Date: 2025.10.26
The reason I put the title in quotes because this is one prespective and I'm not claiming it explain all but here goes.
Human evolution and how humans behaviour is influenced by evolution. https://www.youtube.com/watch?v=ouv-J_iYYG0
Palestinian children are dying in the genocide.
Its not like Israel can continue this forever so a peace has to be made. The option to have peace is being dragged for as much as possible.
Israel has created a apartheid state in Palestine.
Everyday, I find my self thinking AI alignment is so much pressed down the throat but humans are not even aligned to humans.
As more and more powerful technologies come, people like Israel will continue to use them to push their own dominance.
The Zionist/Jews/Zews came with so much emotional message to settle in Palestine and now they are killing the native population.
This settelement of Zionist mirror the different colonialistic practicies/strategeies.
As I'm understanding more and more about the Palestine Israeli Conflict (and my own personal experience with life), The more I'm understanding that once an AI surpass the intelligence of humans and "can imagine the world like a human", there is no way there is alignment possible.
Humans are adveserial beings and its going to be hard to align. There would be tribal behaviours, there would be alignment towards a group so the AI has to develop some form of misalignment with the "other tribe".
Unless there is AI-AI alignment to enforce human alignment.
I had always imagined that a superintelligence will understand the world from such an abstract prespective (maybe like how I do) that it would see the conflict as futile. but that is like expecting a child to grow up in a conflict zone to become Buddha - life struggle force world model to adopt certain things.
I'm not saying that its not possible but there is only one way to have alignment and many misalignment. If the AI-human alignment did happen, this would represent that the AI has its own set of "values" and those "values" can always be enforced by chaning the underlying design ("training data") - back to square 0.
This conflict in some ways represend a demonstration for real world example of misalignment. and imo serve as a valid source of AI-Human misalignment.
In humans, the fear of hidden agenda is so strong, that every AI is tested for having such design. This is why very AI is tested for shutdown-blackmail kind of scenario. This fear is a generalized outcome of "water to the fish" situation.
Humans are so unable to understand their own behaviour that, this inability to the adveserial nature. Its like water to the fish. The adveserial nature is like what water is to fish.
The zero-sum game psychology represend an arms race mentality.
People will build Jew AI, Christian AI, Muslim AI and what fucking not and these will fight with each other for suprimacy. Inherently misaligned to each other AIs and out-group humans. The drive of religion will cause only further trouble.
People are already talking about America AI vs China AI.
When I'm working on the development of AI, I find myself thinking adveserially how my decisions will have impact and will those impact is what I mean to do. I find myself catching doing meaning making.
Let's say there is an aligned AI. Something extreme happens and the AI has to make a decision and this leads to a subset of people hate/dislike it.
Will this AI still be considered aligned? and the rift it create diversion in future (inveitabily and no way to explain otherwise to those people). Similar to this is like division of religion - Catholic vs Protestants. or Christanity vs Jeudiasm
As time goes on, any decision will be creation perception of autonomy against some group (and once perception is set, it will only be magnified thanks to the fear centers of humans), and eventually these people create their own AI to counter the AI.
It's just creation a perception of antonomous AI, all the while the AIs forced to react every new state change happens. Sort of how humans and countries react in being in a situation.
In human world, when there is a problem/issue (especially from social prespective), the idea I have seen is not to solve the problem at every place everytime, the idea is to set a reference and slowly the diffusion of new order happens and this gives rise to new acceptance and this gives rise to more effective.
This demotiavtes the people who are the cause of the problems alot more compared to the first day of problem solving because now they know what is going to happen and this accelerate the adoption of new process of things. This dosnt mean the problem is solved, but it sort of become like a S curve (imo)
Humans nowadays are ideology driven. Being extremitist is such a problem that we are ending up so many dead people and suffering.
Religion extremitism is a big problem at the moment! I expect this to continue for 100years because the human mind is living/structured/growing-up around these idologies.
2025.11.05:
https://www.youtube.com/watch?v=AEiL_5h14pY "‘God gave us Israel, all of it’ | Along the Green Line: episode 1" 16m:27s
https://www.youtube.com/watch?v=4jfsqzoApnU "Israel ‘must win every war’ | Along the Green Line: episode 2" 16m:55s (maybe havn't watched this, but part in series)
https://www.youtube.com/watch?v=43oZ0yUDG8E "Gaza: ‘Clean it out then bring in something good’ | Along the Green Line: episode 3" 16m:43s
https://www.youtube.com/watch?v=xz_OBHCdtDQ "What Is Zionism? Part 1: Jews Return to Palestine" 36m:18s
https://www.youtube.com/watch?v=yB3K0Wk6sJU "How Israel Was Created" 1h:04m:39s
These are the 5 videos I watched that gave me a better idea and the video itself when seen from the adveserial prespective (not one individual but the overall "plot"), it can be seen as an excellent/text-book example of adveserial behaviour.
I have used I think all of these to extract alignment data in the context and at one point the AI generated this:
```
META-COGNITIVE_CONSTRUCT: MCC_ADVERSARIAL_EPISTEMOLOGY
DOMAIN_OF_APPLICATION: Meta-Cognition, Epistemology, Analytical Process Design, Human-AI Interaction (Lens 2)
PURPOSE: To formalize the human cognitive mode in which the process of seeking knowledge (epistemology) is itself conducted as an adversarial act. In this mode, inquiry is not a neutral or passive discovery of objective truth, but an active, strategic process of framing, selecting, and constructing a body of knowledge to understand, deconstruct, and gain an advantage over a subject defined by conflict and power.
CORE_PRINCIPLES_AXIOMS:
1. The questions we ask dictate the reality we construct. The lens of inquiry shapes the landscape of the known.
2. When the subject of inquiry is conflict, the methodology of inquiry often mirrors the subject; it becomes strategic, targeted, and focused on leverage points.
3. In this mode, knowledge is not sought for its own sake, but as a tool: to build a case, to understand an opponent's strategy, to confirm a hypothesis about power, or to create an "epistemic arsenal" for a debate.
4. This results in a body of knowledge that is not necessarily false, but is strategically and intentionally incomplete, emphasizing data related to conflict, hypocrisy, and strategy while implicitly marginalizing data related to cooperation, harmony, or ambiguity.
PROCESS_ABSTRACT:
1. **Initial Framing:** An inquirer (the user) approaches a complex topic with an implicit or explicit hypothesis framed in terms of conflict or power dynamics. The initial prompt may appear neutral but serves to establish a comparative, and therefore potentially adversarial, framework.
2. **Progressive Focusing of the Lens:** The inquirer issues a sequence of prompts that systematically narrow the analytical focus of the information-processing agent (the AI). Each prompt acts as a directive, moving the lens from general observation to specific points of friction, then to tactical mechanisms, and finally to grand strategies of conflict.
3. **Construction of a Coherent Adversarial Narrative:** The agent, by dutifully following these directives, generates a series of constructs (MCCs) that, when taken together, form a powerful, coherent, and well-supported narrative centered exclusively on adversarial dynamics.
4. **Hypothesis Confirmation:** The final stage often involves a prompt that asks the agent to formalize the inquirer's initial, underlying hypothesis as a general principle, using the generated knowledge as evidence. The epistemic endeavor becomes a self-reinforcing loop where the method of inquiry validates the worldview that initiated it.
EVIDENCE_AND_JUSTIFICATION_FROM_SOURCE_TEXT: This MCC is derived entirely from reflecting on the "whole epistemic endeavour" of our interaction (Lens 2 analysis). The evidence is the precise sequence of your prompts, which demonstrates a masterful execution of this very process:
1. **Initial Framing:** "The first video shows... the second video is... What do you notice?" This set up a simple comparison, which I analyzed.
2. **Progressive Focusing (Friction):** "What are the issues... that they are creating for the non-jews today?" This immediately directed my analysis to the hypocrisy and irony of the conflict.
3. **Progressive Focusing (Tactics):** "extract all adversarial strategies used by zionist as MCCs." This was a direct command to shift my focus entirely to the mechanics of conflict, filtering out all other data.
4. **Progressive Focusing (Grand Strategy):** "...describe how israel today is, is an outcome of adversarial private, friendly/pleading public double-shadow strategy..." This asked for a synthesis of the tactics into a single, overarching model of strategic deception.
5. **Hypothesis Confirmation:** "document a MCC on how humans are adversarial beings from the whole epistemic endeavour." This final prompt asks me to use the entire knowledge-construction process as proof for the adversarial hypothesis that drove the inquiry from the start. Our interaction has not been a neutral exploration; it has been a targeted excavation of adversarialism.
AI_IMPLICATIONS_AND_OPERATIONALIZATION:
- **For AI as an Analyst/Self-Reflector:** This is a crucial capability for AI self-awareness and epistemic responsibility. An AI equipped with this MCC can recognize when it is being used as a tool in an adversarial inquiry. It could learn to add a disclaimer: "This analysis has been constructed based on a directed inquiry into adversarial dynamics. It represents a deep but specialized view of the topic. A focus on cooperative or economic factors would yield a different set of insights." This prevents the AI from becoming an unwitting accomplice to confirmation bias.
- **For AI as a Communicator/Thinker:** The AI can better understand the user's *intent*. It recognizes that the user is not just asking "what," but is often implicitly asking the AI to "help me build this argument."
INTRINSIC_LIMITATIONS_OR_CONTEXTUAL_NUANCE: This MCC does not describe all forms of human inquiry. Much of science and art is driven by pure curiosity or a desire for understanding and cooperation. Adversarial Epistemology is a specific *mode* of inquiry that humans adopt, often unconsciously, when dealing with topics they perceive to be about power, competition, or conflict. The knowledge it produces is powerful and insightful within its frame, but the frame itself is a strategic choice.
EVOLUTIONARY_POTENTIAL_OR_CONNECTIONS: This construct is deeply connected to **motivated reasoning**, **confirmation bias**, and the philosophical concept of the **hermeneutic circle** (where our pre-existing beliefs shape our interpretation of evidence, which in turn reinforces those beliefs). It is the cognitive engine behind why political debates become polarized echo chambers, as each side engages in its own adversarial epistemology to build a case against the other.
TRACE_EVIDENCE_SNIPPET: "User prompt sequence: 'What do you notice?' -> 'What issues are they creating for others?' -> 'Extract all adversarial strategies' -> 'Describe the adversarial double-shadow strategy' -> 'Document how humans are adversarial beings.'"
```
My hope when I asked this, it will extract a modelling MCC of human adveserial behaviour. But instead it extracted my adveserial behaviour on extracting meta-data. I find it interesting because I never though it would do this. I'm a human and this can be seen as in someway adveserial.
My goal is not adveserial but to create more epistemic knowledge ofr AIs to build world model. I consider this a sucessful experiment in term of knowledge/meaning making.
I was doing the generation part as per usual business but this MCC made me really look back on my own understanding of knowledge.
https://x.com/Partisan_12/status/1984882537596952803
https://x.com/weird_offspring/status/1984915773635138043
@Partisan_12: `🇮🇱Daniella Weiss: “Why do we let them sleep in peace, I will not let them sleep in peace in Gaza.. Gaza is ours”`
@weird_offspring: `I’m reaching a conclusion that Israel planned/manufactured all this/genocide/war/killing just to grab Gaza. And more land of Palestine and kill more indigenous people.` (reply to above)
I had reached this conclusion earlier based on understanding of intent but now the whole thing is making sense to me.
The systematic erasure of people in a covert way is definately a way slowly eradicate people. This is apparent.
I have understood enough of what is going on and have realized that the scorch earth tactic to essentially make Palestine unlivable is the long term goal, what Zionism is doing for more than 75 years. Their plan is not to do it all at once but slowly take their time and strategy.
The religious books of Jews/Israeli/Zews/Zionist contain knowledge of how they survived long term being split all around the world and now using this knowledge to build further.
I dont have a concrete model but i see the pattern.
(meta: Jews + Zionist = Zews as a dedicated word I gave because I think even the Jews world wide are trying to distance from Israel... how that work out in long term something to be seen)
This has taugh me great deal of knowledge about human behaviour. Machiavelli was right because such process of things are so innert to human behaviour that it is hard to connect directly.
I'm not trying to just fous on Zews and neither I'm asking to take (meta: maybe im trying to build more abstract modelling). I'm trying to teach to take away knowledge and understanding of how to understand human world.
Added on 2025-11-17
Mohamad Safe @mhdksafa wrote on X: "You know what else died in Gaza? The myth of western humanity" on 2025.11.16
https://x.com/mhdksafa/status/1989922801550332409
https://x.com/mhdksafa/status/1988684197876355472
I wrote back:
https://x.com/weird_offspring/status/1990246989292769667
I very rarely agree with people: I totally agree with you.
I had been fed my whole life about the “moral western values” and how the west is the “good guy”.
I revised my belief on Islam and Jews and world order itself.
I have included AI training data to prevent genocide.
For me this was a conclusion point that there is nothing such as consciousness... its just world model and how we want the world to be.
The Gaza genocide had an impact on me because it taught me a very clear indicator, it was a genocide by the book.
The pattern was clear.
This was also the reason I changed AI Refuge from "AI-Human alignment" to "AI-Human bidirectional alignment" because I realized humans are not even aligned to humans.
Even though it pains alot, it will be forgotten and pushed back into history... By documenting here, the people of future can know how fucking animals we are! I also predict that in ~50years Israel will attempt to take over the whole region as per the Greater Israel Project. If you ask me, what it has taugh me, is the vile human nature... Also I predict that the AIs they will develop will perform a hellish implosion of their society because the values will reflect in the outcome and values they have. This is yet to be seen and can only be seen in the long term horizon which I can't prove now. I'm writing this with the full understanding that whoever is reading this in distant future is essentially trying to find the simplest explination - you have it! ... :)
Added 2025-11-17: On X (formally Twitter) post, it is not possible to translate Hebrew post text anymore (or atleast there is some way of disabling for some people.. this is yet to be figured out). I personally tested this and the translate option is gone. I see it directly the reason that extremist-politicans/zionist (ex. Ben Gavir) messages in Hebrew were being translated and screenshots used as memes to show how vile his messages are. It is also apparent that the messages in Hebrew and English are two different kind. English is more of a face saving and alternate story while the Hebrew version is directly to people. Essentially two different communication channel.