Khan Academy - Bitcoin - Proof of work

Short Summary

Main recommendation: Proof of work protocols are a way to prove that someone has spent a significant amount of computational effort, and the video explains how such puzzles are built and verified. The core mechanism is finding a proof string that, combined with a challenge string and run through a cryptographic hash function, produces an output beginning with a set number of zeros.

Key Points

• A proof of work puzzle must be hard to solve but quick to verify; the verifier needs only one hash computation to check the answer.
• Bitcoin's electronic payment system uses a proof of work scheme to create transaction block chains.
• Earlier proposals used the same idea to deter denial-of-service (DoS) attacks, requiring a requester to solve a puzzle before a service responds.
• The original proposed application was deterring spam email: a proof of work is tied to each message, acting like a postage stamp paid in CPU cycles instead of money.
• The approach is cheap for a legitimate sender with few messages but prohibitively expensive for a spammer sending millions.

Details Worth Noting

• Concrete instructions: To solve the puzzle, the prover takes a challenge string (c) and searches for a proof string (p) such that hashing the concatenation of c and p, for example with SHA-256, yields an output whose first set number of bits are all 0. Because the hash output looks random, there is no known shortcut, so the prover must try candidate strings until one works. To verify, the checker hashes c and the proposed p once and confirms the required leading zeros are present.
• Difficulty is adjusted by changing the number of required leading zeros: each added zero doubles the average work, and each removed zero halves it.
• Requiring 40 leading zeros means about 2^40 (roughly one trillion) attempts on average, though individual searches may take much more or less.
• In the spam example, the challenge string represents the email message itself.

Bottom Line

Proof of work turns computational effort into a verifiable signal: expensive to produce, cheap to check, and tunable in difficulty, which is why it can deter spam and underpin Bitcoin.