Ash Koosha made a 2.5-hour film nobody needed to see

Odysseus: The Fall exists because rendering technology is now cheap enough that a single person — Fountain 0 cofounder Ash Koosha, who also directed, wrote, composed. Starred — can generate 2. 5 hours of coherent imagery without a crew. The result is unwatchable. Not in the way a bad screenplay is unwatchable. Unwatchable in the specific way that means the images themselves refuse to hold meaning: they have no depth, no visual texture that suggests anyone made a choice about what matters on screen.

The article frames this as evidence that AI-generated film is creatively bankrupt. It isn't. It's evidence that Koosha's film is creatively bankrupt, which is a smaller, less interesting claim.

Consider digital cinematography in the late 1990s. Directors who had spent careers composing for film stock picked up digital cameras and made work that looked flat, characterless, overexposed. Not because the medium was creatively broken. Because thirty years of craft knowledge didn't transfer. Roger Deakins didn't know yet how to light digitally. Nobody did. The technology arrived before the grammar did. Gatekeepers (film critics, cinematographers, institutions) declared it fundamentally inferior. Then the grammar arrived. Deakins lit Bourne Ultimatum in 2007. Sonya Mead graded Blade Runner 2049 in 2017. Digital stopped looking cheap because people spent a decade learning to see in it, to compose in it, to make its specific weaknesses (no grain, no flicker, a particular kind of color sensitivity) into arguments instead of apologies.

A badly made film made with new tools still reads as a badly made film, full stop. But the diagnosis matters. Odysseus: The Fall isn't evidence that generative imagery can never carry aesthetic intention. It's evidence that Koosha either didn't develop one or didn't know how to translate it through these particular tools. The article doesn't distinguish between those. It assumes the tools are the problem. That's how we've always misread technological arrival.

A badly made film made with new tools still reads as a badly made film, full stop.

The real question isn't whether AI can make art. It's whether enough people will care enough to develop the craft standards that turn capability into vision. That requires patience. It requires failure. It requires years of work by people who aren't treating the tools as permission to make something fast. Koosha made his film quickly. That's visible in every frame. The technology isn't the problem. The haste is.