I don’t see the problem. There seems to be no logical reason that local laws can’t change because of arbitrarily complicated nonlocal rules. You can even see nontrivial examples of this in practice in some modern technology. Various of Microsoft’s operating systems have reportedly contained substantial amounts of code to recognize particular usage patterns characteristic of particular old applications, and change the rules so the old application continues to work even though it depends on old behavior which has otherwise disappeared from the new operating system. Vaguely-similar principles of global patterns changing local decision rules also appear, in less-nauseating ways, in all sorts of software for solving hard optimization problems (optimizing compilers, finding the optimum move in Chess, finding the optimum schedule for a big logistics operation...). What would go impossibly wrong if you rewrote physics with added rules which recognize patterns characteristic of presence or absence of patterns (like “living organism” and “magical incantation”) and which rejigger the local rules as a consequence?
Changing the local rules specifically to stomp out technology without making the rest of the universe’s behavior unrecognizable is a tricky job, since you are correct that everything tends to be cross-coupled in weird ways. But I think one could at least make existing technology pretty frustrating. One way to start would be by making a list of a hundred or a thousand technogically useful patterns (things heating up to combustion temperature, things bending around a fulcrum, sizable things rotating or oscillating many many times without changing shape, lots of energy being stored for a long time in an elastic object) and make case by case hacks to damp them out (spontaneously cooling things when they rise above 100 degrees Celsius, letting the lever soften and bend, etc.) whenever they weren’t preceded by the suitably magically approved pattern of causality. (So, e.g., you can light a fire with a spell, and perhaps by striking suitably hard objects against each other, but not with a match or a magnifying glass. And you can use hinges as long as they are between bones in a living organism.) The result would be a very weird universe, but if I remember correctly (from long, long ago), the universe in those books was supposed to be very weird anyway.
Note that when someone reads your “if people have a right to be stupid, the market will respond by supplying all the stupidity that can be sold” it does sound rather as though you’re making a point about market decisions in particular, not just one of a spectrum of points like “if people have a right to vote for stupid policies, then ambitious politicians will supply all the stupid policies that people can be convinced to vote for.” Also, it’s not too uncommon for people to play rhetorical (and perhaps internal doublethink) games where people’s rationality in market decisionmaking is judged differently than in politics.
Similarly, you could state specifically “when we let members of someparticularethnicgroup vote, they often make uninformed decisions,” and believe yourself logically justified by the general truth that they are people and when we let people of any ethnic group vote, they often make uninformed decisions. But I don’t recommend that you try making that statement, especially about an ethnic group where it’s not too uncommon for people to dump on them in particular, unless you’re prepared to raise many, many more hackles than you would by just stating your general point about letting people in general vote.
And even though I give the parallel of another politically charged statement, I don’t think this is just people getting irrational around politically charged issues. In ordinary, un-charged situations too, it is normal for people to choose reasonably general forms of their statements when possible, so if you make a narrow statement, it conveys a suggestion that a more general statement doesn’t hold. “It’s really cold in the living room” means in practice something like “the living room is colder than the rest of the house” or “I am physically unable to leave the living room and don’t know about the rest of the house,” not “it’s really cold in the house.”
It’s not a completely reliable conversational rule, and it’s probably one of the reasons that some wag said that “communication would be more reliable if people would turn off the gainy decompression,” but it’s not obviously an unimportant or silly rule, either. In fact, if I imagine designing cooperating robotic agents with very powerful brains, very sophisticated software, and very low communication bandwidth, I’d be very inclined to borrow the rule.