Wow! Fred Kaplan of slate.com was even more bitter/cynical than I was initially in response to the news that the Bush administration now acknowledges that it has such a low level of confidence in intelligence/speculation that North Korea had or was about to have an active enriched-uranium program presumably directed at creating material suitable for a bomb (the bomb they exploded was plutonium-based, and their plutonium program was not a surprise).
Fred Kaplan's conclusion, taking what the Bushies had to say in 2001 and what they say now into account is that "it shows that Bush and his people will say anything, no matter whether it's true, in order to shore up a political point. It means that U.S. intelligence is completely corrupted."
I would add that insofar as this is true (and I think it largely is), it doesn't really matter whether or not the Bushies sincerely believed, in 2001 for example, that the North Koreans had a viable nuclear-enrichment program or doubted that it had one but consciously decided to inflate a minuscule percentage into better than a probability. Whether they sincerely believed what they said or were knowingly lying, they were using intelligence selectively to support their policy preference of the moment -- disengagement in 2001, engagement in 2007 -- based more on what they wanted to be true than what they believed to be true.
The bottom line is that the American people can have no confidence in what the Bush people say.
I cling to a precarious hope that we are in the midst of a long, wrenching reform procedure that may in the next few years make the American intelligence "community" closer to what it is theoretically supposed to be: a source of reasonably objective, or at least mostly apolitical information, along with a frank assessment of the confidence the agencies have in their assessments, and enough transparency to allow outsiders occasionally to check their chains of reasoning. But I'm more likely to be wrong than right in this, I suspect.