It sure as hell is different!
First, (1) is NOT "find out what the players want to do".
(1) is "design the 'encounter'."
(2) is tell the players what they're going to do.
The only times I've told players what they are going to do in my games are when they e.g. need to make balance checks to stay standing. You are posting with a caricature in part due to poor modules. (A lot of 4e fans bitch about published 4e modules and how they
ignore the guidance in the DMGs)
1 is set the scene. Normally based in part on previous PC actions. This is unchanged.
2 is let the players tell you
In any RPG I've ever played, the only reason to make a "lock picking" or "pocket picking" or "banjo picking" or "nose picking" roll is because I'm trying to do that thing.
Yes, that's why you make a skill check. Even in a skill challenge.
That has not changed in 4e.
If I decide to shoot a Dark Troll, then I roll for that attack. If the shot kills the DT, then it's dead. If while it's not dead I decide to Hide and then Sneak and Pick its Pocketses, then Sneak back, that's probably several chances to slip up. If they all pass muster, then I've got what I've got; otherwise, I probably got caught!
And in 4e, that whole series of events could be a skill challenge. Assuming you planned that when you went in rather than were just reacting. Or it could be a series of skill checks if you were reacting as you went.
Maybe I'll try a line that warrants a Fast Talk roll at a plus or minus, or spend my time building longer lasting Persuasion.
Maybe I'll do something that calls for another roll, or for none at all.
Maybe at some point I'll try on its boots, or otherwise make an issue of its shoe size. Maybe I won't.
If it really makes no difference, then why pay cash money for it? Why champion and proselytize for it? If the new boss is just the same as the old, then wherefore the revolution?
I think at this point it would be easier to explain by giving you an example of a skill challenge I have actually used on the fly.
The whole thing started off as a combat scene with a dragonrider on a young dragon coming down to rescue a fallen ally. To cut a long story short, and through the dice gods keeping things interesting, the dragon rider killed his ally with friendly fire (rolling a natural 20 to do so). So the PCs now have a scared dragon rider and dragon and need to get them both somewhere safe to avoid trouble from above with the dragonrider's former allies, and from the city being attacked by dragons (and the rest of the army). This was in my second or third session DMing. Ever. In any edition.
The scene was set. I asked the PCs what they were going to do. And of course, they came up with a half-ludicrous PC plan. Throw the fledgeling dragon half onto a cart. Wrap blankets over it to disguise it from the air. Hold lanterns and claim it's a plague cart to keep people from investigating. How the hell do you work out which details to roll for in a plan like that and how much time to give it before it will get boring? Especially as there are hundreds of problems you could throw at a plan like that and not only wasn't the situation something I could have prepared for, even if I had I wouldn't have forseen that plan.
What would you do with a PC plan like that? For bonus marks, what would you do in your third session DMing and having to keep up with that?
My answer was simple and seemed confident (it wasn't, but it was a good enough fake). I made a mental note that this was a level 3 (the PCs were level 1 at the time), 6 or 8 success skill challenge (I forget which) and that gave me all the mechanical side I needed. I made them roll for what I considered the parts likely to fail of what they were doing (such as scavenge the cart - streetwise) and evaluated what they were trying as easy, medium, or hard (all you really need). When they failed a roll, I gave them a couple of curious street-urchins who came to investigate which opened up the rest of the social skills (the PCs got to pick how to deal with the kids of course - and bribing them would have been an automatic success). Whole thing was paced well, easy to evaluate, easy to give the XP reward, and fun for everyone without causing me to wrack my brains. (Of course I did not tell the players that I was running a skill challenge - I use them as a DM-side tool).
For that matter, Skill Challenges work very well for most insane PC plans. And I have some insane planners in my group. (I'm one when I'm not DMing.) Some way of bridging the gap between immediate skill check and mini-quest that drags out over a session or more. Skill challenges are particularly good when you have PCs working together but not in the same place - e.g. two PCs providing a distraction to allow two more to sneak in more easily. Or for group sneaking when someone's wearing heavy armour - the challenge is to keep anyone from hearing him (or bluff the guards with animal noises or just see the guards coming and hide or...)