In a recent tweet thread, I showerthought about “premium execution layers”. Let’s take the thought experiment further…
A premium execution layer is exclusively focused on high-value financial applications; target users: HNWIs, corporations, governments, high-value rollups etc.
For this, the key design goals are:
Economic security & sustainability
Sound money base asset
Stability (pref. ossified)
Quality auditing community
As Sylve points out, a lot of traditional finance still runs on half-a-century year old COBOL code. Fancy new tech matters not when it comes to dealing with money. People just want something stable and proven. The more money there is at stake, the more ossified a solution they want.
Now, there’s nothing in crypto that’s Lindy enough, but let’s start with the best we’ve got. So, it has to be a multi-client EVM execution layer. We’re probably a decade or two away from EVM being sufficiently Lindy, but it’s the best we can do, for now.
L1 or L2? The key requirement will be secure access to liquidity, so it’s gotta be an Ethereum L2. That’ll give you validating bridges with access to $350B worth of assets, plus inheriting full Ethereum security. I won’t rule out L1 entirely, though, it’s not impossible there’s some revolutionary innovation (e.g. a new consensus protocol that obsoletes proof-of-stake; or an groundbreaking money design that obsoletes BTC and ETH) eventually - but it can be considered when it happens. There’s some
It’ll have to be an optimistic rollup - zk rollups are far too complex at this time. So are ORs, but they have a shorter route to ossification. A withdrawal time of 7 days is perhaps too long, but if that can be shortened down to 2 or 3 days (matching T+2 standard) that’ll be perfectly fine. Remember, this is just for withdrawals, which will be fairly uncommon for the target usecase. Once again - high value financial transactions are all about security and settlement assurance, not time pressures. Post-EIP-4844, the rollup will likely pay <1% of fees to Ethereum, so >99% are profits for the network - an excellent bargain!
You don’t need to make it that easy to verify. The simple reason is the target user can easily afford to buy a supercomputer and hire someone to maintain it. That said, I also don’t think much TPS is required, so a balance can be found where all the wealthiest entities in the world are supported. It doesn’t really matter if it’s a monolithic L1, a ZK-L1, or an L2 in this regard.
Last, but most importantly, you have to have an incredible business development team that reaches out to every wealthy entity in the world to use their network. It’s also imperative to get traditional finance players on board and supporting the network.
Does all of this sound like some capitalist dystopia? Based on your worldview, maybe. But that doesn’t matter, this is just a fun thought experiment about the most valuable possible execution layer.