My reading of [1] is that Palantir does data fusion. Their software, when installed on an organization's peripheral systems by their FDEs, centralizes all the org's data (within the org - not at palantir), and allows the org's management to do analyses on the pool.

I'm guessing that people are scared that the state will install one big palantir instance on all its systems. So that anything any part of the state learns about you, in any context or interaction, can be effortlessly used against you in every other context (perhaps via parallel construction in a lawsuit).

Basically, the fear would be that palantir makes mass surveillance data actionable, fuses surveillance programs, and incorporates most IT into mass surveillance programs.

The government would become less like a series of seperate agencies, more like a big consciousness that knows things (knows centrally, everything it was told anywhere).

Note this is just my interpretation of the fear.

Its fuzzy. Others may know more about palantir than me and thus have a more precise and grounded concern.

[1] https://archive.ph/6ljwy#selection-2539.194-2539.400

See also: https://redlib.privadency.com/r/Futurology/comments/4o02p3/o...