Bitcoin is hot-hot-hot, but today I want to talk about how it ends. Earlier, I predicted a variety of possible fates for Bitcoin ranging from taking over the entire M1 money supply to complete collapse, but the most probable one, in my view, is that Bitcoin is eventually supplanted by one or more successor digital currencies which win in the marketplace. I think that successor will also itself be supplanted, and that this might continue for some time. I want to talk about not just why that might happen, but also how it may take place.

Nobody thinks Bitcoin is perfect, and no digital currency (DigiC) is likely to satisfy everybody. Some of the flaws are seen as flaws by most people, but many of its facets are seen as features by some, and flaws by others. The anonymity of addresses, the public nature of the transactions, the irrevocable transactions, the fixed supply, the mining system, the resistance to control by governments -- there are parties that love these and hate these.

Bitcoin's most remarkable achievement, so far, is the demonstration that a digital currency with no intrinsic value or backer/market maker can work and get a serious valuation. Bitcoin argues -- and for now demonstrates -- that you can have a money that people will accept only because they know they can get others to accept it with no reliance on a government's credit or the useful physical properties of a metal. The price of a bitcoin today is pretty clearly the result of speculative bubble investment, but that it sustains a price at all is a revelation.

Bitcoins have their value because they are scarce. That scarcity is written into the code -- in the regulated speed of mining, and in the fixed limit on coins. There will only be so many bitcoins, and this gives you confidence in their value, unlike say, Zimbabwe 100 trillion dollar notes. This fixed limit is often criticised because it will be strongly deflationary over time, and some more traditional economic theory feels there are serious problems with a deflationary currency. People resist spending it because holding it is better than spending it, among other things.

Altcoins

While bitcoins have this scarcity, digital currencies as a group do not. You can always create another digital currency. And many people have. While Bitcoin is the largest, there are many "altcoins," a few of which (such as Ripple, Litecoin and even the satirical currency Dogecoin) have serious total market capitalizations of tens or hundreds of millions of dollars(1). Some of these altcoins are simply Bitcoin or minor modifications of the Bitcoin protocol with a different blockchain or group of participants, others have more serious differences, such as alternate forms of mining. Ripple is considerably different. New Altcoins will emerge from time to time, presumably forever.

What makes one digital coin better than another? Obviously a crucial element is who will accept the coin in exchange for goods, services or other types of currency. The leading coin (Bitcoin) is accepted at more stores which gives it a competitive advantage.

If one is using digital currency simply as a medium -- changing dollars to bitcoins to immediately buy something with bitcoins at a store, then it doesn't matter a great deal which DigiC you use, or what its price is, as long as it is not extremely volatile. (You may be interested in other attributes, like speed of transaction and revocation, along with security, ease of use and other factors.) If you wish to hold the DigC you care about appreciation, inflation and deflation, as well as the risk of collapse. These factors are affected as well by the "cost" of the DigiC.

The cost of a digital currency

I will advance that every currency has a cost which affects its value. For fiat currency like dollars, all new dollars go to the government, and every newly printed dollar devalues all the other dollars, and overprinting creates clear inflation.

New bitcoins are given to miners. Miners do makework computation to demonstrate that a majority of the mining computing power agrees which transactions should be recorded in the public ledger. The miners spend a lot of electricity to do this makework calculation, and they also typically make a profit, in that the value of the new coins they get exceeds their cost of generating them. In addition to this, miners receive transaction fees on the transactions they log, though these fees are minor compared to the value of the mined coins. In the far future, miners will only receive these transaction fees -- and the fees will probably rise.

Other currencies have costs of a transaction as well, usually higher than Bitcoin's. Credit card transaction fees are well known, but large banking system transactions also have surprisingly high fees. This is an area where people hope Bitcoin will compete well with traditional currencies.

There is no free lunch. The cost -- of wasted electricity and miner's profit -- comes from somebody. It comes from the users and holders of bitcoins. (By keeping transaction fees low, it also subsidizes that cost away.) The "market cap" of bitcoin -- currently $7 billion -- was, if a real cap, paid to miners and holders/speculators. For now, most of the holders of bitcoin were miners, so one can argue the cost came from the people bearing it, but this will not be always true.

In part because of this (and just the profligate wastage of bitcoin mining) a number of the altcoins differentiate themselves on being less wasteful in their mining process.

Inter-currency exchange

There are many exchanges that will trade bitcoins for fiat money, and some, including Ripple, that will trade bitcoins with altcoins. It is possible, however, to build a system so that exchange of DigiCs can be done without exchanges. It could be done in the public ledgers (blockchains) of the DigiCs themselves.

In such a system, a transaction could be entered into the blockchain with appropriate scripting commands to indicate it is finalized only when an opposite transaction appears in the blockchain of another system. Only when both components of the transaction occur are both considered finalized. Alternate currencies, as underdogs, have incentive to support this, and there is investigation as to whether Bitcoin's own smart contract scripting language could be adapted to do this. (It may be necessary to enter a tiny extra transaction into the Bitcoin block chain to signal the existence of the counter-transaction in another system.)

The existence of fluid inter-currency exchange with no requirement for trusting a 3rd party helps facilitate competition between different systems. It also means that for merchants, accepting one major DigiC is the same as accepting all the others than can be easily converted it. It becomes possible for users to spend their brand-new currency everywhere. This is good for the marketplace and good for the world. It is bad for Bitcoin as a system, because it removes the barrier to entry altcoins have -- thanks to Bitcoin's head start, Bitcoin, we must recall, is not a company, and does not have the same incentives to promote barriers to its competition.

The endgame

Once we have a world of competing DigiCs, and free and unblockable flow of value between them through automated exchange, the competition will heat up. New altcoins will rise and they will prove popular with those that prefer their attributes. It seems unlikely that Bitcoin, the oldest, will be the long term winner. It certainly doesn't work that way in most other fields.

Bitcoin might fend of other competitors by adapting of course, incorporating their features when it is clear the market loves them. But some of the competitive differences are matters of philosophy, not simple features. Bitcoin will not change these because to change them it would not be Bitcoin. Bitcoin is also not tremendously adaptable, and major changes would require the participation of a large fraction of all miners. Cryptocurrencies with no owner are inherently conservative and resist change -- indeed that conservatism is a feature.

In time, some new altcoin will become particularly popular. This will not be instantaneous. It will grow amid much debate. It may gain its market share not by taking people away from Bitcoin, but by bringing in new users. There may be altcoins with features that nobody has yet dreamed of. There will probably be altcoins which are mixed with fiat currencies. Some nation may even create a DigiC which is their official currency -- and some will love that and some will hate it.

All can coexist, but all can't retain their value. Money is a very useful tool, and today there's about $22 trillion of cash out there. Each currency will have a market share, and that market share, combined with the supply of the currency, will dictate the value of the currency's unit. People will chose which currency to use based on features like ease of use, security and transaction costs and times. People will choose which currency to hold based on predictions of whether the value will rise or fall compared to other currencies. Sadly for the leading currency, it is more likely it will fall and competitors will rise -- that is the path of history. (Though as a counter argument, one could consider that because most currencies are inflationary, gold has not lost to its competitors.)

As long as some portion of the value of a DigiC is considered speculative, it is at risk of losing that value, and losing it quickly. This leads us to the risky endgame for Bitcoin. One day, popular sentiment will grow that another currency is "better." And people will start moving their holdings from bitcoins to the altcoin.

I believe it is possible that this movement could happen quickly, and that momentum for it could grow, and suddenly there could be a rush of movement over a very short amount of time. With nothing backing a bitcoin (or most other DigiCs,) it is only the faith of the public that they are valuable which supports the value. If that faith slips, it can disappear quickly.

This is not a requirement. The old bitcoins still have their utility. But people are fickle and want to be in the currency that everybody else thinks is the winner. Like social networks, you go where the people are.

Smart holders of bitcoins will exchange to the new winner early. Some will exchange to altcoins that end up not winning. Some will wait, and find themselves on a downward slope, eventually unable to exchange at any rate they find reasonable.

The new king or kings will not reign forever, though. It's less clear what stops the chain of progress. What makes the world settle on a DigiC so that it is at low risk of competitive replacement. That new DigiC would have to be very adaptable so that it could respond to competition and remain superior. It's a tall order.

The potential for this quick shift from one currency to another is part of the motivation for my advice to Satoshi to donate a large fraction of (his) otherwise unspendable bitcoins. It would be particularly troublesome if it were observed that Satoshi was moving bitcoins into an altcoin, and would cause serious harm to support for bitcoin. Since the founder of a coin is unable to move out of it the way others can, there is a strong value in donating or selling early.

Could Bitcoin prevent this?

One might ask if "Bitcoin" (which is a movement, not an entity) will sell the altcoins the rope with which they can hang it. If Bitcoin were a company, one could see that argument. In many instruments, there are exchanges and these exchanges have the ability to halt trading when a collapse appears imminent, hoping to give time to prevent such a collapse. The idea of distributed currency exchange, under nobody's control, is very much in line with the philosophy of Bitcoin but it means there can be no control.

The Bitcoin system has been adapted to allow "side chains" which are forks with different rules that typically share mining and perhaps one or two-way currency exchange with the Bitcoin master chain. This is also intended to allow new versions of Bitcoin to test in beta. One can "suspend" a bitcoin amount from the main chain and then use it, under different rules in the new chain, possibly restoring it later. With consensus the two chains an even be merged back together. This is very useful, but is it enough to fight off all potential competition?

A technical outline

Here is an outline of the steps to an exchange transaction with no 3rd party involvement. It does require the existence of "markets" where offers to exchange, in bid/ask format, can be published, but a marketplace is not an exchange, it is just an advertising venue.

Let's say I have 10 "Newcoin" at address BN and at current exchange rates, that's 1 BTC. I have a bitcoin address BB that I want to receive that bitcoin at.

1. I look at the markets, seeking somebody willing to sell 1 BTC at a good rate. In their entry on the exchange, the seller will list either automatic contact details, or in a pure form, simply provide the public address "SB" of their BTC and a public Newcoin address "SN"
2. I publish a transaction to the Newcoin system. It says, "Transfer 10 Newcoins at BN to address SN, if a transaction for 1BTC from SB to BB is published in time window T"
3. Chances are, I inform the seller of this in some way and confirm it, but they might just see it in the Newcoin ledger. Either way, they wait until that transaction is finalized in the Newcoin chain, so they are sure I am not double spending the money at BN. After publishing this, BN is, at least temporarily, unspendable elsewhere.
4. Upon being satisfied the smart transaction is recorded, the seller publishes a transaction to the Bitcoin blockchain that matches the parameters above. In particular, it transfers 1 BTC from SB to me at BB.
5. I, and the world wait for the Bitcoin universe to finalize that and be certain there is no double spending. At this point all this can be confirmed by anybody from the blockchains:

* I can freely spend my new bitcoin * I can now never get back my newcoin. * The seller can freely spend their newcoin. 6. As a courtesy, I may publish a "resolution" transaction to the Newcoin blockchain, indicating that the conditional transaction completed or, after time window T, did not complete. The miners there can verify if that's true, and this saves others in the future from needing to trace everything, but they always can. If the transaction is calculated or marked as non-completed, the newcoins at BN become spendable again, and there are never any coins at SN.

The main problem with this process is that it's slow, it requires double finalization of both blockchains. However, people who want speed can get it either with higher transaction fees, or the use of trusted parties who can make smaller transactions instantaneous if desired, for a small fee. And while communication could take place entirely in the blockchains, it makes more sense if I actually make some more direct connection to the seller and we agree to the transaction in advance, to avoid wasting people's time with double attempts on the same coins or waiting for the blockchain. The long time period might be an issue if exchange rates are fluctuating very quickly -- the seller might decide not to sell because the price changed.

It goes much faster if both currencies support these conditional exchange transactions. In that case, both conditional transactions are published at the same time, and both are complete only when both blockchains have finalized them, and if that doesn't happen, neither one is. This avoids problems with rate fluctuations. This, Bitcoin might resist.

Even if Bitcoin as a community were to resist such exchange, it has little power.
It thus seems that decentralized exchange is possible and likely, and nothing can stop the free competition of DigiCs, to the detriment of the older ones and benefit of the newer ones.

I welcome comments on alternatives to this path of the future.

(1) It is highly unlikely these market caps are "real" in that attempts to sell large fractions of the coinspace would highly depress the price, and so this should not be compared to the market cap of a public company; those often sell for more than their market cap when acquired. A more interesting number to examine would be the real trading volume (not counting internal trades, coin laundry, etc.)