It's Pearl Harbor day today. Like 9/11, Pearl Harbor showed that there were fundamental flaws in our collection and dissemination of intelligence. After Pearl Harbor, we fought the war, then decided how to prevent such an event from ever happening again. The result was essentially the intelligence structure we have today.
The structure we have today is amazingly good at two things: finding and tracking significant military assets anywhere in the world (assessing conventional military capabilities) and determining the economic state of other nations (for example, the CIA was very attentive to wheat harvests in the USSR during the Cold War). Unfortunately, as 9/11 showed, our intelligence system is not that great at discovering intentions, particularly of non-State actors like terrorists. We've also seen numerous leaks of intelligence, in some cases blowing sources and methods (famously, UBL stopped using his cell phone after we let slip we were listening in on all of his calls). Worst of all, perhaps, is the realization that we have almost no human intelligence capability - even now - in the Arab/Muslim world; we cannot penetrate enemy organizations.
I don't think that the current attempt at intelligence reform is a good one. The idea seems to me to be to do something so as to be seen as doing something, rather than a serious attempt to solve the problem of discovering intentions, or to seal the perpetual leaks, or to develop methods of infiltrating enemy organizations. In other words, it's paper shuffling and empire building, rather than any kind of meaningful solution.
What would a meaningful solution look like? It's difficult to say. From an analytical standpoint, our intelligence structure breaks down into collection of information (from human agents to satellite reconnaissance to communications monitoring to reading publicly-available information), analysis of information to create intelligence (bringing together information from different sources to answer questions or uncover unexpected events), and dissemination of information to consumers (like the State Department and the Pentagon). Also, the CIA operates a paramilitary organization that takes direct action (sabotage, assassination, fostering rebellion in enemy countries and the like - Mike Spann, a CIA agent killed in the Afghanistan invasion, was from this organization). But these functions are spread across 15 agencies, with both a lot of duplication (not necessarily bad) and a lot of competition and fragmentation (bad).
It seems to me that the organization is not that great - especially with the current bill, which would add yet another layer of bureaucracy to the mix, but I don't know enough about it to suggest a sensible alternative. More critical, though, is that nothing is being done to address risk aversion. The Congress spent a couple of decades beating up on the CIA for everything it did with human agents, in one case at least forbidding the CIA to use people suspected of undertaking criminal activity (realistically, that means that the CIA could never turn a terrorist or drug runner!). This led to a culture of risk aversion within the CIA in particular, where it was better (for one's career, not for the mission or the nation) to fail to collect intelligence than to go out and get the intelligence. The Congress also put numerous roadblocks in place to slow or prevent the spread of information within the government (such as the infamous "wall" between law enforcement and counter-terrorism, so critical to the 9/11 failure), and created incentives towards parochialism in the various intelligence agencies.
This - the Congressional interference - is the primary problem to be addressed, and also the one problem Congress seems most blind to. And until we can remove intelligence from the arena of political showboating, the problem will continue, and we will therefore continue to be less safe than we could be. Perhaps part the solution here is to make all intelligence hearings closed, so that there is no opportunity for public showboating. Whatever the solution, though, the problem most urgent to address is not within the intelligence agencies themselves, but in the Congressional oversight and lawmaking around intelligence.