The improvements here are likely irrelevant for last mile. If hollow core fully replaces solid core, last mile deployments would use it, but saving 33% of latency in a fiber that's almost certainly less than 5 ms long isn't cost effective if there is any economic cost. The reduction in loss also doesn't provide a benefit for short runs. If there's an improvement in splicing, that might be useful for last mile, if splicing is harder, then it's less likely to be adopted.

On medium and long distance runs, it will provide a lot of benefits. Reducing latency on a cross country link is palpable; reducing latency on a shorter link like LA to SF is valuable too, because some routes have many of those. Reducing the number of amplifiers needed will be apprechiated by cable operators as well, fewer points of failure, likely a lower power budget, etc.

It may obsolete existing long haul fiber. But installed fiber will still be useful even if there's better fiber that could be installed... And existing fiber will be useful for redundancy and additional capacity even if there's better fiber on the same route.