Course

The Cold Hard Sats course

In progress: 19 modules planned, new modules added as they're ready

New to Bitcoin? Watch this first

What is Bitcoin, really? Watch a block get built

The fastest way to see what Bitcoin is: a shared record that thousands of computers keep, one block at a time. In 45 seconds, follow a real payment from the waiting room to the chain before you start the modules below.

Open the video page with the full summary →

Data: mempool.space® API · Visual style inspired by the open-source mempool.space explorer · Not affiliated with or endorsed by mempool.space.

Read what the video shows, scene by scene

0:00The chain. Two blocks of payments are waiting their turn on the left. On the right are the most recent blocks already on the chain, #971,867 back to #971,864, each with its number of transactions.
0:03New payments wait in the mempool. A transaction is a payment someone sent; each square is one, colored by its fee. The mempool is the waiting room for payments that are not in a block yet. At the snapshot, 84,252 transactions were waiting.
0:09Higher fee, picked first. The squares re-sort with the highest fees first. A fee is a small tip so a miner includes the payment, measured in sat/vB: sats per virtual byte (1 sat = 0.00000001 BTC). That day about 3 sat/vB was the fast lane and 1 sat/vB the economy lane.
0:15Packing a block. The best-paying payments fill an empty block from the bottom up until the block space is full. A block is a batch of payments added to the shared record, and its space is limited. A miner is a computer competing to add the next block. Payments left behind wait for a later block.
0:21The hard part: find the winning number. A hash is a fingerprint of the block: change one number and it changes completely. Miners keep changing one number and checking the fingerprint, guess after guess, until it starts with enough zeros.
0:29A miner finds the winning number. The real winning number for block #971,868 was 2,906,539,527, giving a fingerprint with 20 leading zeros. This is proof of work: lots of guesses until one wins. It is hard to find and easy for anyone to check.
0:32Mined, and the reward. The block is sealed and stamped: block 971,868, mined Oct 17, 2026, 9:30 AM ET. The miner received 3.125 BTC of new bitcoin plus 0.01526261 BTC in fees, 3.14026261 BTC in total, taken from the block's real payout.
0:35Snap. It joins the chain. The block snaps onto the chain and the older blocks slide over one place. Block #971,868 is now on the chain.
0:37Each block locks onto the one before. A blockchain is blocks linked in order. Changing an old block would mean redoing its work and the work of every block after it. Meanwhile the waiting room keeps filling, and the next block starts now.

The squares are a sample drawn from real fee data, and the guesses before the winning number are illustrations. The block, its winning number, its fingerprint and its reward are real. Data snapshot taken Oct 17, 2026, 10:02 AM ET.

Part 1

Why Bitcoin

The ideas: money, sound money, and why a scarce, neutral, open network matters.

Module 1
~45 min
What is money?
Explain what makes good money and why fiat inflation leads people to Bitcoin
Basics
- Barter and why money emerges
- Properties of good money: scarce, durable, portable, divisible, verifiable, hard to censor
- Fiat money, money printing and inflation
+ Go deeper (optional)
- Stock-to-flow as a scarcity lens (and its limits)
- Why those problems lead to Bitcoin
+ Advanced (optional)
- Monetary history case studies: gold standard, Bretton Woods, 1971
Coming soon
Module 2
~45 min
Cypherpunk and Austrian roots
Trace the ideas and earlier attempts that Bitcoin builds on
Basics
- Who the cypherpunks were
- Austrian economics and sound money
+ Go deeper (optional)
- DigiCash, b-money, Hashcash and bit gold
- The Cypherpunk Manifesto
+ Advanced (optional)
- Reading Satoshi's 2008 announcement and the first replies on the list
Primary sources: Satoshi's announcement, cryptography mailing list (Oct 31, 2008) · A Cypherpunk's Manifesto (Hughes, 1993) · b-money (Wei Dai, 1998)
Coming soon
Module 3
~60 min
Satoshi's white paper
Read the white paper and summarize the problem it solves
Basics
- The double-spend problem without a trusted third party
- Proof-of-work and the longest chain
+ Go deeper (optional)
- Incentives for honest miners
- Simplified payment verification
+ Advanced (optional)
- Section 11: the attacker probability math
Primary sources: Bitcoin white paper (Nakamoto, 2008)
Coming soon
Module 4
~45 min
Sovereignty, decentralization and “Bitcoin, not crypto”
Explain why decentralization, censorship resistance and a fixed supply set Bitcoin apart
Basics
- Self-sovereignty and censorship resistance
- Why decentralization matters
+ Go deeper (optional)
- Who controls the rules: users, nodes, miners
- Bitcoin vs other tokens: issuance, founders, premines and trust
+ Advanced (optional)
- Soft forks, activation and governance debates
Primary sources: Bitcoin white paper (Nakamoto, 2008)
Coming soon

Part 2

How it works

Deep dives into issuance, difficulty, mining and the security budget, straight from the source code and BIPs.

Module 5
~60 min
Inside a block
Read a block header field by field and trace a transaction into a block
▶ Watch: How a Bitcoin block gets built (45 s)
Basics
- Transactions, the mempool and fees
- Using a block explorer
+ Go deeper (optional)
- The six header fields: version, previous hash, Merkle root, time, bits, nonce
- Merkle trees and why SPV works
+ Advanced (optional)
- The 80-byte header and why the nonce alone is not enough (extranonce)
+ Things most Bitcoiners don't know
- A block header has exactly six fields; the nonce is only 32 bits, so miners also vary an extranonce in the coinbase to get fresh Merkle roots. [Bitcoin Core primitives/block.h]
Primary sources: Bitcoin white paper (Nakamoto, 2008) · Bitcoin Core primitives/block.h
Coming soon
Module 6
~75 min
Issuance and the halving
Derive the supply schedule from the code and explain why it is not exactly 21M
Basics
- Block subsidy starts at 50 BTC
- Halving every 210,000 blocks (roughly four years)
+ Go deeper (optional)
- Integer satoshi math and the right-shift halving in GetBlockSubsidy
- Why the cap is 20,999,999.9769 BTC
- The last subsidy at block 6,929,999 (around 2140)
+ Advanced (optional)
- Unspendable coins: the genesis coinbase, the BIP30 duplicate coinbases (blocks 91842 and 91880), provable burns, underclaimed rewards
+ Things most Bitcoiners don't know
- The halving is a bit shift (nSubsidy >>= halvings) on an integer number of satoshis, so fractions are dropped; summing all 33 non-zero eras gives 20,999,999.9769 BTC, not 21M. [Bitcoin Core validation.cpp, GetBlockSubsidy]
- The genesis block's 50 BTC can never be spent: Core skips connecting its transactions. [Bitcoin Core validation.cpp, ConnectBlock genesis special case]
- Blocks 91842 and 91880 reused earlier coinbase transaction IDs, overwriting those outputs; BIP30 was added so it can't happen again. [BIP30: duplicate transactions]
- Consensus only rejects a coinbase that claims too much. Claiming less is valid, and the unclaimed satoshis never come into existence. [Bitcoin Core validation.cpp, GetBlockSubsidy]
Primary sources: Bitcoin Core validation.cpp, GetBlockSubsidy · Bitcoin Core chainparams.cpp (nSubsidyHalvingInterval = 210000) · BIP30: duplicate transactions
Coming soon
Module 7
~75 min
The difficulty adjustment
Explain exactly how and when difficulty retargets and where the edge cases are
Basics
- Target of one block every 10 minutes
- Retarget every 2016 blocks
+ Go deeper (optional)
- The 4x clamp in either direction
- Median-time-past and the 2-hour future limit
- Testnet's 20-minute minimum-difficulty rule
+ Advanced (optional)
- The off-by-one: only 2015 intervals are measured
- The timewarp vulnerability and the Consensus Cleanup (BIP54)
+ Things most Bitcoiners don't know
- Retargeting measures from block N-2015 to block N, so it times 2015 intervals, not 2016. That off-by-one is part of what makes the timewarp attack possible. [Bitcoin Core pow.cpp]
- Each retarget is clamped: the measured timespan is limited to between 1/4 and 4x the two-week target. [Bitcoin Core pow.cpp]
- A block can't be more than 2 hours ahead of network-adjusted time, and must be later than the median of the previous 11 blocks. [Bitcoin Core chain.h (MAX_FUTURE_BLOCK_TIME)]
- On testnet, if no block arrives within 20 minutes, a minimum-difficulty block is allowed. [Bitcoin Core pow.cpp]
- BIP54 fixes timewarp by requiring the first block of each period to be no more than 7200 seconds earlier than the block before it. [BIP54: Consensus Cleanup]
Primary sources: Bitcoin Core pow.cpp · Bitcoin Core chain.h (MAX_FUTURE_BLOCK_TIME) · BIP113: median time-past · BIP54: Consensus Cleanup
Coming soon
Module 8
~60 min
Mining: hardware, pools and protocols
Follow a hash from an ASIC to a pool payout
Basics
- What mining actually does
- Why pools exist
+ Go deeper (optional)
- Coinbase maturity: 100 blocks before a reward can be spent
- Stratum v1 vs Stratum v2 (job negotiation)
- Pool payout schemes: FPPS vs PPLNS
+ Advanced (optional)
- Selfish mining and the 1/3 hash-power result
+ Things most Bitcoiners don't know
- A coinbase output can't be spent until it has 100 confirmations. [Bitcoin Core consensus.h (COINBASE_MATURITY)]
- Stratum v2 lets miners choose their own block templates instead of the pool choosing for them. [Stratum V2 specification]
- Selfish mining can pay off with less than half the hash power; the paper shows a threshold as low as 1/3 under its assumptions. [Eyal & Sirer, Majority is not Enough (2013)]
Primary sources: Bitcoin Core consensus.h (COINBASE_MATURITY) · Stratum V2 specification · Eyal & Sirer, Majority is not Enough (2013)
Coming soon
Module 9
~45 min
Mining economics and the security budget
Reason about who pays for security as the subsidy shrinks
Basics
- Subsidy plus fees = miner revenue
- Hashprice: revenue per unit of hash rate
+ Go deeper (optional)
- The fee market after the subsidy ends
- How halvings hit miner margins
+ Advanced (optional)
- The security budget debate: arguments on each side
+ Things most Bitcoiners don't know
- Miner revenue per block is capped at subsidy plus fees; consensus enforces it directly. [Bitcoin Core validation.cpp, GetBlockSubsidy]
- Satoshi expected fees to take over once the coins were issued (white paper section 6). [Bitcoin white paper (Nakamoto, 2008)]
Primary sources: Bitcoin white paper (Nakamoto, 2008) · Bitcoin Core validation.cpp, GetBlockSubsidy
Coming soon
Module 10
~45 min
Don't trust, verify: nodes
Know what a full node checks and how to run or connect to one
Basics
- What a full node verifies
- Nodes vs miners: who enforces the rules
+ Go deeper (optional)
- Running your own node or using Tor (see Start9 and alternatives)
- Verifying software downloads
+ Advanced (optional)
- Reproducible builds and signed releases
Coming soon
Module 11
~60 min
Keys, seeds and your threat model
A written one-page threat model
Basics
- Keys, seeds and exchange IOUs
- Golden rules: never type or photograph your seed, verify on device, test first
+ Go deeper (optional)
- Theft vs loss failure modes
- Choosing tier A, B or C
+ Advanced (optional)
- How a seed becomes keys (BIP39 and BIP32)
Coming soon

Part 3

Self-custody hands-on

Hold your own keys: a hardware wallet, a durable backup and safe habits.

Module 12
~65 min
Your first hardware wallet setup →
A device bought direct, verified, and a successful test transaction
Basics
- All 8 built lessons: buy direct, inspect, verify firmware, initialize offline, write and verify the seed, check addresses on the device, send a test transaction
+ Go deeper (optional)
- (to write) Attestation checks and what they prove
- (to write) Verifying the companion app download (PGP + SHA-256)
+ Advanced (optional)
- (to write) Full wipe-and-restore test with fingerprint and first-address match
Available now
Module 13
~90 min
Seed phrase backup
A durable, tested backup and an optional passphrase
Basics
- Paper to metal
- No digital copies
+ Go deeper (optional)
- Passphrase rules and fingerprints
+ Advanced (optional)
- Why firmware updates don't fix old seeds
Coming soon
Module 14
~60 min
Sending, receiving and scams
Safe habits and a scam-spotting reflex
Basics
- Fresh addresses and on-device verification
- Seed requests and fake firmware
+ Go deeper (optional)
- Fees, RBF and CPFP
- Address poisoning
+ Advanced (optional)
- Tampered devices and supply-chain attacks
Coming soon
Module 15
~45 min
Moving off an exchange safely
Your coins in your custody, moved in stages
Basics
- Lock down the exchange account (non-SMS 2FA, allowlists)
- Test withdrawal, then tranches
+ Go deeper (optional)
- Record keeping for taxes
+ Advanced (optional)
- Migrating after a wallet advisory
Coming soon

Part 4

Advanced

Multisig, inheritance, privacy and verifying everything you built.

Module 16
~120 min
Multisig basics
A 2-of-3 multisig wallet plus descriptor backups
Basics
- What 2-of-3 means
- Vendor diversity
+ Go deeper (optional)
- Descriptor backups with every seed
- On-device registration
+ Advanced (optional)
- Test every key pair
Coming soon
Module 17
~60 min
Inheritance
A completed letter of instruction
Basics
- A map, not the treasure
+ Go deeper (optional)
- Roles and a technical helper
- Legal layer with an estate attorney
+ Advanced (optional)
- Rehearse with a test wallet
Coming soon
Module 18
~60 min
Privacy
A privacy routine that fits your threat model
Basics
- Address reuse
- KYC vs non-KYC sources
+ Go deeper (optional)
- Coin control and labeling
- Own node or Tor for wallet lookups
+ Advanced (optional)
- Chain analysis heuristics
Coming soon
Module 19
~60 min
Verifying your setup and recovery drills
A passed recovery drill and a yearly calendar
Basics
- Monthly, quarterly and yearly routine
+ Go deeper (optional)
- Yearly recovery drill on a device
- Incident playbook
+ Advanced (optional)
- Device advisory vs seed advisory
Coming soon

Work through it with the printable checklists and the Security Status tracker. New modules are added as they're ready; get the brief to hear when one is published.