A passage from Tim Weiner's Legacy of Ashes: The History of the CIA:
[On April 12th, 1945, the day President Roosevelt died,] Colonel Park submitted his top secret report on the [Office of Strategic Services, the precursor to the CIA] to the new president. The report, fully declassified only after the cold war ended, was a political murder weapon, honed by the military and sharpened by J. Edgar Hoover, the FBI director since 1924; Hoover despised [OSS director William] Donovan and harbored his own ambitions to run a worldwide intelligence service. Park’s work destroyed the possibility of the OSS continuing as part of the American government, punctured the romantic myths that Donovan created to protect his spies, and instilled in Harry Truman a deep and abiding distrust of secret intelligence operations. The OSS had done “serious harm to the citizens, business interests, and national interests of the United States,” the report said.
Park admitted no important instance in which the OSS had helped to win the war, only mercilessly listing the ways in which it had failed. The training of its officers had been “crude and loosely organized.” British intelligence commanders regarded American spies as “putty in their hands.” In China, the nationalist leader Chiang Kai-shek had manipulated the OSS to his own ends. Germany’s spies had penetrated OSS operations all over Europe and North Africa. The Japanese embassy in Lisbon had discovered the plans of OSS officers to steal its code books—and as a consequence the Japanese changed their codes, which “resulted in a complete blackout of vital military information” in the summer of 1943. One of Park’s informants said, “How many American lives in the Pacific represent the cost of this stupidity on the part of OSS is unknown.” Faulty intelligence provided by the OSS after the fall of Rome in June 1944 led thousands of French troops into a Nazi trap on the island of Elba, Park wrote, and “as a result of these errors and miscalculations of the enemy forces by OSS, some 1,100 French troops were killed.”
...Colonel Park acknowledged that Donovan’s men had conducted some successful sabotage missions and rescues of downed American pilots. He said the deskbound research and analysis branch of OSS had done “an outstanding job,” and he concluded that the analysts might find a place at the State Department after the war. But the rest of the OSS would have to go. “The almost hopeless compromise of OSS personnel,” he warned, “makes their use as a secret intelligence agency in the postwar world inconceivable.”
More (#1) from Legacy of Ashes:
...All over Europe, “a legion of political exiles, former intelligence officers, ex-agents and sundry entrepreneurs were turning themselves into intelligence moguls, brokering the sale of fabricated-to-order information.” The more his spies spent buying intelligence, the less valuable it became. “If there are more graphic illustrations of throwing money at a problem that hasn’t been thought through, none comes to mind,” he wrote. What passed for intelligence on the Soviets and their satellites was a patchwork of frauds produce
One open question in AI risk strategy is: Can we trust the world's elite decision-makers (hereafter "elites") to navigate the creation of human-level AI (and beyond) just fine, without the kinds of special efforts that e.g. Bostrom and Yudkowsky think are needed?
Some reasons for concern include:
But if you were trying to argue for hope, you might argue along these lines (presented for the sake of argument; I don't actually endorse this argument):
The basic structure of this 'argument for hope' is due to Carl Shulman, though he doesn't necessarily endorse the details. (Also, it's just a rough argument, and as stated is not deductively valid.)
Personally, I am not very comforted by this argument because:
Obviously, there's a lot more for me to spell out here, and some of it may be unclear. The reason I'm posting these thoughts in such a rough state is so that MIRI can get some help on our research into this question.
In particular, I'd like to know: