Earlier quoted context omitted.
Any idea how to preserve the ability for people to go to "their" representative under that model?
You're much more likely to get "your" representative in this model. In geography-based districting, you only have a chance to vote for someone who represents your interests if they run in your district. Gerrymandering is often perpetuated under claims that we need to draw districts to group like voters so they can have representation. Example: Black majority districts were drawn to allow for Black representatives, bu…
At the same time that concentrating a constituency (such as black voters) into a district will make it more likely they get a representative that cares about their needs, it can also be used to reduce the total number of representatives that constituency would get in other instances.
For example, four districts with ~40% black constituents may be redrawn into one district of 85% black constituents and three districts of 25% black constituents, resulting in a representative that should be more responsive to the black constituents in that district. Conversely, four 60% black districts can be redrawn into two 80% and two 40% districts, resulting in likely fewer representatives that would cater to those constituents and their wants. The exact same strategy can be helpful or hurtful based on the specific makeup of the districts in question.
The problem with the proposed system is that you lose local representation, and you lose a level of definition in the views. If I live in the inner city, do I support the green party, or do I support the group that cares about inner city issues? I'm guaranteed me vote counts towards the group I put it towards, but I may be much less likely to find a group/candidates that actually encompass most my views and wants well (even if that candidate might not have been elected).
These systems aren't better or worse than each other, they're just different, and depending on who they are representing, they will do better or worse. It's entirely possible that if we could accurately measure how well people felt their views were represented and their will carried out a change could result in those going up, and a decade from now a change back to the current system could do the same (likely after a continuous drop in those numbers over years), because the groups and people represented are continuously changing and any assumption that a single system will work best in perpetuity is itself flawed.