@iserith: So let's start with what seems to be the fundamental basis of your objection, the claim that this is no different than a party acquiring the ability to fly and the OP has simply failed to engage in proper encounter design. Now, flight can put severe constraints on encounter design and there is a point beyond which all encounters have to assume that the party can fly if they are to provide credible challenge. But the first access a party is like to get to flight is the flight spell at 3rd level, which only has a 10 minute duration. So the constraints that puts on encounter design is rather weak. There is a point beyond which wilderness exploration no longer presents a credible challenge to a party with access to magic because they will have magical protection against the elements, magical portable shelters, the ability to conjure food and drink, magical transportation or even teleportation, and as such none of the challenges normally associated with survival in the wild are really relevant. But, again, that point is at a relatively high level, and there is a ton of space in which to design wilderness adventures with survival challenge components where magic might provide some help but doesn't provide a ready solution to every problem.
However, there is a categorical difference between the design constraint on scenarios and encounters that the 5e version of Planeshift imposes and something like the 3rd level flight spell, and that is that the Planeshift spell is likely the first means of reaching the planes the players are likely to have. So there isn't really any room below it. There is no space here between accessing the planes and nigh perfect mastery of access to the planes. This is a reliable, precise, mass means of transportation that makes no demands on the party. It whisks you reliably to where you want to go without even having knowledge of the destination. That's a totally different issue than a PC can fly for 10 minutes if he concentrates. It's more akin to a design where the 3rd level spell conjured a flying carpet for the party to use for 24 hours, resulting in reliable mass flight the first time flight as a resource was likely encountered.
This isn't a problem just for DM's playing 'gotcha' like a DM that wants to insist the players can't use fire against trolls. This isn't just a problem for a novice DM who designs a murder mystery without paying attention to the divination techniques available to the party members, or who is struggling to keep shopkeepers from being robbed by PC's because he failed to imagine a society where magic is prevalent and been in long existence and the PC's use of basic illusions and invisibility is something he can't handle. If my opinion was based on any of those sort of things, you'd expect my opinions on fire use by parties against trolls to be very different than they are.
Certainly you could make an argument that such things are a problem that DMs shouldn't have to deal with and that spells like invisibility or flight or persistently misleveled for historical reasons considering the profound impact those abilities can have. But no such argument is necessary for what I'm talking about.
A reliable precise means of mass instantaneous transportation provides severe constraints on encounter design that goes beyond just needing to deal with flight or create food and water. It's not just exploration that potentially goes away - even though exploration on the outer planes is often a good next step in wilderness exploration once the challenges of more mundane exploration go away and this blocks it. The new design of the spell also challenges investigation challenges since the PC needs no special familiarity with the new location at all.
And this is incoherent because the 5e designers otherwise went out of there way to address this as a problem, making changes to the game which are obviously designed to prevent instantaneous transport spells from easily wrecking scenarios.
Again, consider the changes around teleport. Not only did teleport move up to 7th level, giving DMs more space to design encounters that couldn't be bypassed by teleport, but 'Teleport Without Error' went away entirely. The new teleport isn't reliable and can't teleport to an unknown location. You need some degree of familiarity just to get there even with repeated tries. Attempts to go to an unknown location reasonably go very badly. You need at the very least an accurate description of the location you are attempting to go to.
But Planeshift at the same level has no such restriction. It is reliable and doesn't require an accurate description and is precise to the level of per the description at least delivering you to the general area. So the alterations to how Teleport works are undone by the alterations to how Planeshift works. That's incoherent.
Some indication that what the designers were trying to achieve is eliminate reliable mass pin point instantaneous transportation can be found in the built in ambiguity of the spell. Namely, the spell specifically calls out that the DM by fiat may block pinpoint transportation if such transportation wouldn't make for a fun scenario. But the designer seems to think it sufficient for the purposes of the spell that the DM will be able to place the party in the starting location of his choice - essentially at the entrance of the dungeon. As the OP points out though, this call out doesn't handle the problem of a hidden or secret destination. Teleport is blocked from working in that situation, but Planeshift allows this restriction to be bypassed.
I understand the intention of the DM fiat callout. Planeshift in its utility mode has always been a defacto DM tool for simultaneously allowing access to the planes and controlling it. The DM has always had the option to fudge the 'miss' roll and locate the players in a position best suited for the needs of the scenario. Removing the random factor and empowering the GM to make a ruling is I think intended to empower the GM to do this without fudging. But, if that's really the intention, why leave open the ability to Planeshift to completely unknown locations? The designer could have called out that Planeshift was only semi-reliable and if shifting to an unknown location, it was highly likely that the party would not land within eyeshot of the destination or have a clear idea where it was located.
But then, I suppose that would make too obvious that this spell running on GM fiat was really a GM tool and could be used in an adversarial way? Point is that the design is weird and stands in contrast to the designs applied elsewhere in the addition. It creates problems without really providing much in the way of a solution. The only real benefit is that a GM might not feel hidebound to force the party to land 300 miles away when such a journey doesn't suit the pacing of the game, but again, if that's the case, just go ahead and say, "Party lands at a destination the GM feels is appropriate." or use some sort of random distance depending on the familiarity that the party has with the destination in parallel to and complementing the implementation of Teleport.
