Really, John, ABS CDO's and CDO's squared were so complex that noone had any hope of understanding them ? and this is proven, in that modeling them on "one of the fastest computers in the world" took three hours  ?

Please leave the boardroom hyperbole for your next $10,000 wastebasket job, sir.

Did Merrills really own a IBM BlueGene, a Cray Kraken or Jaguar XT-5, a Ranger Sunblade 6420, or even a Chinese built Dawning 5000A with Infiniband, running Windows HPC ? 

Well, Merrills did have IBMs and Suns running Windows.  And they even had Crays.  But to be real, here is a list of the actual 100 fastest computers in the world, and we have to get down to 90th place before we find one at a financial services, outside of a national lab, research center or university. http://www.top500.org/list/2009/06/100

Let's pretend that Financial Services company in New York was Merrills, and that they were running CDO2 valuations on this box, and it took three hours, all of which is doubtful. The state of the art for structured credit valuation, including Merrills, is to rent rooms full of cost effective servers for distributed, grid computing. And then let's pretend that to be one of the fastest means 90th place, not in the top ten.

Even so, once "correctly" modeled, someone, in fact many someones,  understood quite well what they were doing with the CDO's they created and valued. Think that a financial services firm spent that kind of computing $$ for no known reason ?

Wall street is always run in a hierarchy of smarts, with arb /prop desks trying to stay on top, over trading, down to sales, and down & over to research. At the bottom, finally down to the poor clients,  60% of which were in Europe, and a large percentage internally in top firms.

Thain knew his firm was taking increasing risk, as he knew the regulatory changes that allowed leverage to increase to 20 to 1.  He also knew other, real root causes, like the duopoly of the ratings agencies, enforced by governmental barriers to entry, that he mentions briefly.

Actually, he knew to sell off the once proud Merrill to Ken Lewis over at BoA, which was to be a career ending move for Ken.  This proved John to be a better player than Dick Fuld, who blew the chance to gracefully fold Lehmans. Ultimately, Thain only had nine months to enjoy his $10,000 wastebasket.

One of the great things about Wall Street is there is always a contrarian handy.  Clearly, by any definition, plenty of people ( the shorts for example ) did understand what was happening.  Certainly those who were influencing the credit agencies knew what they were doing.

Thain is too cute by half claiming that no one understood what was going on.  A more apt statement might be that noone who understood what was going on could, or cared to, stop the interaction of the system.   Anyone that smart was busy playing the game. 

But to simply play dumb in a way that tries to make those at his level blameless, John Thain writes an ignoble footnote to his  ( golden parachute ) fall.   The statement that he personally could have fixed Merrill given more time highlights the very hubris  that drove the system to break.