To me, this looks like demand is going to cheaper blockchains instead of scaling on Bitcoin L2 solutions.
As long as there is free space in block it makes sense for miners to include any non-zero free transaction, so it is a race to bottom with only occasional surges, kinda like the ones that happen in Uber.
Almost $800k is burnt hourly. Very few people need to transfer that much, and those who do, can wait for extra confirmations.
I would like to see the block reward reduced even SOONER. Halvenings every 4 years is too slow. But I don't think it's going to happen sadly.
As for fees, they could be driven higher by limiting block size, should the community decide they need more security once the block reward becomes too small.
Point is: the algorithm is not infallible, and would have to adapt. Otherwise Bitcoin becomes brittle.
The second part of the miner reward consists of transaction fees which will compensate. Most small scale transactions will be done on second layers such as lightning, thus being unaffected, while big transactions for which the extra security is worth it will be willing to pay the high transaction fees.