We can divide miners into three types: A. individual CPU miners, B. CPU farm, and C. GPU. Assume their hashes are 50MH/s, 50MH/s, 100MH/s for A, B, C, respectively. The block reward is adjusted so that it has maximum value when net hash = 50 MH/s, while block reward is zero, when net hash > 50 MHs. It will be clear then, when any two of them, or all of them (A, B, C) mining at the same time will cause zero reward. If A is keeping mining and won't stop (even with zero rewards), then B and C must reduce their hash to get their mining rewards (provided that B and C cannot afford days no-mining-returns, since they have higher mining costs than individual miners of A).
In reality, we have to adjust the numbers to better fit in the real case.
Everyone should carefully read the idea behind this update. I've seen people just ignore reading. If you feel something happening, we should be thinking solutions in addition to raising issues.
The above is just a simple description.
XMG won't be minted if network hash > 50 Mh/s (not accurate number but close).
You are punishing all the miners this way, limit individual miners instead of network.
Not like this, why do you think limit individual miners; I would see the exit of the big miners first! See individual miners can mine with few coins in 1-2 days (appreciated miners sticking to the coin so far), can the big miners mine with few coins or nothing in 1-2 days??
This is risky though, but this is better than doing nothing, simply watching big miners.
I've explained this idea to a few experienced people, they think this is innovative and they do worry about limits, but why not try it out?