Wonder in Beyond Earth
|Requires|| 4 Firaxite|
|Notes||Military Units fight at full strength even when damaged.|
The Markov Eclipse will permanently tie up 4 Firaxite resources.
"Your most valuable human asset has one foot in the code and one foot in the world and is still on solid ground." - Suzanne Marjorie Fielding, Transplanetary Management, Methods, and Resources
The exponential evolutions of battlefield AI did not lead to lopsided military victories as many armchair generals had predicted. A counter-trend of "hard pure" military theory enjoyed a certain vogue, accompanied by a penchant for detonating EMP weapons near any metal box with a power feed going into it. But a handful of military scientists looked at how AI and human soldiers could be brought into closer coordination and arrived at the conclusion that with careful training, a human mind could be trained to operate in close harmony with certain classes of AI. From this finding came the Markov Eclipse: The superposition of the human mind on the AI.
A small series of Markovian decision-making AI models had been in use for some generations, providing reliable if somewhat unexciting advantage to holistic analyses of well-described battlespace problems. What was unique about these were their parallel state descriptors, projecting not only the best-guess current state, but a Bayesian array of most likely alternatives. When sudden changes to the battlescape appeared, the AI would simply switch to the alternative chain that best described the current situation, and updated accordingly. This turned out to have a psychophysiological equivalent in certain brain structures associated with alternate-world formulation in the human prefrontal cortex.
From there, it was a simple matter of developing a system of mnemonic training and deep-perceptual visualization interfaces for the human operators. With sufficient training (9 years being modal, with acceptable results occurring as early as Year 4 in precocious operators), an operator could view a highly schematized representation of the AI battlespace model and intuitively process what the AI was seeing. Fine-resolution positron emission feedback from the operator's brain would go back to the AI, which could update some of its assumptions regarding human psychological elements, and quickly update its model. Together, a human and AI system formed a powerful team, capable of processing massive bands of probability data and directing troops with a minimum of wasted time and surprise.
The Markov Eclipse were an elite cadre, each operator and AI team more valuable than a division of soldiers or an entire gunboat squadron. The Markov Eclipse commanded universal respect, capable of overriding the orders of anyone in the line of command, their decisions subject only to the political review of the state. Years of intensive, brain-shaping training and daily proximity to AI "personality" often left them somewhat aloof and detached, and the awe in which their fellow soldiers regarded them merely contributed to their Olympian status.
The exemplar tale of the power of the Markov Eclipse comes from the siege of Drees, where the First Battalion, 3rd Marines, pinned down inside the station's primary reactor control building, managed to link into the security controls and improvise a burst transmitter that linked back to Markov Eclipse. Each marine's weapon AI and soldier com immediately lit up, restored to functionality, and individuals began receiving incredibly specific commands: "Suppress at sector B7. Expend no more than 37 rounds. Fire artillery at sector A3 for effect, fragmentation. Fire A9, guided munition." Certain marines in marginal psychological states reported hearing patriotic music, or messages from old comrades over voxcom. The battle lasted over twenty hours, at the end of which a common message came: "Hostile forces neutralized. End transmission." The Markov Eclipse was gone, the marines of First Battalion were alive, and scattered about Drees were the shattered and burning remains of two armored divisions.