On the Ethereum hard fork

It’s been an exciting few weeks in digital currency. Ethereum went through its hard fork, and Coinbase finally announced support for it.

I think it’s safe to say we underestimated the attention that the minority chain would get as a result of the hard fork.

This meme made the rounds today on reddit, poking fun at my tweets, which I have to say was pretty well done :)

I thought it might be interesting for people to see some of the behind the scenes of how we thought about the hard fork, both before and after it happened.

Leading up to launch

The Coinbase team had been hard at work for a number of months, getting ready to launch support for Ethereum. And our original launch date was set to be July 19th, 2016.

As we got closer to the launch date, it became clear that Ethereum was planning their hard fork right after that (on the 20th). Originally, I was all for launching on the 19th. I was ready to get into the Ethereum game as quick as we could (I felt like we were late), especially if we could find a safe way to operate through the fork. But based on the advice of the team (Charlie Lee in particular) we decided to push the launch until after the hard fork to be safe. This turned out to be a good decision (which I’ll get to later).

We also discussed the possibility of replay attacks at that time. A number of other exchanges had created smart contracts to separate out ETH and ETC as it was withdrawn (I have to give credit to them for being ahead of us on this) and we considered delaying launch to set up something similar. But we knew the vast majority of Ethereum was stored in cold storage (where it would be safe from replays) if ETC ever became real. In addition, the price of ETC would be low and we figured a minority (roughly equal to the hashing power %) of users would be interested in obtaining it.

We were comfortable with that, so we decided to go ahead and launch.

Launch

The launch went well. About $1M in Ethereum traded in the first 24 hours, and we were off to the races.

Over the coming week, as we expected, some replay attacks started to show up. And the ETC chain got a bit more attention that I anticipated. I originally thought it would get 1% or less (both in hashing power and price). But it ended up getting closer to 15%.

We’d now had some more time to think through what our policy would be regarding ETC, and we started getting a number of emails from customers asking about it.

For Coinbase, the decision was fairly easy. Since we had launched after the Ethereum hard fork, there were no customers who had ever had “pre-fork” ether stored with us. We had also never announced support for ETC so we didn’t feel there was any expectation of customers being able to retrieve it. We put out an official statement here to provide some clarity to customers, and make it explicit that they should not transfer ETC to Coinbase.

For GDAX, our exchange for professional traders, the decision was a bit more difficult. Since GDAX had added Ethereum support prior to the hard fork, there were some customers storing ether there who may have wanted to get access to ETC as well. We had never made any announcement about supporting ETC, but we also felt like it wouldn’t be right to deprive them of that option if they wanted to move their ETC off the site (since they had moved it prior to the hard fork). First, we began work to nullify the replay attacks. Even though the vast majority of funds were protected in cold storage and we owned sufficient ether (and it’s corresponding ETC) to cover what was in the hot wallet, we didn’t want this to continue long term. We also went to work on building a feature for GDAX customers (who had a balance prior to the hard fork) to withdraw their ETC. We didn’t have any plans to support ETC long term, so we made it clear that customers should not send any additional ETC there for trading. Official statement here.

So was Ethereum’s hard fork a success?

I think the experiment is still being run. It certainly was not as clean as it could have been, but there were some positives as well.

Someone more eloquent than I summed it up nicely earlier today:

• A persistent chain split is bad, all else being equal (loss of network effect)
• However, the combined value of ETH + ETC is now higher than it was pre-fork (so all else is not equal)
• When there are irreconcilable differences, it probably makes sense to split, and you’ll get a net gain

If ETC persists, there is nothing wrong with that. I’m not ideologically opposed to there being multiple coins. People should be able to create whatever options they want in the world. And if there is genuine interest in ETC, I don’t see why we wouldn’t add it in the future. But I also don’t think it offers anything particularly compelling at the moment. The fact that Ethereum did a hard fork to recover funds from an attack sets a precedent for all digital currencies, not just one chain. Forking does not undo this.

I didn’t have a strong opinion either way on the reason for this particular fork (The DAO hack, I can see it from both perspectives) — I’m more interested in hard forks as a general concept, and how they will be used (or not used) as digital currency evolves.

I do think it’s safe to say that my estimates regarding hard forks were wrong in this case, and I probably deserve to be made fun of by the internet for that :) I would have guessed that within a few hours, the minority fork would be trading at less than 1% of hashing power, and price. Clearly this didn’t happen. According to Vitalik, a substantial portion of people buying ETC where bitcoin small blockers. I have no first hand knowledge if this is true, but if so it is an interesting outcome that I didn’t anticipate. Someone should have taken me up on my bet (I would have lost if this had happened in bitcoin).

It could turn out that I’m directionally correct, but I was wrong on the timeline. Will ETC fade in the next month? Year? Only time will tell.

I will say that I’m glad we got to see a hard fork in action. I think the Ethereum developers learned a lot from the experience, and future hard forks will probably improve on this one to avoid the issue of replay attacks. I admire their willingness to move fast and ship code, even if it isn’t always perfect. I love the saying “action produces information”. In the end, nothing happened that we can’t recover from, and I think the Ethereum network will continue growing — so in that sense it was a success.

Conclusion

I always prefer to look at digital currency from a long term perspective. I’m much less concerned about the ups and downs of any given week, month, or year. I don’t even care which currency ultimately “wins” (there will probably be multiple winners). The end state of the world I want to see is one with an open financial system for the world, where everyone has equality of opportunity around financial services. That vision will probably take 10+ years to be realized, and I want to help it happen as quickly as possible. Most of the decisions we make at Coinbase and GDAX are in furtherance of this goal. For example, I prefer to get digital currency in the hands of more people, than to argue over which one is best.

To create an open financial system, we have a long way to go. We’ve reached less than 1% of the people on the planet who could benefit from digital currency today. Most people don’t have any idea what digital currency is, much less what the difference is between ETH and ETC. They just want to send money home to their family in another country, sell their products to more customers, or be able to get a loan for their business.

I’m glad to see that the industry is continuing to evolve quickly. It feels like digital currency has picked up steam in the last 6 months (we probably hit the low point in the trough of the gartner hype cycle about 6 months ago) and we’re now trending up. “Blockchain” as a concept is probably a bit earlier in the cycle (near the peak and due for a correction, if I had to guess), but it will have it’s own resurgence after falling. I’m looking forward to many more interesting moments along the way.