I think this is a very good point. I want to respond to it in two ways.
First, I think in general that 4e is a game that rewards pre-prep of encounters. At least for me, it's easier to map out combat locations with interesting terrain in advance rather than on the spot. Similarly, it's easier for me to prepare an interesting skill challenge in advance rather than on the spot.
I don't dictate to my players what encounters they will have and what their nature will be - rather, I rely on my intuition as to these things. Sometimes I'm wrong, and have to work stuff out on the fly.
I think the DMG2 discussion of secondary skills for skill challenges has helped me a bit with this, because it gives a better range of ideas about what secondary skills can do, and so has helped me make on-the-fly judgements about how I should mechanically interpret a player's description of their PC's actions in a non-combat encounter (eg are they trying to undo a previous PC's mistake? if so, then it's a secondary check to try and negate a failure).
I don't think that 4e is any worse in this respect than any fairly crunchy RPG, but I could be wrong. Anyway, this is the first time I've really seen someone point to an aspect of 4e that has the potential to push in a railroady direction.
Second, I think it is possible to place pre-prepped encounters in such a way that the prep makes it easier to run the encounter, but the prep is useful regardless of the goal at which the players aim. So a list of likely avenues of social interaction with an NPC, together with DCs, can be useful prep even if it's not clear whether the players will want to treat that enemy as an ally or an enemy. I admit that if they decide to start a fight then all that work is wasted (this is a consequence of 4e having two such different resolution systems), but maybe some of it can be recycled the same way you might recycle a location if the PCs end up having a picnic with the orcs rather than fighting them.
What I usually do (using any system) for regular campaign play is organize prep into categories:
Locations
NPC's/Monsters
Events
I try and work on these independent of each other. If there is a location that is a likely spot for combat then I will note the relevant details about the area.
NPC's have statblocks which contain combat and non-combat
information (skills, motivations, personal details). The amount of detail depends on the importance and influence of the NPC. A shopkeeper might just have line about personality, skills, and share a combat block identical
to dozens of other NPC's of this type (listing it just once to save space). A major player in the area would get a full statblock and more notes about goals/motivations.
Events are either triggered by PC activity or set in motion by an outside force ( an NPC, nature, a deity, etc). Some of these will happen at particular times, some might happen without intervention, and others will only happen if the players take specific actions.
Each element has obvious influence on the others and they combine in various ways but I try and avoid stringing too many together combined with the presence of the PC's. I like to let the players decide what is
important to focus on. Whatever they decide becomes the center of the action. This is why it is difficult to design meaningful skill challenges.
Using my prep method, if a pursuit begins then the relevant info is at hand. I have info about the environment, the NPC(s) involved, and any events taking place in that area at that time. Detailing an elaborate chase challenge that may or may not happen in that particular place or time seems like wasted work.
The structured scene based design that 4E is designed around is (IMHO) more suited for convention/organized play than open ended home campaigning.