What is it for?
The previous lessons showed how blockchains work. This one asks what they are good for, as carefully as the evidence allows. Some uses work and matter to real people, often far from the financial centres where the technology is debated. Some are claimed far more often than they are demonstrated. And the field has a long record of fraud, failure and theft, which anyone working in it should know as well as its successes.
This lesson informs; it does not sell. It contains no price talk and no investment advice, and nothing here recommends buying, selling or holding any asset.
Payments and remittances
Sending money home is one of the largest financial flows reaching low- and middle-income countries, and it is expensive. The World Bank's Remittance Prices Worldwide has put the global average cost of sending the equivalent of 200 US dollars at around 6 to 6.5 per cent in 2024 and 2025, roughly double the 3 per cent target in the United Nations' Sustainable Development Goals, with corridors into sub-Saharan Africa costing more.
A blockchain transfer can cross a border in minutes for a small network fee. But the cost a family pays is rarely just the network fee: money must be converted into a token (an on-ramp) and back into local currency (an off-ramp), each with fees and exchange-rate margins, and recipients need a phone, a wallet and somewhere to cash out. Stablecoins (next section) remove the volatility, and in several countries with high inflation or capital controls, journalists and researchers have reported people holding dollar stablecoins as savings. That use is real, and so are its risks: an issuer's solvency, frozen accounts, and lost keys.
The fairest comparison is with fast payment systems built without blockchains. Kenya's M-Pesa (2007) brought mobile money to millions without bank accounts; India's UPI (2016) and Brazil's Pix (2020) made instant, near-free domestic payments ordinary, run by regulated, trusted operators. Across borders, the remaining cost is mostly compliance, currency exchange and lack of competition, not the speed of the ledger.
Stablecoins
Definition— stablecoin
A stablecoin is a token designed to keep a stable value against a reference asset, usually the US dollar. By how the peg is defended:
- Fiat-backed: an issuer holds reserves (bank deposits, short-term government bills) and promises to redeem tokens one for one. The largest, Tether's USDT and Circle's USDC, are of this kind.
- Crypto-collateralised: tokens are minted against more than their value in other crypto assets locked in contracts, which are sold off (liquidated) automatically if their value falls too far. DAI, from the MakerDAO project (now rebranded Sky), is the best-known example.
- Algorithmic: the peg is defended by a mechanism rather than by assets of equal value held outside the system, typically by letting holders swap one stablecoin for one dollar's worth of a companion token that the protocol can mint.
Each kind fails differently. A fiat-backed coin is only as good as its reserves, its issuer's honesty and its banks: in 2021 a US regulator fined Tether for misstatements about its reserves, and in March 2023 USDC fell well below one dollar for about two days after Circle disclosed that 3.3 billion dollars of its reserves were at the failed Silicon Valley Bank. A crypto-collateralised coin depends on liquidations working in a crash: on 12 March 2020, amid a price collapse and network congestion, some MakerDAO collateral auctions closed at bids of zero.
Algorithmic designs failed most dramatically. TerraUSD (UST) kept its peg by letting anyone burn 1 UST for 1 dollar's worth of the token LUNA, or the reverse. A lending service on the same chain, Anchor, paid close to 20 per cent a year on UST deposits, which attracted billions. In May 2022, large withdrawals pushed UST below its peg; holders rushed to redeem for LUNA, which multiplied LUNA's supply and crushed its price, which made the redemptions worth less, which deepened the run. Within days both tokens had lost nearly all their value, wiping out about 40 billion US dollars of market value and the savings of many ordinary users across the world. Terra's founder, Do Kwon, pleaded guilty to fraud charges in a US court in August 2025 and was sentenced in December 2025 to 15 years in prison.
Regulators have since written rules for stablecoins specifically, in the EU and the United States among others (see the regulation section below).
Decentralised exchanges: constant-product market makers
Decentralised finance (DeFi) means financial services built as smart contracts: exchanges, lending pools, derivatives. Its most original invention is the automated market maker (AMM). A traditional exchange matches buyers with sellers in an order book. An AMM instead holds reserves of two tokens in a contract and quotes prices from a formula, so anyone can trade against it at any time, and anyone can supply the reserves in return for a share of the fees.
Definition— constant-product market maker
A pool holds reserves of token X and of token Y. A trade that adds of X (of which a fraction is kept as a fee) may remove of Y only if the product of the reserves does not fall: The pool's marginal price of X, in Y, is . This is the design of Uniswap's first two versions (2018 and 2020), with .
The program implements the swap and measures two effects. Price impact: large trades get worse prices, because each unit sold moves the price further. Fee accrual: the fee portion stays in the pool, so creeps upwards and liquidity providers' share of the reserves grows.
class ConstantProductPool:
"""Reserves x of token X and y of token Y; trades keep x * y from falling (fees make it rise)."""
def __init__(self, x, y, fee=0.003):
self.x, self.y, self.fee = x, y, fee
def price(self): # marginal price of X, in Y per X
return self.y / self.x
def quote_x_for_y(self, dx):
"""How much Y you receive for selling dx of X, leaving (x + dx)(y - dy) >= x * y."""
dx_after_fee = dx * (1 - self.fee)
return self.y * dx_after_fee / (self.x + dx_after_fee)
def sell_x(self, dx):
dy = self.quote_x_for_y(dx)
self.x, self.y = self.x + dx, self.y - dy
return dy
# A pool of 500 ETH-like tokens (X) and 1,000,000 dollar-like tokens (Y): price 2000 Y per X.
print(" sell X share of pool Y received avg price price impact pool price after")
for dx in [0.5, 5, 25, 50, 125, 500]:
pool = ConstantProductPool(500, 1_000_000)
before = pool.price()
dy = pool.sell_x(dx)
avg = dy / dx
print(f"{dx:7} {dx / 500:12.1%} {dy:10,.0f} {avg:9,.1f} {1 - avg / before:11.2%}"
f" {pool.price():15,.1f}")
pool = ConstantProductPool(500, 1_000_000)
k0 = pool.x * pool.y
for _ in range(100): # 100 round trips: sell X, then buy it back by selling Y
dy = pool.sell_x(10)
dx = dy * (1 - pool.fee) * pool.x / (pool.y + dy * (1 - pool.fee))
pool.x, pool.y = pool.x - dx, pool.y + dy
print(f"after 100 round trips, k has grown by {pool.x * pool.y / k0 - 1:.2%}: fees accrue to liquidity providers")A trade of 0.1 % of the pool costs 0.40 % relative to the starting price (the 0.3 % fee plus about 0.1 % of impact); a trade of 10 % of the pool costs 9.34 %; selling as much X as the pool holds gets barely half the starting price. That is why large trades are split across pools and over time, and why the sandwich attacks of Smart-contract security pay.
Liquidity providers face a subtler cost. When the outside market price moves, arbitrageurs trade with the pool until its price matches, and those trades always leave the pool holding more of whichever token has fallen in value. Compared with simply holding the original deposit, the provider loses; this is called impermanent loss, because it disappears if the price returns to where it started. For a fee-free pool and a price change by a factor , the provider's position is worth times the held deposit.
from math import sqrt
def lp_versus_hold(r, x0=100.0, p0=2000.0):
"""Deposit x0 of X and x0*p0 of Y into a fee-free pool. The outside price moves to r*p0 and
arbitrageurs trade until the pool agrees. Return (value of the LP share, value of just holding)."""
k = x0 * (x0 * p0)
p1 = r * p0
x1, y1 = sqrt(k / p1), sqrt(k * p1) # the only reserves with x * y = k and y / x = p1
assert abs(x1 * y1 - k) < 1e-6 * k and abs(y1 / x1 - p1) < 1e-9 * p1
return x1 * p1 + y1, x0 * p1 + x0 * p0 # both valued at the new price, in Y
print(" price moves by LP / hold formula 2*sqrt(r)/(1+r) shortfall in Y (deposit 100 X + 200,000 Y)")
for r in [0.25, 0.5, 0.8, 1.0, 1.25, 2.0, 4.0, 10.0]:
lp, hold = lp_versus_hold(r)
print(f"{r:13}x {lp / hold:9.4f} {2 * sqrt(r) / (1 + r):23.4f} {hold - lp:12,.0f}")Research has sharpened this: Milionis, Moallemi, Roughgarden and Zhang (2022) measure the loss to arbitrageurs continuously, as loss-versus-rebalancing, which does not vanish when prices return. Later designs concentrate liquidity in a price range (Uniswap v3, 2021) or use curves for tokens that should trade near parity (Curve). Lending protocols such as Aave and Compound, which lend against over-collateralised deposits and liquidate automatically, are DeFi's other pillar, and every risk of the previous two lessons applies to all of it.
Tokenisation of real-world assets
Tokenisation represents claims on conventional assets, such as government bonds, money-market funds, real estate or invoices, as tokens on a ledger. Banks, asset managers and central banks have run serious projects: the European Investment Bank issued a digital bond on Ethereum in 2021; Franklin Templeton has run a US government money-market fund recorded on public blockchains since 2021; BlackRock launched a tokenised fund on Ethereum in 2024; and central banks have run joint experiments, such as the Monetary Authority of Singapore's Project Guardian and the Bank for International Settlements' Project Agorá. The promised benefits are faster settlement, fewer reconciliations between institutions, round-the-clock transfers and programmable compliance.
Be precise about what a token is. It records whom the issuer considers entitled to an asset; the asset itself, and the legal right to it, live off-chain, with a custodian or in a fund's register, under a country's law. If the issuer, custodian or legal structure fails, the token is a well-kept record of a broken promise. Most tokenised securities are also permissioned: only verified investors may hold them, and the issuer can freeze or reverse transfers, as the law requires. Whether shared ledgers beat well-run conventional databases for these markets is an open, empirical question, and many projects remain pilots.
Supply chains and identity: claims weighed sceptically
Blockchains have been proposed for tracking food, medicines, minerals and shipping containers. The recurring difficulty is the oracle problem in physical form: a ledger can prove that a record has not changed since it was written, not that it was true when written. If a supplier mislabels a crate, the blockchain preserves the false label perfectly; traceability depends on inspections and incentives where goods change hands, whatever the records are kept in. TradeLens, a shipping platform built on a permissioned blockchain by the shipping company Maersk with IBM and launched in 2018, was discontinued after its operators announced in 2022 that it had not reached commercial viability; commentators widely attributed this to competitors' reluctance to join a platform run by a rival. The hard part was governance, not the ledger.
Identity is similar. The W3C's standards for decentralised identifiers and verifiable credentials let an issuer, such as a university or a government, sign a credential that its holder stores and presents, and that anyone can verify without contacting the issuer. The cryptography is ordinary digital signatures (Signatures and keys); a blockchain is one possible place to publish issuers' keys and revocation lists, not a requirement, and many deployments, including the European Union's digital identity wallets, do not use one. Putting personal data itself on an immutable public ledger conflicts with privacy law and with common sense.
Central bank digital currencies
A central bank digital currency (CBDC) is a digital liability of a central bank, like a banknote in electronic form, available to the public (retail) or to banks (wholesale). Despite frequent confusion, most CBDCs are not public blockchains; they are run by central banks, often through commercial banks and payment providers, on centralised or permissioned systems. The Atlantic Council's tracker follows well over a hundred countries' work. A few, stated carefully:
- The Bahamas launched the Sand Dollar in October 2020, the first nationwide retail CBDC, partly to reach residents of its many islands who have little access to bank branches.
- Nigeria launched the eNaira in October 2021. Adoption remained low despite promotion, and it has been widely studied as a lesson in how hard it is to persuade people to switch from cash and existing mobile payments.
- China's e-CNY has been piloted since 2020 in a growing number of cities, distributed through commercial banks, and is among the largest trials by number of users and transactions.
- India has piloted a digital rupee for banks and for retail users since late 2022.
- The euro area: the European Central Bank ran a preparation phase for a digital euro from November 2023 to October 2025, and then decided to continue to the next phase, aiming to be ready for a possible first issuance in 2029 if EU legislation is adopted; as of 2026 that legislation was still under negotiation, and the ECB says no decision to issue has been taken.
- The United States: an executive order of January 2025 prohibited federal agencies from establishing a CBDC.
The debates are about privacy (a central bank could in principle see every payment, unlike with cash), the risk that people move savings out of commercial banks in a crisis (hence proposed holding limits), resilience, and whether the private sector already provides what is needed.
DAOs
A decentralised autonomous organisation (DAO) manages shared funds or a protocol through on-chain voting, usually weighted by governance tokens. DAOs govern several large DeFi protocols and fund public goods. Their weaknesses are documented: voting power is often concentrated in a few large holders, turnout is low, and voting by token balance invites attack. In April 2022 an attacker used a flash loan to borrow enough of the Beanstalk protocol's governance tokens to pass a malicious proposal in a single transaction, taking about 182 million US dollars of the protocol's funds. And a DAO's legal status is unsettled: in 2023 a US court held one DAO liable as an unincorporated association, exposing its token-holding voters to potential liability. "Decentralised" describes a voting mechanism, not an absence of responsibility.
Failures: exchanges, frauds and collapses
The largest losses to ordinary users have come not from broken cryptography but from custodians: companies that held customers' coins and failed them.
- Mt. Gox (Tokyo) was at one time the largest bitcoin exchange in the world. In February 2014 it halted withdrawals and filed for bankruptcy, having lost about 850,000 bitcoins, of which roughly 200,000 were later found. The losses had accumulated over years of theft and poor record-keeping. Creditors began to be repaid only in 2024, a decade later.
- Coincheck (Tokyo) lost tokens worth about 58 billion yen to hackers in January 2018, after keeping them in an internet-connected "hot" wallet; Japan's regulators, who had introduced registration for exchanges in 2017, tightened supervision afterwards.
- FTX, one of the largest exchanges, collapsed in November 2022 when it emerged that customer deposits had been used by its affiliated trading firm, Alameda Research. Its founder, Sam Bankman-Fried, was convicted of fraud and conspiracy in November 2023 and sentenced to 25 years in prison in March 2024. The bankruptcy estate began repaying customers in 2025, in dollar amounts fixed at the date of the collapse.
- In the same year the lenders Celsius and Voyager, which had promised high yields on customers' deposits, went bankrupt.
The common thread is old-fashioned: customers trusted an intermediary with no effective audit and no separation of client funds. A blockchain does not help users who hand their keys to such an intermediary, though self-custody brings its own risks of loss and theft.
Energy and environment
Proof of work converts electricity into security by design (Proof of work). The Cambridge Bitcoin Electricity Consumption Index puts Bitcoin's annual electricity use at a level comparable to a medium-sized country's, with wide uncertainty about the energy mix, and researchers have also estimated substantial electronic waste from short-lived mining machines. Supporters point to mining's use of stranded or surplus energy and its ability to switch off at peak demand; critics point to emissions where mining runs on fossil fuels. The numbers change with prices, hardware and policy, so cite current estimates with their error bars. Proof of stake uses orders of magnitude less: Ethereum's switch in September 2022 cut its electricity use by more than 99.9 %, by the estimates of researchers commissioned to measure it.
Regulation as of 2026
Regulation differs by country and changes quickly; what follows is limited to the points that are well established as of 2026.
- International standards. The Financial Action Task Force (FATF), which sets anti-money-laundering standards for its member jurisdictions, has since 2019 required countries to license or register virtual asset service providers (exchanges, custodial wallets) and to apply the travel rule: when a provider sends a transfer, information identifying the sender and the recipient must travel with it to the receiving provider, as it does for bank wire transfers. FATF's own reviews have found implementation uneven across countries.
- The European Union. The Markets in Crypto-Assets regulation, MiCA (Regulation (EU) 2023/1114), is the most comprehensive regime so far. Its rules for stablecoins apply from 30 June 2024, and the rest, including the authorisation of crypto-asset service providers and rules on white papers, market abuse and marketing, from 30 December 2024, with transitional periods for existing firms ending by July 2026 at the latest. A companion regulation applies the travel rule to crypto transfers from the same date.
- The United States enacted the GENIUS Act for payment stablecoins in July 2025, requiring full reserves in cash and short-term government securities and regular disclosure; other questions, such as which tokens are securities, have been shaped by agency action and court cases and remain contested.
- Japan has required crypto exchanges to register with its Financial Services Agency since 2017.
- China banned crypto trading and mining on the mainland in 2021.
- El Salvador made bitcoin legal tender in 2021, and in 2025, as part of an agreement with the International Monetary Fund, amended the law so that businesses are no longer required to accept it.
Two principles help when reading any rule. Regulation mostly attaches to intermediaries, those who hold customers' assets or arrange their trades, because that is where customers can be harmed and where a regulator can act. And whether something is regulated depends on what it does, not on whether it uses a blockchain.
Fraud: how to recognise a scam
Fraud is the field's largest harm to ordinary people. The FBI's Internet Crime Complaint Center has reported losses in the billions of US dollars a year from complaints involving cryptocurrency, with investment fraud the costliest category, and those are only the cases reported in one country. Common forms:
- Relationship investment scams ("pig butchering"): a stranger builds a friendship or romance over weeks, then introduces an investment platform that shows steadily growing profits, which are fictional; withdrawals require ever more "fees" or "taxes". The United Nations Office on Drugs and Crime has documented scam centres in Southeast Asia where many of the workers are themselves victims of human trafficking.
- Ponzi schemes dressed as crypto. OneCoin, which raised billions of dollars from people in many countries from 2014 onwards, had no real blockchain at all; one of its founders was added to the FBI's Ten Most Wanted list in 2022.
- Rug pulls and pump-and-dumps: a token is promoted heavily, then its creators sell out.
- Impersonation: fake support staff, deepfake celebrity endorsements, fake exchanges and wallet apps; and recovery scams, which offer people who have lost money to get it back, for a fee.
The warning signs are consistent and worth learning by heart:
- Guaranteed or unusually high returns. Nobody can guarantee returns; high, steady yields with no visible source are the signature of a Ponzi scheme.
- Unsolicited contact from a stranger, especially one who moves the conversation to a private messaging app.
- Urgency and secrecy: act now, tell no one, your bank will not understand.
- Requests for your seed phrase, private key or remote access to your device. Never give them, to anyone.
- Fees to withdraw your own money, or a platform that shows profits you cannot take out.
- Payment in crypto or gift cards demanded by someone claiming to be a government agency, a utility or a company.
If you are a victim, stop paying, keep records, and report to the police and to national fraud services (in the United States, IC3); exchanges can sometimes freeze stolen funds if told quickly. Victims are not foolish: these operations are professional, patient and psychologically skilled, and they target lonely and trusting people deliberately.
Optional Christian reflection · Just weights
A reflection from the Christian tradition that inspires Lumenary Academy. It is not part of the mathematics and you are welcome to skip it.
Blockchains are sometimes called "trustless", but every lesson in this level found people underneath: developers who write the code, auditors who check it, key holders who sign upgrades, exchanges that hold deposits, issuers who promise reserves. The technology can make some promises checkable; it cannot make people honest. An old proverb puts the standard simply: "A false balance is abomination to the LORD: but a just weight is his delight" (Proverbs 11:1, King James Version). Honest reserves, accurate disclosures and code that does what it says are just weights in modern form. And the words that follow the parable of the unjust steward, "he that is faithful in that which is least is faithful also in much" (Luke 16:10), fit an industry where small shortcuts with other people's money have repeatedly grown into catastrophes. Whatever one's beliefs, those who build or advise in this field are stewards of other people's savings and trust, and are rightly judged by how they care for them.
Common misconceptions
Misconception
"A stablecoin is as safe as a dollar in the bank." A fiat-backed stablecoin is a claim on a private issuer, with whatever reserves, legal protections and redemption rights its terms and regulators provide, which differ from deposit insurance and differ between issuers. Algorithmic designs have no reserves of equal value at all.
Misconception
"Impermanent loss doesn't matter because it is impermanent." It disappears only if the price returns exactly to where it was when you deposited. Prices often do not, and loss-versus-rebalancing, the value lost to arbitrageurs along the way, is not recovered even then.
Misconception
"Crypto fraud only catches people who are greedy or naive." Professional scams are built to defeat intelligent, cautious people, using time, trust and pressure. The defence is a set of fixed rules (never share a seed phrase, distrust unsolicited offers, never pay to withdraw), not cleverness in the moment.
Exercises
1. Applications, limits and regulation
Warm-upHumanA constant-product pool with no fee holds 100 TOK and 400 USD, so . Kemal sells 100 TOK to the pool. How many USD does he receive?
2. Applications, limits and regulation
Warm-upHumanGrace in London sends the equivalent of 200 US dollars to her brother in Accra. The service charges a fixed fee of 5 dollars and uses an exchange rate 2 % worse than the mid-market rate. What is the total cost of the transfer, in dollars?
3. Applications, limits and regulation
CoreHumanWhich statement correctly describes TerraUSD (UST) and why it collapsed in May 2022?
4. Applications, limits and regulation
CoreComputationalWrite the two functions every AMM interface needs, for a constant-product pool with reserves x (of X) and y (of Y) and a fee taken from the input:
amount_out(x, y, dx, fee): how much Y you receive for sellingdxof X (onlydx * (1 - fee)counts towards keeping constant).amount_in(x, y, dy, fee): how much X you must sell to receive exactlydyof Y. RaiseValueErrorunless0 <= dy < y.
The two must be inverses of each other, and no trade may ever reduce .
5. Applications, limits and regulation
StretchComputationalLiquidity providers own a pool through shares. Complete add_liquidity and remove_liquidity:
- The first deposit mints shares.
- Later deposits mint shares in proportion to what they add:
min(dx / x, dy / y) * total(using the reserves before the deposit), so that adding mostly one token earns credit only for the scarcer side. remove_liquidity(who, amount)burnsamountofwho's shares and returns that fraction of both reserves,(dx, dy). RaiseValueErrorifamount <= 0orwhoholds fewer shares.
The tests make traders swap back and forth so that fees accumulate, then check that a provider withdraws more than they deposited, in the sense of a larger product .
6. Applications, limits and regulation
CoreHumanWhich statements about central bank digital currencies are correct, as of 2026? Select all that apply.
7. Applications, limits and regulation
CoreHumanWhich statements about crypto regulation are correct, as of 2026? Select all that apply.
8. Applications, limits and regulation
Warm-upHumanWei met someone online three weeks ago who has become a close friend. The friend now recommends a trading platform where Wei's first small deposit already shows a 30 % profit, but to withdraw, Wei is told to pay a 15 % "tax" first. Which of these are warning signs of a scam? Select all that apply.
Further reading and resources
- Remittance Prices Worldwide · The World BankQuarterly data on the cost of sending money across borders.
- Uniswap v2 Core (whitepaper) · Hayden Adams, Noah Zinsmeister and Dan Robinson · 2020
- Automated Market Making and Loss-Versus-Rebalancing · Jason Milionis, Ciamac C. Moallemi, Tim Roughgarden and Anthony Lee Zhang · 2022A sharper measure than impermanent loss of what liquidity providers give up to arbitrageurs.
- Regulation (EU) 2023/1114 on markets in crypto-assets (MiCA)
- ESMA: Markets in Crypto-Assets Regulation
- Financial Action Task Force (FATF)Recommendation 15 and its interpretive note apply anti-money-laundering standards, including the travel rule, to virtual asset service providers.
- Central Bank Digital Currency Tracker · Atlantic Council
- European Central Bank: the digital euro
- Cambridge Bitcoin Electricity Consumption Index · Cambridge Centre for Alternative Finance
- FBI Internet Crime Complaint Center (IC3)Where to report internet fraud in the United States, and its annual reports.