Can AI keep Iris and Gravis consistent?
Explore a character-consistency study and inspect its published comparisons.
Open the study ↗AI video · Research · Production
What can AI video actually deliver?
A home for the tests, comparisons, and production questions behind Artificial General Video. Start with the public evidence in Navo Bench.
01 / Public benchmark archive
Inspect the work.
The original studies and media remain in Navo Bench. This research index points to those public sources so the evidence stays together.
Can an AI presenter look real?
Review the Maggie presenter work, including the visual and production questions behind the tests.
Open the study ↗From storyboard to usable AI video.
Follow the WHEEL OS production study and examine how the work moves from planning into outputs.
Open the study ↗02 / Capability questions
Look beyond the highlight.
Useful questions to bring to an AI video test. These are evaluation lenses, not a claim that every study measures every category.
01 / Identity
Does the character remain recognizable across shots and conditions?
02 / Direction
Does the result follow the requested scene, action, framing, and timing?
03 / Continuity
Do shots, performances, and audio work together as a sequence?
04 / Production
What revisions and effort are needed to reach a usable result?
03 / AGV Report
Get the latest AGV findings in From Content to Capital.
Weekly field notes on turning ideas into content, leads, cash, and capital.
Get From Content to Capital →The AGV Report is the research lens inside Miguel’s newsletter: findings, production lessons, and questions worth testing next.