Citing Lukasz as the topic of “increasing the gas limit” is more prevalent than ever, thinking of recent pushes like https://pumpthegas.org/ .
Furthermore, there’s already an EIP for increasing the blob count, which might also need a block size reduction as a prerequisite.
In addition, parts of the community have expressed clear demands to increase the gas limit and/or increase the blob count, and they must not be ignored (thinking of the upcoming year(s)).
While the current max block size doesn’t leave much room for such increases, EIP-7623 proposes to increase the calldata costs for calldata-heavy transactions, thus reducing the max block size from 3.5 MiB (incl. blobs) to 1.9 MiB.
The goal of the EIP was to keep the impact on users as small as possible while still achieving the goal of reducing the max. possible block size. Find the latest analysis on the EIP here:
https://nerolation.github.io/eip-7623-impact-analysis/index2.html
The changes needed are very small. Same for the impact (99% of user won’t feel any difference) and the majority of those 1% that are affected are users that put messages into tx (so there are better alternatives than Ethereum L1 anyway). Some non-DA use cases that are affected are big proofs like STARKs (~30% more gas) or on-chain merkle-proofs.
In my opinion, EIP-7623 paves the way for further scaling and should be included in Pectra.