View Single Post
Old 01-01-2019, 11:44 AM   #26
TGLS
 
Join Date: Jan 2014
Default Re: [GAME] Incorporate a Cyberpunk Megacorp

Quote:
Originally Posted by Daigoro View Post
Question 30 [STG] - NAI Kill-strategists
Who created Esperance? Are there similar systems around? What is the basis of its computational core?
Answer 30.1 [STG] - The MASTER Project

Esperance was created by Dr. Nicolás Ruiz Reyes, a noted expert in the field of AI, along with a team of assistants. The AI was created as part of the MASTER Project, an ambitious project to create a series of high-level expert systems to aid in day-to-day operations of the PAC. This ambitious project had purchased a number systems from General Analytics, including a number of systems larger and smaller than the 1m x 1m x 0.5m form factor. The entire project operated out of a research compound deep in the Amazon.

Esperance was the first of the AIs that were to be constructed. Plans indicated a dozen more AI were to be built, including an economics manager, a intelligence gatherer, a cyberwarfare system, a supervisor, and others. There were at least three times as many systems than AIs planned. Recovered correspondence indicates this was for spares and prototype testing.

Esperance was recovered by TIO during a raid on the compound. TIO had received some apparently false intelligence that indicated that the compound was conducting reverse-engineering on TIO's products. Only one member of Dr. Ruiz's research team was captured, who promptly killed themselves. No systems were recovered by PAC, and TIO had noted one of the systems was unaccounted for...

Answer 30.2 [STG] - GA Rationalist Series

The GA Rationalist Series is a is a line of supercomputers designed by General Analytics after the restrictions imposed in the wake of the Raphael disaster. The GA Rationalist Series is primarily intended for use with AI, but has occasionally been used for other purposes where price was no object. The GA Rationalist Series has two key design differences compared to systems from before the disaster:

1) Cryogenic operation: In order to ensure that the system does not and cannot operate without user permission, the system may only operate at particular temperature levels. The temperature control is handled by a microcontroller isolated from all other systems and only operable through a physical keypad interface on the box. There are three modes of operation: Secure off, where the computer is at too high a temperature to operate (-80°C to -140°C). The system remains at a subzero temperature to ensure to damage comes to the system while off (-200°C to -190°C). Isolated, where the computer is at a low enough temperature to operate but too low for the external interface to operate. This mode of operation is advisable while an AI installed on the computer is doing lots of "thinking", and has no reason to poll for further information. Communicating, the computer and the external interface are at a low enough temperature to operate (-180°C to -170°C). This mode automatically returns to isolated after three hours with no user input, to ensure that the user does not accidentally leave the computer in this mode. Cryostatic operation is not used solely for security; superconductors and other components require a low-temperature environment for normal operation.

2) Hardware nodal operation: Unlike many systems from before the Raphael disaster, the GA rationalist series has computation nodes organized at the hardware level instead of virtual hardware nodes organized at the software level. The long and the short of this is simple; without a steady supply of hardware, it is impossible for an AI to create more nodes than the number given to it to begin with (more nodes could be created on previous systems if the AI were to optimize nodes to a smaller load and then create more nodes in the free space). All the AI can do is reassign nodes from one purpose to another. Other advantages of this process include the faster processing and lower operating system overhead.

The GA Rationalist Series has a very active open source community. Much of this is spurred on by General Analytics itself, as a wide variety of free software drives sales of the systems. The software side of General Analytics is growing increasingly frustrated with these developments, and may take action at some point in the future...

Esperance is obviously not open source software. It runs on the GA Rationalist 4350 - Spinoza, a fourth generation system introduced in 2096. It was discovered recently that the cryogenic microcontroller had a flaw in its design that accidentally left a functional radio receiver on the device. Colonel Vieira doesn't know...

Answer 30.3 [STG] - Similar Systems

Most governments and corporations have at least considered the possibility of using expert systems to explore decision spaces for optimal choices. There even exists open source AI designed for use with manufacturing, financial, HR, and other similar decisions. Most corporations and governments have large programming departments that tweak the AI for specific purposes.

Kill-Strategy analysis, on the other hand, is not a common application for AI, and is obviously kept secret. Most of the great powers have at least one AI that could be used for such a purpose, but many do not consider it sufficiently important to dedicate a whole AI to this purpose. PAC's project would have been somewhat unique in this respect.

Counter-Kill-Strategy analysis, is something that many corporations and countries consider to be very important though. In the United States Government, there are at least 3 systems that are operating for this specific purpose; one with FBI for witness protection, one with the Secret Service to protect the president and other leaders, and one with the NSA, for a classified purpose. Many other countries have one to protect other key individuals. Counter-Kill-Strategy analysis is limited, however, and can sometimes fail. Other times it identifies targeted individuals who are not under any kind of protection. Usually, they are left to die as protecting them may reveal flaws in the AI.

Last edited by TGLS; 01-02-2019 at 12:30 PM. Reason: 30.3
TGLS is offline   Reply With Quote