There are serious difficulties with these, particularly the former. There are definitely functional density limits to our hardware. Without getting into esoterica, I think there is a substantial amount of evidence that our brains are capable of not insignificant intelligence enhancement without chemicals (e.g. through environmental stimulation), but are largely limited by plasticity and generation of the computing substrate. At the same time, it would be difficult to coax too much additional capability out of our existing hardware; you can only drive so many synapses off a neuron, and location matters.
I think man-machine interfaces will be the best bet for extending our capability. Fortunately, that is coming along at a reasonable clip all things considered.
It's long been determined that lack of stimulation and neglect in infancy/childhood can produce a mentally-deficient individual. I think there may be things we can yet do in developing children that can more closely approach the genetic limits of their potential intelligence. There is a lot still to do in the area of discovering games and activities that may improve the intellectual development of children