This might be the dumbest thing ever, but what the hell. Here we go:

Is there a reason why Eiichiro Oda isn’t a consultant for the top world model frontier AI labs? I came across Oda through a reply tweet by Eric Glyman, founder of Ramp, who gave a short summary of the obsessive, relentless, unique and consistent approach of the “One Piece” creator. I’m not into anime or manga, even though I’m quite familiar with a few (Naruto, One Piece, Attack on Titan, Hajime no Ippo) because of friends and classmates who were obsessed with them in secondary school. However, I did watch Hajime no Ippo, a boxing-themed manga series, as a teen, and as I recall it was really good, with some amazing fight scenes. David Senra dropped an episode on the Founders Podcast covering how Oda works, and one question kept popping up in my head while listening: “How is someone who has been so deeply entrenched in complex world building for nearly three decades not being consulted on lore by the top frontier startup labs building world models?” To answer this question, I was confronted with yet another question, lol! Are these two things connected, or can they be: a creative storytelling and world-building genius and a scientific generative 3D world model AI institution?

To answer this second question, we need to take a step back and look at world building in the context of storytelling versus the physical world, and herein lies the paradigm gap. Building spatial intelligence for reinforcement learning and robotics requires a deep understanding of low-level physics, mathematical geometry, light and spatial dimensions like depth. Elite storytelling world building, on the other hand, seems to be underpinned by an incessant obsession and thirst for creativity. Still, spatial consistency across different temporal states is crucial, as it mimics key aspects of the real world. For example, I recently came across a vibe-coded world game whose first iteration showed the same location changing drastically after the game character looked at a scene, looked away to a different scene, and then returned to the original one. On the horizon of the first view, the water and sky touched, but when the character turned away and turned back, the water had become land and it was the land that touched the sky. Basically, there were spatial inconsistencies, which are problematic when aiming to mimic the real world. Oda’s artwork has certainly honed him into an elite and meticulous spatial consistency manager, but his expertise as an uncanny creative genius lies in macro-level geography and persistent thematic continuity, a narrative technique rooted in keeping every aspect of a creative work consistent with its central theme across different times, settings and characters. This is what holds a story together and makes it make sense on a deeper emotional level, beyond logic and facts, when considering the bigger picture.

You can see where I’m going with this. Oda won’t be needed for next-token geometric rendering in a physics-based 3D world model, for autoregressive frame generation, or for training a robot to navigate an office. The AI that these frontier labs are investing millions in will handle those, because it is being trained to understand the intuitive laws of physical reality (collision, gravity and perspective). But I bet he can provide value in gaming (which I think is about to explode with 3D generative AI), generative media and simulation styling. If World Labs wants its flagship multimodal generative AI model, Marble, to generate infinitely expansive, self-consistent fictional universes for video games or even virtual production, then Oda seems like a high-level authority on “internal logic continuity” to consult. As a loose analogy, a lack of internal logic continuity, or groundedness, leads to factual hallucinations in ChatGPT. Internal logic continuity sets the premise for a fictional universe to adhere strictly to its own defined physics laws, constraints and cause-and-effect mechanics, even at the expense of defying the real world. It seems to me that this phenomenon underpins the benchmark for the “holy grail” of advanced AI world models at the frontier of spatial intelligence. Minecraft’s global appeal and popularity are rooted in its consistent systemic rules. Games like League of Legends, The Legend of Zelda and Elden Ring are hugely popular because they create fictional worlds that their users get absolutely lost in. I believe we’re about to enter an era of gaming explosion powered by generative AI world models, and it’s not insane to think Oda could have a massive role to play if summoned to participate. The chances are low because he seems insanely entrenched in the world of One Piece, so good luck trying to whisk him away from that.

Anyway, this rabbit hole has led me to commit to starting the One Piece live-action series on Netflix. Could this be my otaku era?