Since the Mincoin distribution curve is flat and there are as many rewards per minute now as there were in the beginning and will be near the end, there are benefits for non-early adopters who want to start mining and still obtain a decent amount of MNC. (ignoring the first 4,320 blocks for argument's sake)
I don't necessarily have a particular favorite to propose in regards to a CPU algorithm with characteristics to make it GPU / ASIC resistant but I am inclined to think there must be some particular hardware extensions on the various newer AMD / Intel chips which could be featured in a proof-of-work algorithm where plain chips see a hash rate of x +/- 35% (due to clock frequency) and the good chips see a hash rate of somewhere in the ballpark of 3x.
For comparison, think processors from the late 90s encoding an MP3 before MMX extensions (with integer math optimizations) or video encoding before the various floating-point SSE extensions improved the throughput dramatically.
Since I haven't looked too deeply at potential candidate algorithms, I am open to discussion and listening to the thoughts from others about algorithm change related suggestions.
More than likely a "good" candidate will already be included in pooler's or tpruvot's cpuminer and have seen some action in prime time on a relatively secure chain.
At the same time -- people have their Scrypt ASICs from 2014 and might want to have a chance to find profitability with them in the not so distant future on the Mincoin chain.
Is the impetus to change the proof-of-work algorithm to secure the network or to level the playing field with one cpu one vote?
Or for the sake of argument, does one cpu one vote actually achieve the goal of better securing the network? (I think it does very much.)