A lot of MMT talks only about exploring better options for how that existing effort/flows is redirected, and if we would get testably better results from other more interesting configurations, not about creating some new tools out of thin air simply because they "want" to.
I have no idea how many proponents of MMT find interest in such similar engineering-oriented views of economics, but general economic theories of inflation/deflation from that point of view seem so primitively useless versus what we know of, say, electric resistance, inductance, capacitance, conductivity, etc.
I cannot tell you if my armchair theories are any better in the long term than the status quo or your preferred models, but only that MMT at least seems to be one of the few areas of economic study currently attempting to address them (albeit indirectly in many cases). That is also not necessarily proof that MMT is correct, just as my models don't ipso facto disprove status quo "wisdom".
Runaway inflation can make even intra-year saving impossible.
If this is a feature of MMT, it seems they'd be willing to give up on a lot of civilization since be began farming.
Yes, it's a very different view of economics from a lot of traditional approaches. It's not a fundamental "give up on a lot of civilization", even where it is a general rethink of "what does it mean to be a member of a civilization", including such things as what should be an individual's guaranteed safety floor, it's still mostly just civilization as we know it today swapping what we today think of as edge cases (social security, medicare, et al) for more generalized principles (in theory at least).
Arguably, to some proponents of MMT/in some versions of MMT, this is a feature and not a bug. Economically speaking, stored value at rest isn't "interesting", it's a potential indicator of waste or inefficiency of resource utilization in current flows. I've heard the argument that the real "bug" here is the human psychological need to "hoard" stationary value/"keep score", which admittedly is a very hard problem to solve, but having hard problems to solve isn't necessarily a disproof of a hypothesis such as MMT.