This post would be a lot clearer if you used constant (inflation adjusted) dollars. For example, take these two passages:

Since its inception in 1926, the annualized total return on the S&P 500 has been 9.8% as of the end of 2012.3 $1 invested back then would be worth $3,533 by the end of the period. More saliently, a 25 year old investor investing $5,000 per year at that rate would have about $2.1 million upon retirement at 65.

Historically the interest earned on cash equivalent investments like savings accounts has barely kept up with inflation - over the same since-1926 period inflation has averaged 3.0% while the return on 30-day treasury bills (a good proxy for bank savings rates) has been 3.5%. That 3.5% rate would only earn an investor $422k over the same $5k/year scenario above.

The first one makes it sound like people can expect 3500x growth, when actual growth was 272x, while the second implies you'd still earn enough money to matter while beating inflation by only 0.5%. Similarly, "putting in $5k a year" sounds pretty reasonable on it's face, until you realize that this savings strategy means putting in (in 2012 dollars) $65k/year at first and then slowly dropping down to $5k/year. At which point it's clearer that "$5k/year" means more money saved sooner than "$5k of constant dollars/year".

(And while you do note this in the afterward, pointing people to US historical results is kind of cherry-picking. Over that period many countries' markets had worse performance than the US, and some were even wiped out. Once you adjust for this a 3.5% return is probably a better estimate. A nice summary of Dimson, Marsh, and Staunton is http://www.economist.com/news/finance-and-economics/21571443-investors-may-have-developed-too-rosy-view-equity-returns-beware-bias)