Michael E. Mann has released a TED talk on the web in recent days. You can watch it here: http://tedxtalks.ted.com/video/TEDxPSU-Michael-Mann-A-Look-Int
The talk dates from 13 November 2011 (according to this page: http://www.meteo.psu.edu/~mann/Mann/lectures/index.php)
Mann includes an interesting comparison between Hansen's models made in 1988 and actual temperatures. Here is the graphic which Mann shows:
Another comparison can be viewed here: http://www.c3headlines.com/climate-models/
This comparison shows poor agreement between the predictions and the actual temperatures.
Why the discrepancy?
It turns out that:
1) Mann truncated actual temperatures at 2005 (strange given the fact that the talk was recorded in late 2011)
2) Mann says that the 'medium model' was 'pretty much spot on'. Actually the 'low model' was as close to observation as the 'medium model' - and including years after 2005 shows that the 'low' model is in better agreement with observation.
3) Mann misrepresented the 'high', 'medium' and 'low' models. These were actually assuming increased, constant, or reduced CO2 output. We have not reduced CO2 output since 1988 (quite the opposite) so the appropriate model for comparison would be what Mann called the 'high model'.