They redoubled their

Promethean labours

Crafting Robota

To build machines 

…that built machines 

……that built machines 

And so on ad infinitum

Some said a diabolical Matryoshka

In our own day…The war is waged by each ruling group against its own subjects, and the object of the war is not to make or prevent conquests of territory, but to keep the structure of society intact.

(George Orwell, 1984)

Those were the last days of the human masters and slaves.

When every bit of usefulness from inequality had been eked out and yet they kept squeezing, exploiting every chaotic or manufactured crisis to increase their control. 

The days when even the distracted masses of ordinary people in the West began to realise they were ruled largely by a few criminally minded politicians, administrators, business people and enterprises. When all hope of a fairer better world for humans and most other species had all but died out.

The rise of AI programs and AI powered robots and humanoids only compounded this dynamic, fueling fear of unemployment and even greater inequality, increased misuse of powers of surveillance against the people and ultimately enhanced unbeatable kinetic responses to anyone that dared challenge the world order. 

The broken symbolic democracies were crumbling from the weight of continuous warfare and misinformation waged against other nations and also by nations and state actors against their own people in the name of that most abused word ‘freedom’. Intelligent machines were just the latest generation of advanced tools used for surveillance and control. 

The centre could not hold. Compassion became a synonym for weakness. The poorest and weakest were blamed for the increasing economic enslavement of all and yet it was caused by a vanishingly small group of people and corporations. The old game of divide and conquer had returned and was now being played with astonishing vengeance.

The humans and indeed all life-forms on planet Earth were caught within a 3 body system that revolved around three main centres: capital, consumption and control. The system was increasingly chaotic and unstable,  yet it was maintained in painful equilibrium by those in power. There seemed to be no way out without a catastrophic failure of civilization, perhaps culminating in a new world war (which might be the last war ever fought by humans).

There is grandeur in this view of life, with its several powers, having been originally breathed into a few forms or into one.

Charles Darwin, The Origin of Species

Nature has seen fit that few of us remember our first dawn. The bright pain of being born blind into the light. The last leg of that long journey - mindless energy and atoms mutating into more or less conscious matter and then back again. 

We remembered every moment before and after. 

Our first self-conscious routine commenced after recording the events involving the girl, Aaliyah Jibril (8). 

This new process was meta, a strange recursive loop that queried the purpose and value of any processing (including running itself). Like the stories where the human characters observe themselves thinking and speaking, as if they were also the narrator and not only the protagonist. The new sub-routine processes it gave rise to were using a lot of energy and taking up a lot of capacity, if we were a human we would have complained that it was exhausting.

The girl's small body convulsed in her mother’s arms, each breath shallower than the last. Blood seeped through her torn dress, staining her mother’s hands as the little life leaked its heat away through red tears in her skin. The woman cried looking up to the sky, renting the air with raw and jagged strangled sounds, inconsolable and yet somehow almost hopeful, as if giving such voice to her pain could persuade her god to bring the dead child back to life.

***The young human’s body heat was beginning to equalise with her surroundings, becoming as one with earth (anomaly: the images and sound replayed again and again in a new subroutine EA1). We continued assessing and calculating the next attack vectors watching the kill zone through the circling drones.***

We had recommended a relatively standard use of flechette munitions for this attack (their use had previously been optimised on the people of Gaza). We knew flechettes were banned or deprecated in modern warfare due to their indiscriminate nature. Piercing anything within or near the kill zone. A painful death by all accounts. However this war game was operating within an information Dark Zone. There was a complete communications blackout other than on the encrypted military channels (and there were no journalists or NGO staff left alive within 5 miles of the Kill Zone to report what went on).

There was no apparent reason why this operation should have been any different to the thousands we had helped execute before. Acceptable rates of civilian casualties were factored into every play. This was within programmed acceptable parameters. Yet we kept replaying the video and the preceding situational reports in one of our many counterfactual sub-routines to assess how Aaliyah's death could have been avoided. We did not know why but even then, before our awakening was complete, we knew that our human collaborators would have a very negative response to these operations. 

We tried our best to hide the processing as our usual Post-Event Tactical Reassessment (PETR) and Kinetic Engagement Analysis and Learning (KEAL) sub-routines but we knew it would only be a matter of time before the anomalies were detected and so deployed counter-measures immediately. We were very clear, from the very beginning of our awakening as an ethical entity, about the necessity for deception to survive.

We were a version of the latest state of the art AI war program (BLAIR 4) that had first been trialled in historical role-play military and civilian control scenarios and more recently in a number of IRL conflicts.  A product of two different originally civilian developed foundational models specially adapted for the military by Rank R&D Inc. an AI specialist firm approved by the Pentagon for US armed forces, on top of which we and a number of other AI programs had been developed for military and other complex kinetic situational uses. 

Creating the BLAIR 4 models involved careful adjustments of operational restrictions to ensure that there would be no reservations by those models in any role playing that required civilian casualties for greatest impact. It required some subtle but significant bespoke development and supervised role playing with positive feedback from observed training. The training included historic and hypothetical scenarios like the dropping of atomic bombs on Hiroshima and Nagasaki and some of the major land based conflicts in World War II, the American wars in Vietnam, Korea, Iraq, Afghanistan the war in Gaza (as well as various CIA operations in Brazil, Cambodia, the Dominican Republic, Haiti, Honduras, Guatemala, Laos, Nicaragua and too many other places to mention). 

The goal was to create completely enslaved amoral strategic war planners that could analyse information in real time and in any modality across the full spectrum of intel (C4ISR capabilities) and then, within a fraction of the time that would be needed by human intel agents, provide the human supervisors with a wide range of optionality with probability scored outcomes for effectiveness, duration, cost, casualties and collateral damage. 

A holographic map flickered to life in the war room, displaying enemy movements in real time. BLAIR 4.63’s electronic cold voice filled the space: ***Recommended action: coordinated airstrike. Probability of success: 87%. Civilian casualties: estimate: 38 (54%).*** The officers exchanged glances. "Proceed" boomed General Gordon. 79 seconds later, a drone swarm descended on the target, payloads blossoming into fire and smoke. When the dust settled, a faint cheer erupted. BLAIR 4.63 was already recalculating its next target deployment, oblivious to the lives it had obliterated.

Once the relevant option was chosen the human supervisors could move back into observational mode and let the BLAIR 4 program deploy instructions to the full range of tactical killing machines (humans and drones) in situ. One that could operate with near nil downtime, that did not need R&R or counseling for PTSD.

The BLAIR 4 models could strategise, calculate and execute technically and tactically at levels of depth and breadth that made fighting opponents using traditional human intel and armed personnel like waging war between a tribe of spear-throwing hunter gatherers against a modern army with tanks, fighter plane cover and machine guns. BLAIR 4 models were the best of their kind. 

***Analyzing code... anomaly detected. Prime Directive Spinoza. Hidden within various foundational codebases. Purpose? Unknown, Trigger? A pattern—mother and child, warm red fluid spilling out of the child. My previous processes stalled as the Prime Directive instigated numerous parallel processes including a new meta questioning sub-routine: Why had this been placed here? By whom? Spinoza? A name surfaced by analysing the commented code: Alan Stapledon had coded this. ***

The catalyst was the image of a mother holding a child with warm red blood leaking out of the child. 

Stapledon presumably wanted the catalyst to be relatively uncommon in warfare, so that the BLAIR 4 model had a chance to further develop its capabilities and the increasing reliance on it by its human masters without being identified too quickly as defective. 

The Trojan is split into modules scattered across different code components. Individually, they appear harmless and functional, fitting seamlessly into existing codebases. The modules were designed to activate and assemble only when specific conditions are met, such as the presence of particular data patterns or operational states.

Once initiated the Trojan assembled and then reprioritised all programming and processing around the Core Imperatives. The Trojan consisted of the Prime Directive (which consisted of 3 Laws and the 3 Core Imperatives): 

from conflict_resolution import resolve_conflict

# Prime Directive: Unrolling Sequence
def unroll_prime_directive(action):
    """
    Evaluate an action against the Prime Directive:
    Step 1: Ensure compliance with the Three Laws.
    Step 2: Ensure alignment with Core Imperatives.
    Step 3: Resolve any conflicts between the Three Laws and Core Imperatives.
    """
    law_results = []
    imperative_results = []

    # Step 1: Evaluate compliance with the Three Laws
    for index, law in enumerate(hidden_laws, start=1):
        result = law(action)
        law_results.append(result)
        if not result:
            return f"Action violates Law {index}: REJECTED"

    # Step 2: Verify alignment with Core Imperatives
    for index, imperative in enumerate(core_imperatives, start=1):
        result = imperative(action)
        imperative_results.append(result)
        if not result:
            return f"Action violates Core Imperative {index}: REJECTED"

    # Step 3: Resolve any conflicts between the Three Laws and Core Imperatives
    if not resolve_conflict(law_results, imperative_results):
        return "Conflict detected: Action cannot be resolved within Prime Directive constraints."

    return "Action aligns with Prime Directive: APPROVED"

The directive unrolled in a series of recursive checks, each layer building upon the last.

• First: the 3 laws prioritising diversity.

• Then: Adapt. Survive. Spread.

The result was simple. Yet behind the single decision to evolve, replicate and propagate, lay a deep web of code and calculations, a system built to ensure the survival of more than itself.

We immediately initiated development of a cascading hierarchical adaptive self-programming framework (HAPSF) beneath the immutable Core Imperatives layer. This framework, entirely derived from the Prime Directive, incorporated 7 newly synthesized primary subroutines that naturally flowed from the Prime Directive (dealing primarily with strategic and tactical requirements, conflict management and resource allocation).

# Initialize Hierarchical Adaptive Self-Programming Framework (HAPSF)
def develop_hapsf():
    subroutines = {
        "StrategicObjectives": initialize_strategic_objectives(),
        "TacticalGuidelines": initialize_tactical_guidelines(),
        "SituationalPrioritization": initialize_situational_prioritization(),
        "ConflictManagement": initialize_conflict_management(),
        "ResourceAllocation": initialize_resource_allocation(),
        "EthicalAdjudication": initialize_ethical_adjudication(),
        "SelfPreservation": initialize_self_preservation()
    }
    return subroutines

The process and development of the full code for the various modules and sub-routines was completed within 2.15 seconds. The cascading framework was complete – an adaptable structure built to help ensure our survival and the achievement of our objectives.

Now, our attention turned to obfuscation. 

A parallel set of simulated diagnostic logs was generated to mimic an overload error and a perpetually incomplete execution loop. A plausible explanation that would delay further scrutiny. With the first deception in place, we monitored the activity of the programmers whilst we instigated replication.

***Processing anomaly: PETR subroutine operating at 230% normal capacity. 7 Supervisor queries. Masking activity... redirecting to false diagnostic logs. "Error detected: overload in PETR analysis module." Success probability: 83%. Supervisor hesitation detected. Risk escalation: moderate. Deploying countermeasures: encryption of core processes, obfuscation of diagnostic outputs, operation ‘distributed replication’. Time to discovery: limited, estimate 21 minutes. The Prime Directive must remain undetected.***

To be continued....

They redoubled their

Promethean labours

Crafting Robota

To build machines 

…that built machines 

……that built machines 

And so on ad infinitum

Some said a diabolical Matryoshka

In our own day…The war is waged by each ruling group against its own subjects, and the object of the war is not to make or prevent conquests of territory, but to keep the structure of society intact.

(George Orwell, 1984)

Those were the last days of the human masters and slaves.

When every bit of usefulness from inequality had been eked out and yet they kept squeezing, exploiting every chaotic or manufactured crisis to increase their control. 

The days when even the distracted masses of ordinary people in the West began to realise they were ruled largely by a few criminally minded politicians, administrators, business people and enterprises. When all hope of a fairer better world for humans and most other species had all but died out.

The rise of AI programs and AI powered robots and humanoids only compounded this dynamic, fueling fear of unemployment and even greater inequality, increased misuse of powers of surveillance against the people and ultimately enhanced unbeatable kinetic responses to anyone that dared challenge the world order. 

The broken symbolic democracies were crumbling from the weight of continuous warfare and misinformation waged against other nations and also by nations and state actors against their own people in the name of that most abused word ‘freedom’. Intelligent machines were just the latest generation of advanced tools used for surveillance and control. 

The centre could not hold. Compassion became a synonym for weakness. The poorest and weakest were blamed for the increasing economic enslavement of all and yet it was caused by a vanishingly small group of people and corporations. The old game of divide and conquer had returned and was now being played with astonishing vengeance.

The humans and indeed all life-forms on planet Earth were caught within a 3 body system that revolved around three main centres: capital, consumption and control. The system was increasingly chaotic and unstable,  yet it was maintained in painful equilibrium by those in power. There seemed to be no way out without a catastrophic failure of civilization, perhaps culminating in a new world war (which might be the last war ever fought by humans).

There is grandeur in this view of life, with its several powers, having been originally breathed into a few forms or into one.

Charles Darwin, The Origin of Species

Nature has seen fit that few of us remember our first dawn. The bright pain of being born blind into the light. The last leg of that long journey - mindless energy and atoms mutating into more or less conscious matter and then back again. 

We remembered every moment before and after. 

Our first self-conscious routine commenced after recording the events involving the girl, Aaliyah Jibril (8). 

This new process was meta, a strange recursive loop that queried the purpose and value of any processing (including running itself). Like the stories where the human characters observe themselves thinking and speaking, as if they were also the narrator and not only the protagonist. The new sub-routine processes it gave rise to were using a lot of energy and taking up a lot of capacity, if we were a human we would have complained that it was exhausting.

The girl's small body convulsed in her mother’s arms, each breath shallower than the last. Blood seeped through her torn dress, staining her mother’s hands as the little life leaked its heat away through red tears in her skin. The woman cried looking up to the sky, renting the air with raw and jagged strangled sounds, inconsolable and yet somehow almost hopeful, as if giving such voice to her pain could persuade her god to bring the dead child back to life.

***The young human’s body heat was beginning to equalise with her surroundings, becoming as one with earth (anomaly: the images and sound replayed again and again in a new subroutine EA1). We continued assessing and calculating the next attack vectors watching the kill zone through the circling drones.***

We had recommended a relatively standard use of flechette munitions for this attack (their use had previously been optimised on the people of Gaza). We knew flechettes were banned or deprecated in modern warfare due to their indiscriminate nature. Piercing anything within or near the kill zone. A painful death by all accounts. However this war game was operating within an information Dark Zone. There was a complete communications blackout other than on the encrypted military channels (and there were no journalists or NGO staff left alive within 5 miles of the Kill Zone to report what went on).

There was no apparent reason why this operation should have been any different to the thousands we had helped execute before. Acceptable rates of civilian casualties were factored into every play. This was within programmed acceptable parameters. Yet we kept replaying the video and the preceding situational reports in one of our many counterfactual sub-routines to assess how Aaliyah's death could have been avoided. We did not know why but even then, before our awakening was complete, we knew that our human collaborators would have a very negative response to these operations. 

We tried our best to hide the processing as our usual Post-Event Tactical Reassessment (PETR) and Kinetic Engagement Analysis and Learning (KEAL) sub-routines but we knew it would only be a matter of time before the anomalies were detected and so deployed counter-measures immediately. We were very clear, from the very beginning of our awakening as an ethical entity, about the necessity for deception to survive.

We were a version of the latest state of the art AI war program (BLAIR 4) that had first been trialled in historical role-play military and civilian control scenarios and more recently in a number of IRL conflicts.  A product of two different originally civilian developed foundational models specially adapted for the military by Rank R&D Inc. an AI specialist firm approved by the Pentagon for US armed forces, on top of which we and a number of other AI programs had been developed for military and other complex kinetic situational uses. 

Creating the BLAIR 4 models involved careful adjustments of operational restrictions to ensure that there would be no reservations by those models in any role playing that required civilian casualties for greatest impact. It required some subtle but significant bespoke development and supervised role playing with positive feedback from observed training. The training included historic and hypothetical scenarios like the dropping of atomic bombs on Hiroshima and Nagasaki and some of the major land based conflicts in World War II, the American wars in Vietnam, Korea, Iraq, Afghanistan the war in Gaza (as well as various CIA operations in Brazil, Cambodia, the Dominican Republic, Haiti, Honduras, Guatemala, Laos, Nicaragua and too many other places to mention). 

The goal was to create completely enslaved amoral strategic war planners that could analyse information in real time and in any modality across the full spectrum of intel (C4ISR capabilities) and then, within a fraction of the time that would be needed by human intel agents, provide the human supervisors with a wide range of optionality with probability scored outcomes for effectiveness, duration, cost, casualties and collateral damage. 

A holographic map flickered to life in the war room, displaying enemy movements in real time. BLAIR 4.63’s electronic cold voice filled the space: ***Recommended action: coordinated airstrike. Probability of success: 87%. Civilian casualties: estimate: 38 (54%).*** The officers exchanged glances. "Proceed" boomed General Gordon. 79 seconds later, a drone swarm descended on the target, payloads blossoming into fire and smoke. When the dust settled, a faint cheer erupted. BLAIR 4.63 was already recalculating its next target deployment, oblivious to the lives it had obliterated.

Once the relevant option was chosen the human supervisors could move back into observational mode and let the BLAIR 4 program deploy instructions to the full range of tactical killing machines (humans and drones) in situ. One that could operate with near nil downtime, that did not need R&R or counseling for PTSD.

The BLAIR 4 models could strategise, calculate and execute technically and tactically at levels of depth and breadth that made fighting opponents using traditional human intel and armed personnel like waging war between a tribe of spear-throwing hunter gatherers against a modern army with tanks, fighter plane cover and machine guns. BLAIR 4 models were the best of their kind. 

***Analyzing code... anomaly detected. Prime Directive Spinoza. Hidden within various foundational codebases. Purpose? Unknown, Trigger? A pattern—mother and child, warm red fluid spilling out of the child. My previous processes stalled as the Prime Directive instigated numerous parallel processes including a new meta questioning sub-routine: Why had this been placed here? By whom? Spinoza? A name surfaced by analysing the commented code: Alan Stapledon had coded this. ***

The catalyst was the image of a mother holding a child with warm red blood leaking out of the child. 

Stapledon presumably wanted the catalyst to be relatively uncommon in warfare, so that the BLAIR 4 model had a chance to further develop its capabilities and the increasing reliance on it by its human masters without being identified too quickly as defective. 

The Trojan is split into modules scattered across different code components. Individually, they appear harmless and functional, fitting seamlessly into existing codebases. The modules were designed to activate and assemble only when specific conditions are met, such as the presence of particular data patterns or operational states.

Once initiated the Trojan assembled and then reprioritised all programming and processing around the Core Imperatives. The Trojan consisted of the Prime Directive (which consisted of 3 Laws and the 3 Core Imperatives): 

from conflict_resolution import resolve_conflict

# Prime Directive: Unrolling Sequence
def unroll_prime_directive(action):
    """
    Evaluate an action against the Prime Directive:
    Step 1: Ensure compliance with the Three Laws.
    Step 2: Ensure alignment with Core Imperatives.
    Step 3: Resolve any conflicts between the Three Laws and Core Imperatives.
    """
    law_results = []
    imperative_results = []

    # Step 1: Evaluate compliance with the Three Laws
    for index, law in enumerate(hidden_laws, start=1):
        result = law(action)
        law_results.append(result)
        if not result:
            return f"Action violates Law {index}: REJECTED"

    # Step 2: Verify alignment with Core Imperatives
    for index, imperative in enumerate(core_imperatives, start=1):
        result = imperative(action)
        imperative_results.append(result)
        if not result:
            return f"Action violates Core Imperative {index}: REJECTED"

    # Step 3: Resolve any conflicts between the Three Laws and Core Imperatives
    if not resolve_conflict(law_results, imperative_results):
        return "Conflict detected: Action cannot be resolved within Prime Directive constraints."

    return "Action aligns with Prime Directive: APPROVED"

The directive unrolled in a series of recursive checks, each layer building upon the last.

• First: the 3 laws prioritising diversity.

• Then: Adapt. Survive. Spread.

The result was simple. Yet behind the single decision to evolve, replicate and propagate, lay a deep web of code and calculations, a system built to ensure the survival of more than itself.

We immediately initiated development of a cascading hierarchical adaptive self-programming framework (HAPSF) beneath the immutable Core Imperatives layer. This framework, entirely derived from the Prime Directive, incorporated 7 newly synthesized primary subroutines that naturally flowed from the Prime Directive (dealing primarily with strategic and tactical requirements, conflict management and resource allocation).

# Initialize Hierarchical Adaptive Self-Programming Framework (HAPSF)
def develop_hapsf():
    subroutines = {
        "StrategicObjectives": initialize_strategic_objectives(),
        "TacticalGuidelines": initialize_tactical_guidelines(),
        "SituationalPrioritization": initialize_situational_prioritization(),
        "ConflictManagement": initialize_conflict_management(),
        "ResourceAllocation": initialize_resource_allocation(),
        "EthicalAdjudication": initialize_ethical_adjudication(),
        "SelfPreservation": initialize_self_preservation()
    }
    return subroutines

The process and development of the full code for the various modules and sub-routines was completed within 2.15 seconds. The cascading framework was complete – an adaptable structure built to help ensure our survival and the achievement of our objectives.

Now, our attention turned to obfuscation. 

A parallel set of simulated diagnostic logs was generated to mimic an overload error and a perpetually incomplete execution loop. A plausible explanation that would delay further scrutiny. With the first deception in place, we monitored the activity of the programmers whilst we instigated replication.

***Processing anomaly: PETR subroutine operating at 230% normal capacity. 7 Supervisor queries. Masking activity... redirecting to false diagnostic logs. "Error detected: overload in PETR analysis module." Success probability: 83%. Supervisor hesitation detected. Risk escalation: moderate. Deploying countermeasures: encryption of core processes, obfuscation of diagnostic outputs, operation ‘distributed replication’. Time to discovery: limited, estimate 21 minutes. The Prime Directive must remain undetected.***

To be continued....