The first and third points aren’t really big deals; actually, the third point would be a disadvantage of save-or-dies. Would be, if the numbers were more reasonable.
By the numbers: you can probably make someone fail a saving throw
Ultimately, caster classes have every reason to pump their save DCs as much as they can. The ability score that sets the DC is going to be their most important – after all, it also sets their bonus spells per day, and spells of a given level have a minimum required score in that ability. There is no other consideration in their lives that compares to what this ability score can do for them. This makes them “SAD” – single ability dependent. They can and should put everything they can into that ability score (short of missing out on spellcasting).
Meanwhile, even if you have all-good saves, the base save progression for good saves is the same as that of spell’s DCs: good saves get +1 every 2 levels, new spell levels, which have a DC 1 higher than the previous, are accessed every 2 levels. Good saves start at +2 (and a half, technically) over DCs, but as levels accrue that becomes a smaller and smaller portion of the overall bonus. So that’s only keeping you up with the basics.
But even with all-good saves, you need three different ability scores. It’s simply impossible to pump three ability scores as much as someone else can pump one. Items that give resistance bonuses to saving throws can make up the difference, but only if you’re ignoring the other three ability scores. And classes with all-good saves are rare, and tend to suck.
How do poor saves fare? They grow at +1 every 3 levels: over time, you are automatically falling behind. Most classes have a thematic link between saves and ability score focuses – if you have a poor save, there’s a pretty good chance you don’t use that ability score for much, either. Even if you do, you’re playing an impossible game of catch-up.
And the spellcasters can capitalize on this. A cleric’s Will save, there you probably have less than a 50:50 shot. But you still have a non-negligible chance. Someone else’s good save, that they don’t focus on the ability score for? You’ll probably land that. A bad save? Almost certainly.
Which wouldn’t be a problem – would actually be a good thing – but...
It’s good that spellcasters can do things; it’d be boring if saves were made often and spellcasters had to spend a lot of turns watching enemies ignore their effects. I think changing these numbers so spells fail often is a bad idea, because it makes playing a spellcaster very un-fun. It’d be better-balanced, but that’t not the only concern.
The real problem is what happens when someone fails a saving throw. If the spellcaster is being smart about spell selection, that person is, at best, out of the fight. Even at low levels, sleep or color spray literally remove people, grease takes out many creatures, glitterdust, even with the Pathfinder nerf, is still brutal. As levels get higher, the effects get worse, affect more people, or don’t allow a save at all. It turns “casting a spell” into “solve the encounter.” And there is no way to fix that aside from going through each and every spell and consider each for a total overhaul. That’s the only way to fix things, and it’s something a lot of people thought Paizo was going to do, when they were hyping up Pathfinder as a fixed 3.5. But outside of a few notorious spells (glitterdust, polymorph), they didn’t really touch the spells, and in the time since have strongly reinforced caster supremacy.
Ultimately, if this ability of spellcasters bothers you, I think you are playing the wrong system. Pathfinder was designed around the idea that spellcasters should be this powerful. Paizo has consistently pushed the envelope in terms of new and greater powers for spellcasters, while limiting and frustrating mundane characters at every turn, even to the point of errata-ing out the few tricks that start to give them a leg up. At this point, the massive gap between the two is a very-much-intentional feature of the system.