📰 Key Takeaways

Three months after Musk announced a “historically accurate” feature-length version of The Odyssey, Grok has released only about seven minutes of it, calling it “episode one of the Odyssey series.” The project was made with Wonder Studios and director Dev Walker. Musk had earlier publicly attacked Christopher Nolan’s Odyssey for casting non-white and transgender actors. Nolan calls himself a “technology skeptic” and said that before AI, he had never seen a technology “so thoroughly rejected by the public.” Nolan’s version relies mostly on practical filming, shot across six countries with thousands of cast and crew. It opened in July to strong box office, sold out IMAX screenings for months, and is seen as a frontrunner for next year’s Oscars.

As for the Grok version, the author admits it’s not as ugly as expected and looks better than other videos Musk has posted recently. Some shots even made them briefly think real actors were sailing on a real sea. Every character has a skin tone Musk approves of, and the narration uses the British accent common in fantasy games and Hollywood epics. But the video has no artistry, inspiration, or humanity: long expository narration drags on, then one bearded face after another delivers stiff lines, making even Nolan’s dialogue sound Shakespearean by comparison. A few individual shots are convincingly realistic, but the whole thing feels lifeless, has no rhythm, and gives you no reason to care about anyone or anything on screen. The author found Nolan’s nearly three-hour version effortless to watch, but kept checking the time remaining before the Grok version was even halfway through, and can’t imagine anyone wanting another seven minutes, let alone the whole film.

💬 JudyAI Lab Take

This story matters because it directly compares what AI-generated video can actually do against what its creator publicly promised. Three months ago Musk said he’d make a “historically accurate” feature-length Odyssey. What he delivered is about seven minutes of a first episode, and the verdict is that it lacks humanity.

One realistic frame doesn’t make a finished work. The original piece notes that some shots even made viewers think real actors were sailing on a real sea, but the overall mix of long narration and stiff lines gives you no reason to care about any character. That shows the current gap in generative AI: it’s good at making individual clips look real, but struggles with pacing, narrative, and emotion, the things that need to be designed as a whole. Another thing worth noticing is that the project’s driving logic was to set a position first and then have the technology back it up, and the technology wasn’t mature enough to match the ambition. For AI builders, a product’s value isn’t in showing off its strongest single capability. It’s in whether users still want to stay after experiencing the whole thing.

Next time you evaluate an AI feature you built, have someone go through the full experience, and watch for the moment they start checking the time.

📅 Source Info

🔗 Further Reading