When Fannie Mae and Freddie Mac not just pioneered but mandated the use of automated underwriting in the mid-nineties, some feared the 'black box' making the decisions. Minority advocates, for instance, queried whether models would make middle-class existing homeowners the standard against which they would be found lacking. No, the agencies insisted, pointing early to improved credit quality and less biased credit decision making from automated agencies. Now we learn that the government-sponsored enterprises -- long deemed "too big to fail" are essentially getting a cash injection to ensure they don't (access to the Fed's Discount Window). It has to make one wonder how technology did so little to prevent Fannie and Freddie from massive mortgage losses. We now know loan quality is bad. We know Hispanics, for example, are among the worst hit by foreclosures. Yet, it's been known since 2003 that foreclosure rates were at then record highs. Though tiny compared with today, five years ago there were signs that subprime loans were witnessing their first test in economic down times and failing. Where did automated underwriting go wrong? Did corrupt brokers or borrowers commonly falsify information, like claimed income, fed into these systems when everyone was too busy making loans to notice? Or, did the industry keep changing the rules of these rules-based systems so long as everyone was making money? In other words, did technology just served to rationalize a collective insanity?