This conversation highlights the need for the Lightning Network to make progress so that more activity can be pushed to layer 2. The fact that we are having this discussion shows that more infrastructure is needed to take transactions off the main chain. The reality is, the main chain is necessarily limited by design, so high fees are inevitable with widespread adoption.
It seems to be mostly ready, but it'll probably take the most popular wallets implementing it for it to have widespread usage. We already know that it can be dumbed down because Eclair has done pretty much that. I still expect the main chain to be busy, but since LN would likely take a lot of the low-amount high-priority transactions (real world transactions basically), the mining fees hopefully won't end up too high.
The mining pools are responsible for the high fees. I still remember what happened in 2017, when some of the mining pools rejected a lot of transactions with low fee, although they were having a lot of space left with the 1 MB block size. Instead of using the full 1 MB size, some of them were using only 300KB or 400 KB. I don't have anything against prioritizing high fee transactions. But rejecting the low fee transactions and asking for a higher fee looks like plain blackmail for me.
There are a few below-capacity blocks, but they've been mostly full the past days.