And why?
Summary
Bitcoin is likely best today
Arbitrum and Optimism are perhaps next
Outlook
Bitcoin is likely best today
“Cheap” is relative, but with transaction fees of $10 and above I don’t consider Ethereum to be cheap, although I think it is quite secure.
Today, Bitcoin is the cheapest and most secure blockchain:
It’s not really cheap, but transactions cost around a dollar or two.
Ethereum today is still proof of work based, and Bitcoin is bigger, so I’m thinking Bitcoin has better security there.
Maybe big PoS chains like Cardano and Avalanche have decent security, but I don’t understand the details of validator centralisation and whether there are many full-nodes being run on the network. So, I don’t fully rule out that I’m overlooking something here.
The big issues with Bitcoin are: i) slow confirmation (10 minute block time), ii) only one currency (Bitcoin) that is highly volatile. Also, Bitcoin isn’t Turing complete so it’s not a direct comparison with blockchains that allow for computation.
Bitcoin Lightning addresses the first issue, but certain payments fail on lightning (I have a node myself) and the value locked in lightning is not even $1B.
Arbitrum and Optimism are perhaps next
If you think Ethereum is the second most secure blockchain, then it makes sense to consider roll-ups (blockchains that rely on Ethereum for security) as a good option for cheap, secure transactions.
The idea with roll-ups is to move computation off-chain but post a list of transactions to Ethereum (and also let Ethereum arbitrate if there is a dispute about off-chain computations).
There are two broad types of roll-ups:
Optimistic roll-ups, which send a transaction list to Ethereum assuming that all transactions are valid – but then allow a 7 day window for disputes
Zero Knowledge (or zk) roll-ups, which include a mini-proof that can be checked by anyone to show that what is submitted to Ethereum is indeed valid. (so there is no need for an arbitration step later).
So far:
Optimistic roll-ups are live (Optimism and Arbitrum are two implementations)
zk-roll ups (like zksync and starkware) are on test nets but not widely used on Ethereum mainnet. This is because zk-rollups are more complicated because of the mini-proof.
Another thing about roll-ups is that they aren’t necessarily EVM (ethereum virtual machine) compatible, so there is work for developers to make them easily work with common wallets (like Metamask). Right now, Optimism and Arbitrum work pretty well with common Ethereum tools, but I understand zk-rollups has more dev to do before being ready.
BTW, there’s a great way to think about transaction fees on different networks by looking at the costs of transferring assets cross-chain with Stargate. Calculator here.
Outlook
On the Ethereum front, I’m tracking closely what happens on the security front as it moves to Proof-of-Stake and I’m also tracking how things are going with different roll-ups, especially as zk roll-ups come online.
On the alt-L1 front, I want to learn more about security models – particularly models that might prove to be superior to Ethereum. My sense is that alt-L1s have roughly thought about optimising for transaction fees rather than security, which I think is a bad choice.
Further recommended reading: