1) The rate of change not being constant is not significant as long as its variation is normally distributed about some mean. I expect that changes in environment are randomly distributed.

2) This is too small scale to have any impact on the trajectory of the numerical growth in number of base pairs.

We don't need to guess, life isn't special. The growth of complexity under favorable conditions is observable and occurs all over the place at the same rate, at scale.