Vol. 1 — June 2026
L1 LIBRARY

Built from MIT curriculum · Every chain. Explained.

L1
CONCEPT · SMART-CONTRACTS

Smart Contracts

Programs that execute agreements on-chain without a trusted intermediary. The ledger already disintermediates payments; smart contracts extend that to arbitrary agreements — provided they're done properly.

Last updated: June 10, 2026

Why It Matters

A blockchain that only moves coins is a payment rail. The larger prize, in Micali’s framing from the casebook, is this: the ledger already disintermediates — it executes transfers with no bank in the middle — and the ultimate goal is to do the same for agreements in general. Escrow without an escrow agent, insurance that pays out without a claims department, markets without a clearinghouse.

The casebook is equally clear about the qualifier: provided they’re done properly. A contract that runs exactly as written runs exactly as written — bugs included — and there’s no one to call when “as written” wasn’t “as intended.”

How It Works

Beginner

A vending machine is a contract that enforces itself: money in, snack out, no shopkeeper, no trust. A smart contract is a vending machine made of code and placed on a blockchain: it holds funds, watches for conditions, and releases value according to fixed rules that no one — including its author — can bend after deployment. The blockchain guarantees the machine can’t be unplugged or tampered with.

Intermediate

The primitive form already exists in Bitcoin: Bitcoin Script, via the UTXO Model, makes spending conditions programmable — multisignature requirements, timelocks — a limited, early form of contract. Generalized platforms make the contract a full program with its own state and balance, invoked by transactions.

The casebook’s two cautions (Micali): the DAO hack — the famous 2016 theft from an Ethereum contract — was an implementation bug, not an inherent flaw of the idea; and contemporary smart contracts are too slow, with performance and careful language design named as the open challenges. Both cautions aged well: the field’s history since is largely the story of those two problems being worked.

Builder

The dominant model became Ethereum’s EVM: gas-metered execution, Solidity contracts, state in a global trie. The bug problem matured into a discipline — audits, formal verification, standardized libraries (OpenZeppelin), bug bounties — after roughly a decade of expensive lessons (reentrancy, oracle manipulation, bridge exploits). The performance problem moved to Layer 2 and alternative runtimes (Solana’s parallel execution, Move’s resource types, CosmWasm). Composability — contracts calling contracts — proved to be both the killer feature (DeFi’s “money legos”) and a compounding risk surface.

Examples

  • Bitcoin Script — Multisig and timelocks via the UTXO model; the limited early form.
  • Ethereum — The general-purpose platform; DeFi, NFTs, DAOs all live here.
  • The DAO (2016) — The casebook’s cautionary example: implementation bug, $60M, a chain-splitting bailout.
  • Algorand — Contracts paired with on-chain self-governance for protocol upgrades.

Tradeoffs

Strengths

  • Disintermediation generalized — agreements execute without trusted middlemen, fees, or discretion.
  • Credible neutrality — the rules are public, deterministic, and apply identically to everyone.
  • Composability — contracts build on contracts, compounding what the platform can express.

Limitations

  • Code is the contract — bugs are binding; there is no court of “what we meant.” The DAO is the standing example.
  • Performance — the casebook’s “too slow” still holds at base layer; execution everywhere is metered and costly.
  • Oracle dependence — contracts only know on-chain facts; real-world inputs reintroduce trusted parties.
  • Immutability cuts both ways — unstoppable also means unfixable without governance machinery.

Sources & Last Updated

  • MIT BLC Module 2: Maintaining Blockchain Integrity (primary source; Micali lecture)
  • Vault note: Smart Contracts (M2 cluster)

Freshness note: the vault marks this a stub pre-dating modern EVM platforms; post-casebook material is flagged inline and should be expanded when later modules are processed.

Last updated: June 10, 2026