The platform sets the rules. Takes a cut. Decides who gets in. Decides who gets shut out. Onchain, the rules are code, the cut is whatever you write into the contract, and nobody can shut down a contract that nobody owns.
The intent stage is small. It is one sentence; written in a notes app at 2am.
i want to build a thing where [X] is impossible to break.
i want to build a thing where [Y] gets paid automatically forever.
i want to build a thing the platform can never delete.
Not the speculation. Not the airdrops. The fact that you can build a thing that runs without you, pays you forever, and cannot be deplatformed. That is the seed.
Frustration with closed systems — you got rate-limited, deplatformed, or watched a platform extract more value than it created.
Seeing a primitive that does not exist anywhere else — onchain royalties, programmable streams, conditional payments. Things money cannot do off-chain.
Economic alignment that does not work off-chain — your earliest users becoming co-owners, paid forever as the protocol grows.
Aerodrome launched on Base in August 2023. Their seed sentence was not “build a DEX.” It was: be the central liquidity layer for a new chain, with emissions that route capital where the chain actually needs it.
That sentence is why they exist. Every decision they have made since then maps back to it. We will follow them across all six chapters.
Not a pitch deck. Not a Y Combinator application. Not a Discord with 500 members. Not a token. Not a roadmap. Those come much later. The intent is the sentence. Everything else is downstream of whether that sentence is sharp enough to survive the next year of doubt.
A builder's first instinct is to imagine the entire system. The platform. The marketplace. The community. The token. The actual move is the opposite. Find the smallest thing that, if it worked, would prove the rest could exist.
A primitive that does not exist anywhere else — onchain royalties, conditional payments, programmable streaming. Things money cannot do off-chain.
A user experience an old industry refuses to ship — faster settlement, cheaper fees, open data.
A community-owned version of something that exists — where the users are also the owners.
Hayden Adams started Uniswap as a learning exercise in Vyper. The first version was a single trading pair. No frontend. No fees. Just a way to test if a constant-product market maker could actually function. That was the wedge.
Ask 5 people who would actually use it. Not friends being polite. Strangers in the use case.
Mock it up in Figma. Send it. Watch reactions.
Open a waitlist. If it cannot get 100 sign-ups in a week, the wedge is wrong.
Ship a fake door. A page that looks real. Track who clicks “buy.”
Ship the dumbest version. One contract. No frontend. Watch if anyone calls it.
Aerodrome did not try to outcompete Uniswap on Ethereum. The wedge was sharper: Base was a brand-new chain with no incumbent liquidity layer.
They forked the ve(3,3) model from Velodrome and bet that a coordination mechanism for emissions would matter more on a new chain than raw fee competition. The smallest version was the contracts plus a launch-day distribution to bootstrap voting power.
Sharp wedges identify a moment, not just a market.
You will start with twenty ideas. Most of them are good. The discipline is killing nineteen of them while the kill still hurts. Builders who refuse to kill end up with ten half-products in eighteen months.
Most Base builders use roughly the same shape: smart contracts in Solidity (Foundry), a wallet integration (Coinbase Smart Wallet, Privy, or Dynamic), a frontend in Next.js with OnchainKit handling wallet flows, an indexer (Goldsky), and Vercel for hosting. Base testnet first. Mainnet second.
Write a function. Deploy it. Something breaks. Fix it. Deploy again. Each iteration teaches you something about how onchain code actually behaves under real conditions — gas, MEV, edge cases, race conditions.
Contracts: Solidity + Foundry (forge test, anvil for local fork).
Wallet: Coinbase Smart Wallet for embedded UX, Privy or Dynamic for email-first onboarding.
Frontend: Next.js + wagmi + viem. OnchainKit for wallet flows, identity, and transaction status.
Account abstraction: Base Paymaster for gas sponsorship — let users transact without ETH.
Indexing: Goldsky or a custom subgraph for reading events at scale.
Hosting: Vercel. Mini-app via Farcaster Frames if relevant.
The romance of “building onchain” ends. The reality begins. The builders who push through the build stage share one habit. They ship something every day, even if it is broken.
Every action costs money. A failed deploy on mainnet is real ETH. A bad migration costs gas. A reentrancy bug can cost the protocol. The build stage is the only stage where a bug is private and recoverable. Use it.
Aerodrome did not write contracts from a blank file. They forked Velodrome v2 — a battle-tested ve(3,3) implementation with two existing audits. The customization was parameter tuning and Base-specific integrations.
The lesson is not “always fork.” The lesson is: when proven contracts exist, ship the differentiator on top, not the boilerplate underneath.
The honest version of launch is unromantic. Most onchain products go live to an audience of one: the builder. The first transaction is a test from another wallet they own. The next is a friend they messaged directly.
The first 100 real users come from one of four places. Farcaster mini-apps. Telegram groups. Twitter threads. Direct partnerships with another protocol. If you have not built one of those four channels before launching, the launch will be quiet.
247 saw the tweet. 31 stopped to look. 8 came back the next day. That is a normal launch. The number on the right is the only one that matters.
Quiet beta. Mainnet, no announcement. DM ten people. Watch them break things.
Token launch. Genesis distribution to a target audience. Fast TVL. Thin organic afterward.
Mini-app. Ship inside Farcaster or Base App. The distribution surface is the launch.
Partnership launch. Co-launch with a protocol whose users you want. Their distribution is your launch.
Aerodrome went live in late August 2023 with a coordinated genesis: an initial veAERO distribution to specific OP holders, early Base users, and select partner protocols who would commit liquidity.
The launch was not a tweet. It was a distribution event with weeks of pre-coordination with the partners who would actually move money on day one. The first $100M of TVL did not come from organic discovery. It came from chosen distribution.
No distribution work before the contract is live. The builder spent six months on contracts and one weekend on the launch tweet. The math collapses in public. Then they quit.
The launch begins three months before the launch. Build the audience while you build the product.
Launch is the start of the loop. Real builders ship the product, watch what people actually do with it, adjust, ship again. The hardest part is killing things that do not work — features you spent two weeks on, channels that brought ten users but no revenue, personas that loved the product but never paid.
Listen. What are users doing that you did not expect? Where do they drop off?
Adjust. Which signal do you respond to? Which do you ignore?
Decide. Are you on the right wedge — or is the data telling you to pivot?
Active wallets, not total wallets. Total grows on launch day; active tells you the truth a week later.
Repeat usage rate. Of the wallets that touched the contract once, how many came back without you nudging them?
Revenue per user. Real fees flowing to the protocol — not token incentives.
Cost to acquire vs lifetime value. The only number that tells you whether the loop is sustainable.
Uniswap was a learning project that became a protocol. OpenSea was a CryptoKitties competitor that became a marketplace. Optimism started as Plasma research. Each one watched the data and changed direction.
Aerodrome shipped in 2023 with a single AMM model: vAMM (volatile) plus sAMM (stable). Solid. Working. By 2024 the data showed users were leaving the stable pools for concentrated-liquidity venues offering tighter spreads.
Their iteration was Slipstream — a concentrated liquidity layer added on top of the existing protocol, integrated with the same ve(3,3) emissions. They did not rebuild. They added the missing primitive the data was asking for.
Iteration that ships beats redesign that doesn't.
The hardest skill: deleting features your users said they wanted but never used. Stated preferences lie. Revealed preferences do not. The kill list is real metrics, not user interviews.
Scaling without signal means burning capital to inflate metrics that will collapse the moment the incentive ends. Three signals, all green, before you spend.
Repeat usage — wallets coming back week over week without you nudging them.
Real revenue — fees flowing to the protocol that did not require a token incentive to exist.
Organic distribution — new users showing up because old users told them.
Scale changes the work. The builder is no longer writing code; they are writing job descriptions, term sheets, partnership emails. The product is no longer a thing they ship every day; it is a system other people maintain. Most builders quit here.
By late 2024 Aerodrome was the largest DEX on Base by TVL, processing meaningful daily volume and routing fees to veAERO holders. The three signals were green: returning LPs, real fees independent of emissions, and organic protocol integrations.
The work that made scale possible happened in chapters two through five. Scale itself is mostly listening — to users, to integrations, to the data — and resisting the urge to ship a feature for every loud opinion.
The builders who survive the transition share one trait. They keep talking to users every week, even at $50M of TVL. The day a builder stops listening to the iterate loop is the day the product starts to die.
Premature token launch. The token becomes the product instead of the product producing the token.
Stop listening. The team gets bigger, decisions move further from users, the product drifts.
Over-hire. The runway burns faster than revenue grows. Suddenly the loop is captive to a fundraise.
Forget the seed sentence. The protocol that solved one problem now tries to solve five. None of them well.
the tools rot. the channels pivot.
the discipline of the journey does not.
everyone you have heard of
started where you are right now.
10 questions. No going back.
Your result is shareable.