But for me, here's the reason why less is more.
Look at kobolds.
In 3e they can in one variety, add a class if you like, but a lot didn't bother, because it was work.
Compare that to 4e kobolds. GOLLY DAMN they have a -society- of kobolds in there, each with a purpose and skillset that make kobold with spear different then kobold with sling. Kobolds have an identity, and yet have varience. Pages of material that would be wasted on 'use Rope' and 'This guy has the spell contact other plan' are instead used to flesh out monster races and make them into -actual races-.
See, I have the exact opposite reaction.
In 3.5, I had a kobold race, and I could easily create kobold wizards, kobold fighters, kobold rogue/barbarians, kobold whatevers, give them whatever items they needed, give them the skills that made them seem "real". The kobold chief had a high charisma, diplomacy, and bluff, because he needed them to be a leader. It's what a leader would logically have.
In 4e, I don't feel there's a kobold race. There's a kobold hive with kobold being born into castes from birth -- he's a slinger, he's a dragonshield. They feel as if they exist only for the fight; there is no society, no culture, no baby kobolds and old kobolds. There's just kobolds who spawn into the encounter and despawn when the encounter ends.
Now, this can be mitigated. The rules make it easy enough to build kobold NPCs. You, the DM, can tinker with trained skills and modify equipment and attributes -- give the Dragonshield more charisma and make him Trained in Bluff if you want him to seem more leaderly, etc. But the rules as written do not actively encourage this. Instead of a kobold race, you have a box of miniatures. Pick the ones you want for the fight, and get on with it. Fights over, put them back in the box.
4e CAN be used to create a world as deep and believable as any 3e world, but it requires a lot more effort, as you're fighting the encounter-centric design paradigm all the way. (Even your kobold NPCs will lack feats and have a much smaller selection of powers)
Maybe it's because I always used programs like PCgen and Herolab, but all of my humanoids are classed, which meant I could use them across all levels pretty easily, and they never felt boring of 'same old', because any kobold (or orc or hobgoblin or whatever) could have any of the abilities a PC could -- magic, psionics, incarnum, whatever. You couldn't say, "Oh, it's a kobold, it has X". You could have kobold merchants who were shrewd bargainers, kobold sages with kick-ass knowledge skills, and so on. With 4e, for most monsters, you quickly learn its "shtick", then you optimize to deal with it. "OK, this one does X if you get within Y. That other one has a one-shot burst power, once it's used it, it's harmless." Again, this can be avoided by making NPCs or "reskinning" monsters, but that's not the presumed default behavior.
4e is not nearly as bad as I feared it would be, and I am even contemplating switching my current campaign over to it once PHB2 comes out (as it will add in the classes my current PCs are using), but it really lacks a lot in detail and creating the feel of a "real" world. The "Gaminess" of it is constantly in your face and requires constant effort to overcome -- it's more work to describe game effects in dramatic terms, or to design interesting non-combat NPCs.
Someone else commented that the 4e system for monsters is great because they're more unique in that they have custom abilities, not spells -- but then goes on to say that, off stage, monsters can do "whatever the DM wants" -- which means they all have the SAME powers. I'd like to see Grazzt have very different "plot powers" than Orcus -- he should be forced, by his nature, to use different means to further his aims. Reading those massive stat blocks full of non-combat powers often inspired plot ideas in me. The current crop of demons, devils, and what-not, as described, can't do basic domination/control type plots, because those abilities in 4e only really work in combat and are very limited. There's no (official) long-term domination rituals, no way other than handwaving to say "The demon has total control over the king". (And if your argument is "Well, handwave it, then! Who cares?" the problem is that without any mechanics, you're left to be completely arbitrary when determining how the control can be BROKEN, and if such methods succeed or fail. Players do not like to feel their actions are limited by the DMs whims. They want know "OK, we do THIS, and then we roll THAT, and then we see if we break the control." It might be a skill challenge (research the ritual, perform the ritual, keep the demons agents from disturbing the ritual), it might be an attack, it could be anything -- but it should be quantified BEFORE the players try it.)