dh5a cells is what I am typically using and the 30-45 minute outgrowth or recovery period is a nuisance and I have heard that it is not always necessary depending on the mode of action of your antibiotics resistance gene.
Yes. Colonies derived from cells without recovery are just as "good" as those derived from cells with recovery. In my hands, that's true for simple plasmids such as pUC, as well as long and difficult constructs, like lentivector. As long as you get enough colonies to do what you want, you should be fine.
It will probably depend on the type of competent cells you are using, but the Zymoresearch Mix N' Go kit specifically states "When selecting with Kanamycin, Tetracycline, etc., an outgrowth performed in SOC medium is required for efficient transformation. In most cases, this step can be omitted when selecting with Ampicillin."
In the end it will also depends on whether you are transforming a pure plasmid (high transformation efficiency) or a ligation or Gibson reaction (low transformation efficiency) and how competent youre cells are (TSS, Mix N' Go or other chemical competence vs electrocompetent)