The Audacity of Ambition: Can AI Really Vibe Code GTA 6?
There’s something undeniably captivating about watching someone attempt the impossible. Ziwen Xu, a 25-year-old AI startup founder, has decided to take on a challenge that’s equal parts audacious and absurd: vibe coding his own version of Grand Theft Auto VI before its official release in November. Personally, I think this is the kind of bold, slightly reckless endeavor that makes the tech world so fascinating. It’s not just about the outcome—it’s about the sheer audacity of the attempt.
What makes this particularly fascinating is the method Xu is using: vibe coding. For the uninitiated, vibe coding is a relatively new approach where developers lean heavily on AI assistants to generate and debug code using natural language prompts. It’s like having a conversation with your computer, but instead of small talk, you’re building a game. What many people don’t realize is that vibe coding is still in its infancy. While high-profile figures like Jack Dorsey have experimented with it, even its inventor, OpenAI’s Andrej Karpathy, has called it “net unhelpful.” So, Xu’s decision to use this method for a project as complex as GTA VI feels like trying to paint the Mona Lisa with a toothbrush.
From my perspective, the real intrigue here isn’t whether Xu will succeed—let’s be honest, he probably won’t—but what his attempt reveals about the current state of AI and human ambition. On day one, his game consisted of a 3D blue oval jumping around gray blocks. By day two, he had a human-like character running through an urban landscape, albeit one that mistakenly included LA skyscrapers instead of Florida’s skyline. If you take a step back and think about it, this is both impressive and hilariously flawed. It’s a reminder that AI, for all its advancements, still struggles with nuance and context.
One thing that immediately stands out is Xu’s commitment. He’s posting daily updates, sharing his code repository on GitHub, and even admitting his mistakes openly. This level of transparency is rare in the tech world, where failure is often swept under the rug. In my opinion, this project isn’t just about building a game—it’s a public experiment in creativity, persistence, and the limits of AI.
But here’s the deeper question: What does this say about our relationship with technology? Xu’s project was inspired by a post from another AI founder, Matt Shumer, who suggested using Anthropic’s Claude Fable 5 to create a GTA VI-caliber game. This isn’t just about one person’s ambition; it’s part of a broader trend of tech enthusiasts pushing AI to its limits. A detail that I find especially interesting is how this reflects our growing reliance on AI as a creative partner. Are we using AI to augment our abilities, or are we outsourcing our creativity?
What this really suggests is that we’re at a crossroads. AI is no longer just a tool—it’s becoming a collaborator, a co-creator. But as Xu’s project demonstrates, it’s far from perfect. The AI built LA skyscrapers in a game set in Florida, and Xu burned through 33% of his weekly AI usage in a single day. These aren’t just technical challenges; they’re reminders of the gaps between human vision and machine execution.
If you ask me, the most compelling aspect of this story isn’t the game itself—it’s the human behind it. Xu’s willingness to take on an almost certainly impossible task, to share his progress publicly, and to embrace failure as part of the process is what makes this story resonate. It’s a reminder that innovation isn’t just about success; it’s about the courage to try.
As we watch Xu’s project unfold, I can’t help but wonder: What will we learn from this? Will it inspire others to push the boundaries of AI, or will it serve as a cautionary tale about overreliance on technology? Personally, I think it’s a bit of both. Xu’s attempt to vibe code GTA VI is more than just a quirky tech project—it’s a reflection of our collective ambition, our fascination with AI, and our enduring desire to create something extraordinary, even if it’s destined to fall short.
In the end, whether Xu beats Rockstar to the punch or not, his project has already achieved something remarkable: it’s sparked a conversation about what’s possible, what’s practical, and what it means to dream big in the age of AI. And that, in my opinion, is a win in itself.
