The posts advocating moderation with hill work are correct. Some easy running in hills during the off season (shallow grades at first) is good to develop the structural integrity for faster or steeper hill running later. Then the occasional faster run in hills (or hill reps - up shallow to moderate grades and down shallow grades) can foster left ventricular hypertrophy to promote higher stroke volume (uphill running at a strong pace) and can improve neural recruitment (rate coding) for stride frequency and length and further improve the ability to withstand eccentric forces (downhill reps).
But stride mechanics and stride frequency are different when running at similar efforts (and even at the same absolute speeds) on the flats than on uphills, and too much running on downhills (particularly on moderate or steep grades) can cause DOMS that takes about 5 days to resolve. So be sure to balance hill work with running on level surfaces in order to train for specificity and to reduce the likelihood of DOMS from the downhills.
Re: Shorter's training speeds: It is true that he remarked that he found it difficult to locate training partners who would run as slowly as he liked to go on easy days and he also had trouble finding people who could keep up on his fast days (no surprise - the guy was an Olympic gold medalist after all). One workout I got from him in 1974 was 12 x 880y plus 12 x 440y, with the half miles being run at around estimated 15k pace and the quarters run at about 10k pace and using recovery periods that are pretty much equal to the run periods (or whenever you're ready to go again while keeping the effort below the laboring level). Some people might not consider this "speed" work, but it's very effective. Real-world trial and error bears this out, and real-world results override any theory. Here's an excellent quote that encapsulates this notion:
"It makes no difference how smart you are, who made the guess, or what his name is. If it disagrees with real-life results, it is wrong. That is all there is to it." - Richard Feynman
That said, research since Shorter's heyday shows that running short or medium-length repeats for a cumulative total of 25 to 35 minutes at the target pace (using the minimum necessary recovery intervals to keep the pace on the reps without struggling) at about the pace you can run for 40-45 minutes in an evenly-paced effort will foster improvement in using lactate as a fuel (it does this by increasing expression of the "monocarboxylate transport proteins" which help shuttle lactate). So Shorter theoretically (raised eyebrow) might have done just as well with 10 880s and 10 440s (@ 34-35 minutes of total "focused" effort), but he did run slower than many elites on his easy days (allowing more recovery, both metabolically and structurally) and he did have an extremely extensive base (perhaps having the natural structural ability and stride mechanics to handle it), so that much accumulated time at a strong effort might have been manageable for him in the long term.
In any case, balance the hill workouts with level-surface workouts (vary the running surface, too - road, grass, track) and learn how much you can handle, managing the stress/recovery equation so as to only push the boundaries often enough to avoid hypertraining or injury.