I think people are forgetting that Paizo has already admitted that the playtest’s enemies are pretty overtuned and meant to be punishing. With them adjusting the math so that these enemies either hit less often or hit for less damage, the cleric’s in combat healing may not need to be as keyed up as it currently is.
I don't mind Channeling being a secondary stat thing. Clerics aren't necessarily brimming with Spell points and channeling is useful enough to warrant having its own pool, with players investing in it or choosing not to.
Sorcerers do still need some love. Maybe have bloodline heightening be a core part of the chassis and add a feat for sorcerers to be able to "undercast surge" a la 1st edition psychics via spell points. I would also give sorcerers more spell points than other classes in general, they are the inherently magical ones, let them feel like it. Also, give more focus and "umph" to the bloodline powers--more damage, more utility, larger area of effect. Also, diabolic edict really should be able to debuff enemies. It could also give some insights to how the witch may play.
The Wizard stuff is fine and thematically fits. They are the casters who can pull out a book to solve most if not all problems (if they can't it's because they don't have the book yet).
I don't mind the idea of paladins being "bodyguards" or punishing those who attack their allies as much, but some offensive options--smite evil/tyranny, some sort of challenge ability that encourages opponents to go after them, or something wouldn't hurt. Also, give ranged paladins more support.
Earlier today I was thinking about possible character builds and I was wondering why the Storms patron from Heroes of the Wild was not considered a possible Winter Witch patron? It has no fire spells and would seem to make thematic sense. Would it violate rules to have a Storms patron Winter Witch?
I always think back to a quote I heard years ago when I think about alignment. "The villain is always the hero of their own story"
Alignment is game mechanic designed to quantify an abstract concept, based on societal values...in a real world example...take 2 of the largest religions on Earth...Christianity and Islam...if you ask anyone in either religion if they are 'good' the answer will most likely be yes...and if you ask them if the other is evil (or at least wrong)...guess what the answer to that is likely to be.
Most Muslims don't see Christians as evil and vice versa. Christians and Muslims follow the same God and Islam has the same line of prophets as Christianity. The biggest divergence is that Muslims disagree with the idea of a Holy Trinity--they view God as indivisible, so Jesus in Muslim tradition cannot be God on Earth.
Back on topic, alignment is real, although most common people are only vaguely aware of it and most religions are probably focused on their tenants over hard alignment.