> Why 2 MB ?
>

Why 20 MB? Do you anticipate 20x transaction count growth in 2016?

Why not grow it by 1 MB per year?
This is a safer option, I don't think that anybody claims that 2 MB blocks
will be a problem.
And in 10 years when we get to 10 MB we'll get more evidence as to whether
network can handle 10 MB blocks.

So this might be a solution which would satisfy both sides:
  *  people who are concerned about block size growth will have an
opportunity to stop it before it grows too much (e.g. with a soft fork),
  *  while people who want bigger blocks will get an equivalent of 25% per
year growth within the first 10 years, which isn't bad, is it?

So far I haven't heard any valid arguments against linear growth.