People predicting vastly lower fees after Dencun are setting users up for disappointment. Yes, whichever rollup(s) will switch to blobs first will enjoy very low fees for a while, but in a few weeks fees will be back closer to today's levels. 4844 is not yet DA scaling, just a more efficient separate rollup fee market.

Ethereum protocol R&D. Admirer of emergent order.

Basically, post-Dencun, each Ethereum node still downloads all the rollup data (now via blobs). So throughput constraint is still node bandwidth. To go beyond this limit, over the next year or so we will start rolling out sampling techniques (DAS), where nodes can check blob availability much more efficiently.

https://hackmd.io/@Jesserc | research @rareskills | Golang | https://github.com/Jesserc

Yes definitely! That's what I was trying to say with "closer to today's levels", I should have worded that clearer. Basically, it will be lower than today (all things being equal - easy to see L2 demand pick up enough to negate that), but only moderately so. Exact change depends on demand elasticity, hard to predict.

Shouldn’t there be a little permanent drop in fees?
Since L2 data is now separated from block space?

Arbitrum swaps are like 1 dollar. Right now the blocker on people using Ethereum is more bad marketing and bad UX. 

I told someone to go buy a coin on Arbitrum, so they went and wasted a lot of time and money bridging to Arb from mainnet instead of just directly depositing from a CEX. They missed the 1000% pump.