1/ Ethereum will scale this year using L2s.

But which L2s should we pick?

Let's explore some of the tech and governance issues around the choice of L2s.

TL;DR imo, getting most of DeFi on Optimism stands out as a great option

2/ This year, optimistic rollups may make DeFi faster and 100x cheaper. To get there, ethereum needs as much of DeFi as possible on the same L2 instance. That L2 instance needs to be supported by MM and coinbase. Then, DeFi will thrive on L2, and users will rarely need L1 txns.

3/ Stepping back, what tech capabilities are necessary for DeFi to thrive? Which L2s or chains can satisfy these necessities? What are the governance implications of locking $100b in a single L2 instance?

4/ DeFi may thrive where generality, composability, and permissionless innovation exist, which is on ethereum, sidechains, other L1s, or within the same L2 rollup instance.

5/ Generality gives us smart contracts. Composability gives us flash loans. And, permissionless innovation gives us flash loan customers.

6/ Among the options for where DeFi may thrive, only the same rollup instance stands out as suitable to scale DeFi this year. The ethereum base layer isn't scaling any time soon, sidechains lack ethereum's security, and other L1s are competitors.

7/ Within the space of rollups, only Optimism and Arbitrum are ready this year and meet the necessary tech criteria. This means the ethereum community's primary good option to scale DeFi this year may be to pile onto a single rollup instance using tech from Optimism or Arbitrum.

8/ Where do zk-rollups fit into all of this? zk-rollups are expected to eventually be better in every way but won't be ready for one or more years. That's why it most likely has to be Optimism or Arbitrum's tech to scale ethereum this year.

9/ Optimism and Arbitrum's tech is (or will be, presumably) open source and can be deployed by anyone any number of times, each as an independent L2 network instance. We only get composability within the same L2 instance.

10/ Moreover, each L2 instance is free to design its own ownership strategy, including potentially launching a new token, who runs the nodes, how software upgrades work, etc.

11/ When a lot of DeFi is on the same L2 instance, that L2 instance seems likely to accrue significant network effects. Its ownership and community will be stewards of the current-generation de facto scaling solution for ethereum.

12/ So the situation is that we want DeFi to scale, it scales generally only if a lot of it is on the same L2 instance, and then that L2 instance may accrue great power.

13/ I think it'd be great for ethereum if as much of DeFi as possible joined the same L2 instance. The right L2 instance. We'd achieve the milestone of having effectively scaled ethereum via an L2 instance that's a reasonable substitute for the base layer.

14/ If we're all joining the same L2 instance, which one do we pick? Which one is the right L2 instance? For starters, I think that some L2 instances will end up being much better than others from the standpoint of trusting them with $100b+ and significant power.

15/ When I think of an L2 instance that we might consider trusting with $100b+ and significant power, some criteria I've been able to come up with are:

- publicly owned
- just uses ETH. No new token
- has a centralized governing body, something like the EF.

16/ Optimism and Arbitrum may both be reasonable choices. Each is approaching mainnet and must decide how to position themselves as the right L2 instance for as much of DeFi as possible.

17/ Optimism might have a bit of a leg up over Arbitrum because Optimism is a public benefit corporation and was always intended to be a single global instance. imo, migrating or co-deploying to Optimism stands out as a great option.