What we will never invest in.
A short list of categories we won’t touch — not because they can’t make money, but because they can’t make money the way we know how to make it. Knowing where you don’t play is half the job.
Everyone publishes investment theses. Almost nobody publishes anti-theses. We think that is backwards. The list of things you won’t do is more disciplining than the list of things you will, because saying yes is easy and saying no is what protects the model.
So — the no list. Subject to revision, but not lightly.
Hardware
Bills of materials, supply chains, returns logistics, capex cycles. Every assumption that makes our four-person model work breaks. We respect hardware companies. We are not one.
Pure marketplaces
Marketplaces require liquidity on both sides before either side will pay. The path to profitability runs through multi-year subsidization. That is a venture-fund category, not a projects-builder category.
Heavily regulated B2B (banking, insurance, healthcare core)
The compliance overhead is its own full-time job. The sales cycles are quarters. The team you need to ship is large by construction. Anything we’d build here would dismantle the rest of the studio.
Anything that requires a unicorn outcome to break even
Some businesses are great if they win and a complete loss if they don’t. We don’t fund those. The off-switch test eliminates them at the gate.
Crypto-native primitives
Token-launch dynamics, regulatory uncertainty, audience that rotates between platforms every twelve months. Some of it will work and we will be wrong about some of these — we just don’t have the muscle to compete in this category and don’t intend to build it.
Audiences we don’t actually understand
Enterprise IT buyers. Industrial procurement. Niche professional verticals where we are not the user and have no shortcut to the user. We can read the research; we can’t feel the texture. The community-first method doesn’t function without that.
Anything that needs us to grow the team
A project that requires us to be ten people instead of four is a project that breaks the operating model. The model is not negotiable. The project is.
Categories where the only moat is brand
Brand-only moats need decades of compounding spend. We can build brand around a product, but we will not build a product whose only defensibility is the brand. The studio doesn’t have the time horizon for it.
The list will move. Some items will leave it as the model evolves or the market changes. New items will arrive. The discipline is to keep one, in writing, and to update it on purpose.
What is left, after all of these are removed, is a narrow category: profitable B2C products with owned distribution, short time-to-revenue, and unit economics that work without a unicorn outcome. That narrow category is where we live.
← all essays and notes