Though I'd suggest that while Bob is correct on the tactical versus the strategic focus, he and the problem seem to forget that our political system is not conducive to long term strategies. That said, he is correct that our threat-centric intel focus is just really dumbb -- with two 'b's.

Ergo, a policy reform is required...

Entropy is correct in that many 'intel' errors are a result of lazy thinking and counterporductive (i.e. excessive classification, parochialism and turf battles, political expediency among other facets) actions by some analysts and many Bosses.

Seems like a policy reform might help...

Steve's quote from Night Watch has this gem:
"Its weak point is that implementation relies on the same people who failed, twice."
Yet another case of a policy error IMO. Fire a few "pour l'encouragement d'les autres..."

Lastly, WM hits a nail squarely:
"The breakdown occurs in my opinion when one moves from position 'the more one knows, the more one can know' (which is fine) to the position 'the more one can know, the more one must know.' "
There is no policy that explicitly says do that, rather, our policies -- and our Congress -- lead us to do that because the system has developed numerous rules and even laws to protect itself from accountability. Our deeply flawed budgetary process leads to a winner takes all approach and a 'go along - get along' attitude and set of turf allocations all too often that create a series of very discrete stovepipes that foster the idea that more is better when we should instead establish and encourage competition between agencies and units to produce meaningful intel. Reward those who get it right to spur the competitors to better efforts.

There's little doubt in my mind that sharp analysts in many agencies are delivering good product to their Bosses. The problem is they are being constrained by politically (in all senses of that word) oriented supervisors and / or units or agencies who do not want their Honcho to get upset by hearing things he or she would prefer not to hear.

That too would seem to indicate a needed policy change -- fire about half the senior people in order to get the rest to do what they should be doing instead of what they think the Boss might want.

And foster competition. While centralization will always be more efficient it will also always be less effective. In my view, effective intel trumps the 'efficient' production of something that is not really intel but is instead all too often platitudinous garb -- er, information -- of marginal value...