Social media was awash this week with demos of the new "Kling" model. Some are unbelievably good, some pretty good, and others just janky.
Many artists are quick to dismiss it because it lacks control. "Can you provide masks? Can you refine fx/animation? Provide 32b EXR for review?"
And I get it. AOVs, EXRs and fine control obviously lead to a better result. You won't prompt Dune.
But that's ignoring the big advantage that AI has: the longtail of niche stories.
Look at how podcasts replaced late night TV. Despite having no fancy sets or crane dollies over a live audience, millions prefer them because they offer topics that are more niche to their interests. I recently enjoyed 4 hours of two dudes discussing the history of Coca-Cola. An idea so terrible, it would be laughed out of every pitch room. But it works because it's low cost and there's 8 billion people with specific interests.
Similarly, AI doesn't need to create perfect movies, it just needs to create more niche content.
I don't think Hollywood will be "replaced" anytime soon. There will probably always be a desire for higher budget movies that we watch together, created with the control we all know today. But I think investments into these high-budget productions will shrink as increased competition make it too risky. In its place will be a burgeoning community of creators that use low-cost tools like AI to publish niche stories direct to Youtube.
I don't know where traditional 3D software fit into this future. Some days I feel like Phil Tippett witnessing the death of practical fx. Other days I'm excited to see tools that let you create without needing to know what a fucking vector is.
What do you think? Where will 3D be in 10 years? Hit reply. I'll read everything.