Mr Cooper Group Inc (COOP)

[955]

AI questions search from available servers they have access to. When you think about data (information) it's hot/warm/cold. Or recent, month to 2, 2-6, longer. When institutions, governments, entities want data "scrubbed", they locate all the sources of the information known, then eliminate it from those sources, modify it, or intercept the requests electronically. Usually the mediums are tape, hard drives, solid state drives, NVRAM, or RAM on constant power sources. There is also paper, microfische, other forms of physical data (for old or very old data). Unfortunately, redacting is made too easy and this is the preemptive method lawyers and institutions prefer. Once data is out, it is extremely hard to control as storage is relatively cheap and many institutions and government just make copies of everything these days. The National Archives digitized their data on 200 Million pages which is nothing on todays storage. ~400GB uncompressed. The other major issue is data reliability. If an algorithm is piecing information, it can create answers that are somewhat broken or misleading. The chatGPT response that Rag caught (post 709673) is an example. It takes a lot of power and access to actually "scrub" data completely so when using AI tools you are really just creating search strings that are interpreted by algorithms. Now AI that can reproduce redacted data... an area having resources poured into it, that may be where answers provide value. TC DBD

It's all about deception and directing the narrative for control and financial gain.