Book teardown · DeFi
DeFi and the Future of Finance
Finance rebuilt from the bottom up as open building blocks anyone can stack. What that fixes, and what it forgets.
If you read nothing else
Centralized finance has five chronic flaws: gatekept control, limited access, inefficiency, opacity, and parts that don't connect. DeFi's bet is that you fix all five at once by rebuilding finance as open, composable building blocks anyone can use, fork, and stack. The power and the peril turn out to be the same property.
This is the most rigorous of the books I've torn down, and the least hype-driven. Harvey is a finance professor at Duke; there's no token being sold here. The setup is a clean diagnosis: traditional finance is gatekept, exclusionary, slow, opaque, and siloed. Around 1.7 billion people are unbanked, a small business loses roughly 3% on every card swipe, and a wire still takes days. The authors' answer is not to patch any one of those. It's to rebuild the whole stack on open protocols, where a ten dollar transaction and a hundred million dollar transaction are handled exactly the same way.
01The model worth stealing
The idea worth carrying out of this book is composability, what the authors call money Legos. In traditional finance, products are sealed boxes that barely talk to each other; moving a position from one institution to another is slow or impossible. In DeFi, every primitive is an open, shared interface: a stablecoin, a swap, a loan, a derivative. Because they all speak the same standard, any product can plug directly into any other. A lending position becomes a token you can post as collateral somewhere else. A stablecoin becomes a brick a hundred other apps build on. Products stop being things a company builds and become recombinations of public building blocks that anyone can snap together.
That is the genuinely new idea, and it travels well beyond crypto. It's the same instinct behind API-first design and embedded finance: open primitives plus permissionless composition. Hold that lens and a lot of where financial infrastructure is heading snaps into focus.
In DeFi, every financial product is a public building block, and anything can plug into anything.
02Three ideas I took from it
- The five flaws are a portable audit tool. Control, access, efficiency, interoperability, opacity. That checklist works on any financial product, crypto or not. The 3% a merchant loses on every swipe (the thing my Swipe teardown was about) is just the "inefficiency" flaw made concrete. Run any system through these five and the gaps light up.
- Some primitives simply can't exist in the old world. The flash loan is the cleanest example: borrow millions with zero collateral, because the smart contract atomically erases the whole transaction if you don't repay inside it. There's no traditional-finance analog. That means DeFi isn't only cheaper finance, it's a new design space, with new instruments that were previously impossible.
- Composability cuts both ways. Stack open blocks and a flaw in one propagates into everything above it. The book's risk chapter is candid and, in hindsight, the most valuable part: smart-contract bugs, oracle manipulation, governance attacks, and the algorithmic-stablecoin "bank run" it flags as a real danger. That last one is exactly what later played out.
03How I read it
My take
This is the most ambitious version yet of one idea: frictionless access for everyone. And I read it the way I read the last two, looking for the back end. DeFi is almost entirely a story about the front door. Open access, instant liquidity, a ten dollar user treated the same as a hundred million dollar one. It is the furthest anyone has pushed the project of removing friction from getting money.
So watch what happens when you can't pay. In DeFi credit there is no hardship plan, no forbearance, no grace period, no one calling to work something out. The book says it flatly about MakerDAO: no broker contacts you, and liquidation can happen immediately. If your collateral slips, a keeper liquidates you in seconds for a fee. That is the entire back end of credit, the part I work on, deleted and replaced with an instant margin call. DeFi does not solve hardship. It removes the part of the system that was built to handle it.
And the inclusion story has a hole I can't unsee. DeFi lending is overcollateralized: to borrow, you have to already own more than you're borrowing. That is the opposite of the problem credit access is supposed to solve. The 1.7 billion unbanked the book opens with mostly do not have surplus crypto to lock up. A system that treats a small and a huge transaction identically still only serves people who already have the asset to post. It underwrites collateral, not people, and the people who most need credit are precisely the ones without the collateral.
As a primer it's rigorous and refreshingly even-handed, and the risk chapter aged the best. But it's a 2021 document written near the top of the cycle, and its headline claim, that DeFi will replace essentially all centralized financial infrastructure, reads as overconfident now. The risks the authors themselves catalog largely came to pass. Read it for the mechanics and the framework, not the prophecy.
04Verdict
The clearest technical explanation of how DeFi actually works, written by people with no token to sell, which is rarer than it should be. If you want to understand open financial infrastructure rather than cheer it or dismiss it, read the introduction, the five-problems chapter, the risks chapter, and the conclusion, and skim the protocol-by-protocol deep dives unless you need them. Keep one hand on the fact that its boldest prediction hasn't aged as well as its plumbing.