I think even this overlooks a middle ground. One of the potentially greatest breathroughs in low input high tech agriculture is no-till farming. I say "potentially" because in practice it is often used as an execuse to pour on huge amounds of Roundup, and other herbicides.
But it has been demonstrated that if you do no-till farming with a multi-crop rotation, you can drastically reduce pesticides, herbicides and artificial fertilizers. The trick is to rotate a a fiber criop, a grain, a green manure and a legume. (For example, kenaf, grain, rye+hairy vetch, soybeans). The fiber crop pretty much outcompetes the weeds with few or no herbicides, clearing the field for the grain. The grain takes up the excess nitorgen that was added to the soil by the last round of green manure and legumes. The green manure and legume of course both add nitrogen. Note that three of the four crops are harvested for human use. The green manure is planted during the part of the season where the soil would be left fallow in any case.
Because it is no-till, all of the crop not actually harvested for human use is returned to the soil. This not only retains a lot of nutrients that would otherwise become agricultural waste, in builds soil structure to help keep the nutrients availabel for plant use and reduces runooff. You eliminate the use of nitrogen fertilizers, reduce the use of other fertilizers by about 80%. With the rotation of the fiber crop you greatly reduce the need for herbices otherwise required in no-till farming; optimum results have been obtained with about 80%-90% fewer herbices than conventional industrial farming. The variety of the rotation lends itself to integrated pest management - reducing pesticide requirements again by 80% to 90%. Labor costs are LOWER than with conventional industrial farming.
Eliminating nitrogen fertilizer and drastically reducing the use of heavy machinery, and reducing chemical use by 90% lowers energy drastically below that of industrial agriculture and substantially below that of much organic agriculture. (Many forms of organic agricultures use 70% of the energy of industrial farming. This particular type of no-till cuts energy by as much as three quarters.)
Why not go further and do true organic no-till? Because t hat lowers the yield per acre drastically. A small amount of pesticide and herbice seems to be the optimum for this type agriculture (though Rodale is doing interesting work on high-yield no-till organic farming). This smaller amount of herbicide and pesticide really will break down before it enter the food chain. Because no-till retains such chemicals much longer they also don't end up in run-off WHEN USED IN THESE QUANTIES. (The quantity is important. Conventional no-till without extensive rotation produces a lot of run-off -though still a bit less than traditional industrial agriculture). Run-off is reduced a great deal more than input in no-till farming. For example water use is only cut by 30%, run-off by half. And the greatly reduced run-off is actually visibly cleaner - in some cases actually clear. And the amount of pesticides and herbicides stay small. Because they are targeted, used only as needed, you don't get immunities developing that require large amounts each year. Similarly because you are building rather than destroying soil, fertilizer demand never increase. (In fact we might predict that in the long run int will drop.) If conventional agriculture hooks the soil on heroin, you might compare this type of carefully managed no till rotational agriculture to using the occasional apirin.
This has labor yields better than traditional industrial agriculture, and it turns agriculture from a carbon source back to a carbon sink. The yield per acre is as good or better than industrial agriculture. Combine it with other tricks like using crops suitable to the local climate (not native plants - just avoiding stupidity like growing cotton in the Arizona desert), and drip or sprinkler irrigation where irrigation is needed, and you can cut water consumption by 75%.
Note that this is still industrial agriculture - very high tech. You need to test to soil to avoid over-application fertilziers, target pesticides and herbicides to particular pests and weeds.
Of course this is suitable to a nation with high labor costs and good access to capital - like the U.S. But don't we want all nations to end up in that position eventually?
I've left out a lot of complexity here. For example the importance of no-till in mainting soil fungi that help cause the soil retain nutrients (and act as a superior carbon sink to tilled or ploughed soil). Or that fact most grains produce more straw than should be left on the ground. Excess straw robs the soil of nitorgen. But the straw that has to harvested in any case can be used directly as building material in areas where land is cheap. (straw walls are inexpensive and sturdy - and insect resistant and fire resistant. But they are also thick and support structures of three stories or below - so are wasteful where land is expensive.). They can be used partical board and fiber board that in can substitute for all conventional forms of those things, and that may substitute in some cases for inferior grades of plywood - reducing the need for timber harvesting.
Similarly management intenstive rotational grazing raise meat animals, cattle, sheep, goats, bison and such with higher production per acre than feeding them grain if take into consideration the acreage used to grow that grain. Done properly such grazing builds the soil rather than eroding it - (as the great plains were built up by millenium of grazing by the buffalo). So animal husbandry can become a carbon sink rather than source as well.