The story being told is this : Everyone will start making films now. Higgsfield the company behind one of the popular generation platforms and builds one of the better harnesses is creating some originals to showcase how to do this right.
They've created a 22 min episode of a sci-fi show called Hell Grind. The cost of it, 69,000$ - and thats just in credits.
If the team that has built the platform and knows the model incurs this cost, someone who has a learning curve will probably get to 3-4x more. If they don't give up before that ( they picked 253 shots out of 16,181 shots that they generated ).
And they have a 15 people broken into small teams, each one led by a director to accomplish this. Which means you are talking about something like 200K - 300K for a 22min output.
And this is for a sci-fi film - where spectacle carries the story. If someone is going to try to do comedy, or drama or romance, it will probably cost more as audiences will pick up discrepancies and will need to finetune every single shot.
As much as the hype is there, it is helpful to do a reality-check as to where things stand now.
https://blog.quanten.co/ai-filmmaking-is-not-cheap-as-you-think/
1 person likes this
Love this breakdown! It goes to show that the "easy way" is not always the most efficient (economically and productively!) It also goes to show that the some of these new processes can become obsolete!
Sam Rivera There is definitely a use case for it - i really like the fact that it doesn't cost millions to do stories that are visual effects dependent, and can bring down some of the ATL costs that gets attached with actors - it will go back to the strength of the story. But there is definitely an entry barrier - independent film makers who are struggling to raise 20-30K to make their first short, arent going to be able to afford these tools. And this is when much of the credits is subsidized.
But i can see how studios could leverage this to enhance things.