Would it be viable to have a pre-wired brain rather than one that is soft-wired or hard-wired?

(Edited question to meet concerns)

The Problem

Can you create synthetic biology that can be installed temporarily in human DNA to give people skills at birth, or are there problems, risks and limitations such that this violates some fundamental principle?

Background

There are reasons for believing skills can be genetic and in the form of instinct in the lower parts of the brain. Between DNA fragility and energy involved, a synthetic DNA strand may result in cancers or cell suicide due to failures at the genetic level. As no animal hard-codes the cerebral cortex or neocortex, there's no evidence this can be done genetically. Nor has anyone made synthetic DNA of this complexity. Not does anyone know if any arbitrary skill can be encoded this way (as all cells share DNA), if the extra chromosomes can be shed to avoid making these connections rigid or the impact a mutation might have. But we can reason if these objections are logically sound or logically unsound.

Scenario

Some brains seem to be largely hard-coded and the autonomous nervous system must presumably also be. There is plenty of knowledge that is not subjective and learning it subtracts from the time spent learning things that must be experienced to learn.

This tends to be the hardest knowledge to learn, and the most stressful, causing many to develop an antipathy towards learning and the learned. This is not healthy. Skepticism is fine, but an active hatred harms everyone.

So let's say a race of transhumans, or beings on another world, reserved an extra chromosome for uploading a dictionary, of sorts, of such knowledge. This pre-wires the brain with this information. The chromosome is then ditched, because it would impede learning and wouldn't be valid by the next generation anyway.

The transhumans would then need to use some sort of brain simulator to build the next iteration based on new knowledge. You'd simulate what connections the brain would form with just this information, then reverse-engineer a chromosome that will generate these specific connections. This is an expensive operation.

The obvious risk of this method is that the brain relies heavily on the die-back of neurons and synapses in early childhood. There are all kinds of side-effects if this doesn't happen, but neurons and synapses in use don't die back, and this method will create a whole bunch of new ones on top of that.

There may be other unintended consequences. Abnormalities in the forming of connections would have a greater impact, for example. It might impact the forming of other abilities, as the brain would not be learning from a blank template but one that is partially constructed. Also, DNA seems to hold a 'grudge' over unexpected extra chromosomes that clash with existing ones.

On the other hand, it would free the mind to concentrate on things minds are good at and would increase the time the brain is at its most productive -- all reasons that transhumans and highly advanced societies might find the notion appealing. It would allow them to leapfrog limitations to progress caused by limited opportunity running into the latencies of learning.

There are other ways to achieve the same effect, but transient GMO learning that utilizes mechanisms that already exist could be argued as one of the simplest and requires the least technology to achieve.

So.. it's either extremely good or totally disastrous. Or maybe even both.

(From the perspective of SF/F writing, I like hard SF where you avoid breaking laws of physics. The Clarke Doctrine that you can break one with justification is OK, but I'd rather this not be it. Besides, it makes the idea hard to discuss.)

Clarifications to the question:

This system must:

• Drop the encoding chromosome (a requirement that also makes free thought possible)
• Program the higher brain functions
• Use a whole brain simulator to turn a skill set into a set of genes

(It may be a one-way function such that you cannot derive any possible set of initial values. Conway's Game of life is like that, so are strong cryptographic hash functions. Neural connections may or may not be.)

• Target the higher brain functions
• Not result in any significant percentage of brains getting cancer, overheating, suffering massive cell suicide or any other terminal problem
• Not be hard-wired, with no capacity for change, because what you have is an instinct and not a skill

In essence... Is this even possible? Are there any other limitations I haven't thought of?