I spent most of today going down the rabbit hole of Garry Tan’s G-Stack after it kept popping up all over my YT and Reddit feed. If you’ve seen the same clips and hot takes, you’ll know the tone: a lot of people dismissing it as “just a repo of Markdown files for Claude,” or saying AI has finally gotten to Gary’s head.
At first glance, I get why people say that. Open the repo and it doesn’t scream “revolutionary product.” It’s structured, text-heavy, almost deceptively simple. But the more time I spent with it, the more it clicked for me: this isn’t about the md files. It’s about the frameworks encoded inside them.
And that’s exactly why I’ll be adopting parts of the G-Stack and evolving it inside my venture studio, Foundry One Group.
As a solo founder building a portfolio of products—while also assembling a system of agents, automations, and workflows to go from idea to market and eventually to scale—I’m constantly looking for leverage. Not shortcuts, but structured leverage. The kind that increases your odds of building something that actually works.
That’s what the G-Stack offers.
What stands out immediately is the depth of experience baked into it. You can tell this isn’t theory. It’s a reflection of patterns that have played out across hundreds of companies, many of them shaped through Y Combinator. The prompts, the structure, the sequencing—they all push you toward one thing that most founders overlook:
building the right thing.
Because here’s the truth: most founders are using AI wrong.
They bring an idea to an AI tool and get instant validation. “Great idea.” “This could be huge.” “You’re onto something.” AI, out of the box, is designed to be helpful—not critical. It doesn’t naturally challenge your assumptions or force you to confront weak thinking.
What G-Stack does differently is introduce guardrails. It reframes the process. It forces better questions at every stage—from idea, to validation, to MVP definition. Even just the CEO-level thinking layer is powerful enough to justify using it. It creates a kind of friction that most AI workflows are missing.
Can you build this yourself? Of course. You can create your own agents, your own prompts, your own systems.
But the reality is, very few people have the pattern recognition that Gary brings to the table. This is someone who has seen what works—and more importantly, what doesn’t—at scale. G-Stack feels like a distillation of that experience, packaged into something founders can actually use.
For context, if you’re not familiar: Garry Tan is the CEO of Y Combinator, one of the most influential startup accelerators in the world. And G-Stack is his open-source framework designed to help founders move from idea to MVP using structured AI workflows.
So, does it work?
In my experience so far—yes, but with nuance.
I ran a few ideas from our venture pipeline through parts of the stack. What came back wasn’t just refined product thinking; it surfaced market questions and framing that are easy to miss when you’re too close to your own idea. The kind of questions you should be asking early, but often don’t because you’re already emotionally invested.
That alone is valuable.
But let’s be clear: this is not a replacement for real-world validation. No framework, no matter how well designed, replaces speaking to users, getting feedback, and building with your customer in mind.
And this is where I think G-Stack, in its current form, has gaps.
Right now, the stack is heavily focused on market validation and product development. That’s important—but building an app is not the same as building a company.
And that distinction is where many founders fail.
We’re in an era of “vibe coding,” where people can spin up products faster than ever. But speed without differentiation leads to a flood of AI-generated products that no one really trusts, uses, or remembers.
What’s missing is everything around the product.
Not just what you build—but how it lives in the world.
If I were to evolve G-Stack into something that supports the full lifecycle—from idea to scalable company—these are the layers I’d add.
Branding would be the first. In a world where AI makes everything look and sound the same, brand becomes your edge. It’s not just visual identity—it’s tone, positioning, emotional connection. The brands that win today are the ones that feel distinct and intentional. That doesn’t happen by accident.
Then comes marketing and launch strategy. One of the biggest misconceptions among new founders is that great products “find their users.” They don’t. Awareness is engineered. It’s built through repetition, storytelling, and timing. And it starts long before launch. Pre-launch campaigns, waitlists, and narrative-building are not optional—they’re foundational.
There’s also the business layer. Incorporation, ownership structures, compliance, financial systems—these are not glamorous, but they are critical. Founders often delay this thinking, and it creates friction later when they try to scale. A strong framework should guide not just product decisions, but company formation itself.
Scaling is another missing piece. Most MVPs are built with a short-term mindset. Few founders think systematically about how their product evolves over one, two, or three years. What is the roadmap? Where is the moat? How does the product stay relevant? Scaling is not just growth—it’s survival over time.
And finally, fundraising. This is where many startups stall. Building an MVP is one challenge; turning that into a fundable company is another. Investor narratives, due diligence, documentation—these are skills and assets that need to be developed intentionally. A truly complete stack would support founders through this phase as well.
Now, I understand why these elements aren’t part of the current G-Stack. The focus is clearly on product and engineering workflows.
But if the goal is to help founders succeed—not just ship—then the scope needs to expand.
Because the biggest gap isn’t getting from idea to MVP.
It’s getting from MVP to a real, scalable company.
At the end of the day, ideas don’t fail only because they’re bad. Many fail because the approach is incomplete. Turning an idea into a company requires more than building an app. It requires structure, strategy, and an understanding of everything that happens after launch.
That’s where I see the real opportunity.
G-Stack, as it stands, is already a powerful foundation. It captures a way of thinking that most founders don’t naturally have access to. And for anyone serious about building, especially in the age of AI, it’s worth studying closely.
If anything, it feels like a kind of cheat sheet—one shaped by years of experience inside Y Combinator.
Maybe even a subtle signal to future founders applying there.
I’m curious—what’s your take on the stack?
How would you adapt it to your own workflow, and more importantly, how would you evolve it beyond just building products into building companies?