Smart Contracts 101: What Actually Runs On-Chain

Smart Contracts 101: What Actually Runs On-Chain

People hear "smart contract" and picture something closer to a legal document than code. It's neither smart in the AI sense, nor a contract in the legal sense. It's just a program.

A program that:

• Lives at a fixed address on a blockchain
• Runs exactly the same way for every single node that executes it
• Can hold and move funds without a human approving each transaction
• Can't be edited after deployment (usually — more on upgradeability another day)

That last point is the one that should scare you a little. In a normal backend, a bad deploy means a rollback and an apology in the incident channel. On-chain, a bug in a contract holding real funds can mean funds that are gone. Permanently.

That's why the actual engineering work in smart contracts isn't the happy path. It's the 20 ways a function can be called that you didn't plan for — reentrancy, integer overflow, a malicious caller pretending to be someone else.

Writing the logic is the easy 20%. Defending it is the job.