From auditless research by Peteris Erins

It's an exemplary post and one that I hope inspires the Ethereum community to action.

Surprisingly, he’s actually advocating for improving the L1 for apps:

It’s a pretty cool thesis coming from the guy who wrote The Complete Guide to Rollups and other essays that helped all of us get excited about rollups in the first place.

I'd like to build on the conversation and consider a couple of additional questions.

What is the value of a North Star in the absence of use cases and moats?

There are several operational benefits of having a North Star, e.g.:

But on balance, it’s also important to pick the right one.

For example, EOS, Tron and Neo were brought up as credible threats to Ethereum as recently as 2019.

Now they don't even appear in the same sentence.

Do we know enough at this point in time to commit to a lane?

Have we gotten to a point competitively where we need to make a bet and run with it?

I think a good North Star for a blockchain is informed by two things:

Regarding use cases, we don't know.

Few dare to articulate whether DeFi or something else is more important to Ethereum’s future.

And regarding Ethereum's moat?

People disagree; here are some examples:

If we don't know the use cases and don't yet understand the moats, we may pick the wrong North Star.

Of course there’s still a valid argument that we should pick anyways and get to building.

Should we deprioritize rollups in general or just third-party rollups?

In our previous post, we discussed the possibility of enshrined rollups and it’s not clear to me from Jon’s arguments that this option is untenable.

There seem to be two arguments against going all in on rollups.

The parasitic argument

This argues that the relationship with the user/developer is most important and Base and other successful L2s will eventually get powerful enough where they won't need to settle to Ethereum.

It's certainly interesting when you compare the success of Base and Ethereum:

Base Onchain Assets, L2Beat

Galaxy Digital

The anti-Celestia argument

This argument states that Ethereum should avoid competing with Celestia as it will be forced to compromise all of its advantages and values in a bid to offer cheaper and more abundant DA for rollups.

Whichever argument you feel is stronger, both arguments basically lead to the same conclusion.

Ethereum’s customer is not the rollup but the app developers.

One more strike for aggregation theory :).

From there, Jon concludes that Ethereum should invest in L1 app adoption by making the base layer more performant without compromising reasonable decentralization.

But is enshrining rollups also a viable architecture that could achieve the same objectives?

This seems to boil down to technical belief and a bet on the specific use cases.

For instance, if we believe that great apps can be built on a multi-rollup architecture with fast interop, maybe enshrining them is the way to go.

If instead composability will work a lot better for a single atomic state machine, that’s an argument for Jon’s proposed path.

The layouts can get quite nuanced when you reason about apps.

For example, is DeFi better on a monolithic state machine? Or is there too much contagion risk?

Perhaps it works best when a single rollup (e.g., Unichain) serves as a liquidity hub while other rollups onboard capital and generate revenue for DeFi collateral.

In summary, I've got more confidence about the merits of staying close to the developer but I wouldn't be comfortable betting on which architecture will prove best for 90% of use cases.