I really like Magness' book Science of Running, and agree with much of it, including some things which some are sceptical of, such as the importance of muscle/tendon tension ('reactivity' in the sense of Klaas Lok).
However, in the training programmes which he sets forth, there are almost no simple workouts like 5 x 2000m at threshold, or 5 x 1000 at 5km pace. He instead suggests a lot of blend/alternation type workouts, and uses sets a lot, with quite little rest internally in the sets.
My feeling is that runners, especially of a more 'slow twitch' type, are better served by the more traditional, simple workouts. Indeed, I think that for many runners, the more complex workouts provide more of a crux, and hence less of stimulus (which is presumably not what Magness intends. I also think it is much harder to design an overall programme with more complex workouts.
Thus certainly it is great to be aware of these kind of complex workouts and have them in the repertoire, but I'd use them only sparingly and judiciously.
In this way, though people tend to be a bit scornfull, I actually think that McMillan's programmes are a very good starting point for many.
https://www.mcmillanrunning.com/best-5k-workout/
https://www.mcmillanrunning.com/the-best-10k-workout/
No need to adhere exactly to them, and I'd typically jump in later (I think it is probably unrealistic to progress from 400s to 1kms unless the goal race is very close to current fitness, i.e. one just needs to sharpen a bit) and add more workouts at other paces, but I think that, though very simple, they do provide a very solid basis for running a good race, especially for a more 'slow twitch' type. In particular, I am not a believer in 'prediction' workouts; I don't think there is any need to run 5 x 1km with minute or 90s rest, again, at least for a more 'slow twitch' runner.
Thoughts welcome!