Firewall often forms teams ad hoc, simply because some sentinels happen to be in the neighbourhood of a developing situation or have the right skills. Many never meet again. There are more cohesive teams among the proxies, such as the servers and the erasure squads, but even here many members are rotated. Sometimes particularly good working relationships or specialized tasks produce more permanent teams. This thread is about them. St. Cyril's Librarians "Ignorance is strength. Or at least a some safety." -"St. Cyril", founder of the Librarians The Librarians are a team of specialists drawn from erasure squads, social engineers and the vectors. Their target is bad information. Rather than trying to get rid of everything bad, they focus on what can be deleted effectively. It is nearly impossible to erase all copies of a blueprint for a WMD, viral code or plans for a seed AGI from the mesh. Worse, the act of trying to suppress something makes it interesting and might draw attention to it. The Streisand Effect is alive and kicking. But many kinds of information exist in just a few places, making it possible to quietly remove it. The Librarians spend much time monitoring just where and what information is stored in the solar system. Using front companies and indexing organisations they know or can find out who has databases of pre-Fall genetics, collects traffic data or is trying to reconstruct arXiv.org. When they discover a possible target (often a result of other Firewall operations) - an obscure scientific paper outlining a method to implode stars, meeting notes from an Argonaut meeting that reveals Firewall complicity in a "research accident", a recording of a basilisk hack - they begin to find ways to bury it forever. Or at least make it noticeable if somebody tries to access it. Just like physical firewall operations deniability is everything. Often it is easier to get rid of the information by crashing the context than attacking the data directly: no evidence suggests what the real target was. It is of course possible (and sometimes suitable) to blow up or corrupt the entire server, but merely scrambling the index may be just as effective. Especially if the scrambling is subtle and hard to initially notice, so that backup copies also get erased. Adding a new, very useful search algorithm to an open source project that just happens to penalize data of the "wrong" type subtly can both prevent discovery and earn more rep within the project - which can be used for better monitoring. Similarly discrediting sources works both for people and data. Spreading disinformation that hints that the source of the data is untrustworthy or has a bad agenda can be very useful. If a few references mentioning that the author of the star collapse paper made up many of his results and was under investigation for academic fraud when the Fall occurred (now impossible to check further) pop up together with searchers for the paper, much is won. Add evidence that someone at the Argonaut meeting was trying to pin the blame somewhere else, and most paranoid investigators will conclude that Firewall was likely just framed. And given the sheer mass of fake basilisk hacks, what is the chance of finding a real recording online? Most Librarians find their work interesting but distasteful. Many have technoprogressive leanings that make them dislike sabotaging the free flow of information and spreading disinformation. Some delight in sneakiness: drilling holes to send specialized nanoswarms into paper archives to rewrite hardcopy or hacking into project Ozmas files to corrupt little-used encryption keys is so *fun*! But they all recognize the need of limiting access to certain information. It might not be as urgent as stopping Exsurgent infections or the return of the TITANs, but it gives more breathing space for Firewall if people cannot whip up bioweapons in their kitchen or reactivate the TITAN code hidden in certain common appliances. (For some ideas of just how diverse bad information can be, see Nick Bostrom's "Information Hazards: A Typology of Potential Harms from Knowledge" http://www.nickbostrom.com/information-hazards.pdf )