As I understand it, Kraydak's argument - from the two posts I've quoted - is that if a caster does 50% or more of a fighter's damage, then given the utility benefits that a caster brings, the marginal benefit from adding another caster to the party will significantly outstrip the benefit of adding another fighter.
I'm not sure that marginal gain is the only relevant measure - for instance, there is also the experience of the individual player of each PC (no player is playing his/her own PC as the marginal increase to the party's capacity).
My issue with making cantrips closer to fighters in power would be that, for casters to have an incentive to use their non-cantrip spells, then those spells would have to be notably better than fighter attacks. Which seems broken. From what [MENTION=43019]keterys[/MENTION] is saying, it sounds like the control effects of D&Dnext cantrips don't compensate for low damage as much as is the case with at-wills in 4e.