I'm not saying that companies should shun the CIA, so that's irrelevant. The distinction I was making was between providing generic products which can be used to implement a policy and products specifically designed to that policy's business requirements. My understanding is Palantir produces the later, and if the policy in question is problematic then they are more culpable than if they developed something generic.
Also, your link doesn't really support your characterization that Oracle Database was developed for the CIA:
> Ampex was developing a giant database for the CIA, and Ellison was assigned to the project, along with Bob Miner and Ed Oates. The CIA's codename for the project was Oracle. The project was something of a disaster, but it led eventually to the formation of Oracle Corporation and its current hugely successful range of databases.
If you pay attention, it says Ampex was the company with the CIA contract, and that project inspired Ellison to found Oracle.
"Ellison was wrong about the success of massively parallel processing computers (though he invested his own money in the idea), wrong about the success of so-called video-on-demand (as Oracle found when it ran a pilot with British Telecom), wrong about network computers replacing PCs (or at least, there's no evidence yet), and wrong about several other things - though we might have to revise that view in 10 years and say he was right but too early."
looks like history has proven that he was right about all of these things!
> Isn't Palantir's assistance clearer and more direct?
Are you aware that the name "Oracle" comes from the original CIA contract?
(crappy reference I found via wikipedia in case you didn't realise this, but it's not a secret part of company history: https://www.theguardian.com/g2/story/0,3604,215072,00.html )