Just now I took another look at the L2 community arguing about TPS, transaction fees, and subsidy ecosystems, it feels like a marketplace where everyone is shouting louder… Back to the topic of modularization, what exactly has it changed for end users? Honestly, you won't suddenly earn an extra penny just because of "modularization"; more likely, it's about whether the failure rate of the same transaction/interactions can be lower, whether there are explanations when things get stuck, and whether issues can be diagnosed faster to identify which layer is causing the problem.

In the past, everything was integrated; if the chain had a hiccup, you could only curse "this broken chain." Now, splitting into execution/data/consensus modules, in theory, problems are more diagnosable: is it sorting congestion, data publishing delays, or validation nodes falling behind? The benefit is that a failure won't take down the entire network at once; the downside is quite real: more bridges, longer paths, crossing back and forth, and the risk surface expands accordingly. Anyway, I now see "lower fees" as already immune; first, I look at whether stability and exit/rollback mechanisms are well implemented, then decide whether to use it.