If Satoshi Nakamoto had a big gripe it was this: commerce on the internet relied entirely on financial institutions to process electronic payments. What was needed was an electronic cash system that would allow people to bypass third parties and transact directly with each. Enter bitcoin.

Nakamoto’s endeavor was wildly successful, but in an unintended way. Rather than being adopted as an online payments platform, bitcoin ended up as an incredibly popular peer-to-peer gambling product. The very feature that makes bitcoin so exciting and addictive—its volatility—has prevented it from becoming a serious alternative to payment networks like Visa or PayPal.

But Nakamoto’s dream of electronic cash isn’t dead. A second attempt is being made with stablecoins. Like bitcoin, stablecoins are run on a decentralized ledger. But whereas bitcoin’s price has no mooring, stablecoins are pegged to a fiat currency—usually the dollar. This peg can be achieved in several ways, the most common of which is for issuers to redeem tokens at par with funds held at a bank.

Having solved bitcoin’s volatility problem, will stablecoins go on to conquer online commerce? Long before they start challenging Visa or PayPal’s share of online retail payments, some important issues will have to be dealt with.

The first generation of stablecoins—which included BitUSD, Nubits, and Tether—debuted in 2014-2015. These early issuers should have been obliged to get licensed as regulated money service businesses. But they were too small to attract the attention of the authorities. The newest generation of stablecoins has proactively sought regulatory approval, which is why they are known as “regulated” stablecoins. These include Circle’s USDC, the Winklevoss’s Gemini dollar, Trust Token’s TrueUSD, and the Paxos Standard.

For now, the use of stablecoins is confined to a closed loop that includes cryptocurrency exchanges, speculators, and stablecoin issuers. Rather than going through the hassle of providing actual fiat accounts to customers, exchanges often choose to “simulate” a fiat account by allowing customers to house stablecoins on the exchange where they can trade them for other cryptocurrencies.

These new entrants aren’t shy about their larger ambitions. TrueUSD claims to have solved the problem of buying coffee with a cryptocurrency, and Paxos’s white paper says that in the future “large populations of consumers may look to Paxos Standard to serve as their primary currency.”

Before people start to buy stuff with stablecoins, a number of technical issues must be addressed. Most regulated stablecoins are built on top of the Ethereum blockchain. But if Ethereum gets too slow and expensive, then stablecoins won’t be able to serve as a legitimate way to make online payments. Reliability is another problem. The recent 51% attack on Ethereum Classic shows how easy it is to re-organize transaction histories. Will large retailers want to risk accepting stablecoins if there is a chance that payments can be unwound?

Even if these technical problems are solved, there is a more existential question: what genuine advantages do stablecoins offer that other types of payments don’t?

The first page of Nakamoto’s white paper is an ode to non-reversibility. Companies like Visa can’t avoid mediating disputes between buyers and sellers, wrote Nakamoto. When the seller is deemed at fault, these transactions are reversed. Mediation is expensive and increases the cost of making a payment. Since bitcoin payments are non-reversible, and therefore cannot be mediated, it opens the door to cheap online payments.

Bitcoin’s irreversibility is modeled after banknotes. You can ask your card network to reverse a $20 payment, sometimes months after it has been made. But if you’ve bought something with a $20 bill, it’s just not possible to phone up the issuer of that note—the Federal Reserve—to cancel the payment. Merchants appreciate the certainty of cash.

Stablecoins could differentiate themselves from incumbent card networks by bringing the non-reversibility of banknotes to online commerce. Unlike central banks, however, stablecoin issuers may not be able to achieve sufficient separation from the coins they issue, in which case they won’t be able to avoid mediation and reversibility.

For instance, to comply with anti-money laundering laws, regulated stablecoin issuers like Paxos have already been obliged to include code that gives them power to freeze coins. If enough crooked merchants were to abuse non-reversibility, say by selling non-existent products, then surely lawmakers will pressure stablecoin issuers to guard against regular fraud too. If so, issuers like Paxos and Gemini will have to set up processes for determining if transactions are legitimate. Source code will have to be modified to allow for transaction reversals.

But if stablecoin transactions no longer have the quality of a banknote transaction, then they cease to have the advantages Nakamoto advocated. They would be just one reversible payment option among many.

Even if stablecoins are able to flawlessly execute non-reversible internet payments, adoption isn’t a slam dunk. A payments network needs to solve a chicken-and-egg problem. Sellers will provide a given payments option only if consumers have adopted it. But this option will be adopted by consumers only if sellers offer it.

Gaining merchant adoption should be relatively easy for stablecoins. If stablecoins are cheaper to process than card payments,merchants can increase their profits each time customers come to the cashier. But what about the consumer? No shopper is going to want to pay their $200 Amazon bill with stablecoins given that their Visa card settles the same $200 bill while also providing a dispute mechanism and points.

To encourage consumers to adopt stablecoins, merchants will have to share some of their cost savings with customers. One way to do this would be to offer a 1 percent discount on any purchase made with stablecoins. Another way would be to surcharge all card payments while allowing for free stablecoin purchases.

Offering these sorts of incentives isn’t always possible. In Canada, where I live card companies have inserted legalese into their terms of service that prevent merchants from surcharging credit cards. In the U.S., states like Massachusetts disallow surcharging altogether. Europe also prohibits the practice.

As for discounts, they may not be a very effective way to encourage consumer adoption. Behavioral finance theorists have found that people don’t treat gains and losses symmetrically: the pain of a $1 loss outweighs the joy of a $1 gain. Since a 2 percent discount on stablecoin purchases would be treated as a gain, it may not be as effective as a 2 percent surcharge on cards, which would be treated as a loss. Ikea found that when it levied a credit card surcharge on Brits from 2004-2010, it was successful in steering shoppers to cheaper options like debit card payments. But Ikea’s attempts to use discounts were ineffective.

Stablecoin issuers will have to devise some way to solve the tricky chicken-and-egg problem while at the same time negotiating the complicated politics of discounting and surcharging.

Satoshi Nakamoto designed bitcoin as way to buy and sell online without relying on financial institutions as trusted third parties. Stablecoins may stand a better chance than bitcoin in addressing Nakamoto’s criticism of online commerce, but they have a long way to go.