This is a complex issue. One of the problems is that Ardi is supposed to have lived 5.5 million years ago, according to John Hawks, but all estimates of the human-chimp divergence say it occurred between 3 and 5 million years ago. If that's true then Ardi is not in either the chimp or human lineages.
The human-chimp divergence is based on calibrating the molecule clock and that's what John addresses in his post. He seems to think that this calibration is accurate [Reviewing the clock, and phylogenomics] but I'm not so sure. Many of these studies (but not all) require calibrating the rate of change by using fixed time points inferred from the fossil record. For example, if you assume that primates and rodents last shared a common ancestor 100 million years ago then you can get a rate of change by adding up the number of changes in each lineage and dividing by 100 (substitutions per million years). Then you look at the number of substitutions in the human and chimp lineages and calculate the years since they diverged.
This is an over-simplification, as John explains on his blog, because the calibrations are also based on known mutation rates and population genetics. The theoretical models agree on a human-chimp divergence time of 3-5 million years.
I've been skeptical of the fossil record calibrations for many years because they give some very unreasonable divergence times and because the so-called "fixed" standards also seem unreasonable. The molecular clock ticks at an approximately constant rate but we just don't know what that rate is. I would have no problem accepting that humans and chimps diverged 6-7 million years ago.
[Reconstructions: Copyright 2009, J.H. Matternes.]