life of the character.
Also, maybe someone who will never be able to own a Canon C300 will get a shot at exploring their vision, and maybe they can communicate it more fully to a team, or maybe the work created with the generative AI tools will stand on it own.
Maybe the next time they watch the practical effects in a Gilliam film they will marvel in awe at their palpability and preciousness.
Maybe they will more deeply appreciate the subtle twitch responses of an actor in a closeup from a Cassavetes film.
I think it is incredible you shot your own movie in the desert, Christina — what an incredible mountain to climb — a deeply collaborative, multidisciplinary art that is the example of many talents coming together as one.
Interestingly, as someone who has done QA for a company that vets the quality of AI responses to requests, I have also reflected on the fact that these algorithms and machine calculations and predictors and recommenders are also, using reinforced learning from human feedback (RLHF) a massive collaboration of the best a group of humans can do to help themselves with new tools.
I am not conflating advanced chatbots with the output of a human hand-crafted film — an era that we are seeing also gradually disappear — economically, and in terms of its perceived value — but perhaps these new generated lead to a desire to examine them anew and marvel at their complexity and depth.
They can all co-exist, and our appreciation can deepen. We will discover, and have already seen contemporary auteurs using GANs, gaussian splats, process digital works, for inspiring output guided by their intrinsic experiences and ideas.
(Yes, also, LLM based writing is inherently derivate, a collaborative tool at best and one that I would not include in the list above for reasons I’ll address in a later article — not having to do with IP and copyright so much as inescapable patterns).
And, of course, there are those will take the ‘easy way out’ and fill the trough with lazy slop that will form another glut.
The generative era is here. It’s not going to rewind to a time before it arrived, and let’s see what we can say with it, but also, as you so beautifully closed in your article Christina, let’s foster AND communicate to those artists — musicians, costumers, sculptors, basket weavers, dancers, our appreciation of their unique expressive nature and the huge investment they make to cultivate it as a creative output, with a renewed understanding of their value.
