Technological progess is an iterative progress. There will be thousands, with no exaggeration, of "breakthroughs" which will lead us to better things. AGI is not around the corner as in next year. However what is around the corner are more iterative improvements building upon the breakthroughs made this year.
Artificial Intelligence has come a long way from the 1980's but it still needs a way to go. Of course the good thing is that the pace is quick and the improvements many. Look at the artificial intelligence in your phone (partly in the cloud i suppose) it is absolutely incredible. Look at the increasing capabilities of robotics. Look at what Tesla is doing with robotics. Astounding. Look at self driving cars. Insane.
The progress is very quick but we are impatient and short sighted creatures (for the most part). AGI is not around the corner but they say that what we have now is narrow artificial intelligence which can only perform a few tasks. Well with each year the list of tasks that it cannot perform gets smaller.
Watson, AlphaGo, Autopilot, Siri/Cortana/Google, Robotic Manufacturing, and so many other things are quite literally breakthroughs but seem to us to be simple steps. They are steps but they are major steps. 5 years from now we will have new "breakthroughs" which will make those things seems like chess programs look today. You will be able to tell when we are nearing AGI. It will be apparent by the immense increase in capabilities from these artificially intelligent machines.
Kurzweil puts it at 2029 even though he puts his singularity date at 2045. Vinge puts his singularity date at 2030 (at the latest) and so i think that it is obvious that he believes massive improvements will be made in that time frame. Ben Goertzel said that we could achieve the Singularity in 10 years with enough investment and research and what not. Obviously a very aggressive time frame (though not actually considering the previous estimates of 2029/2030). Even other more conservative AI researchers put the date from 20-40 years away. Well within a young person's lifetime.
I say give it 3-5 years for a more clear estimate.
The growth of computation is doubly exponential growth.