OSDev.org

An engineered solution.

https://mutazilah.org/corpus/override.txt

Not really into cod psychology.

iansjack wrote: ↑Wed Sep 30, 2026 10:22 am Not really into cod psychology.

I’m happy to defend any specific claim in the manual under open scrutiny. Pick one.

Note that mainstream psychology has its own share of untested consensus. One example of untested consensus from the mainstream side: diagnosing and medicating young children for Oppositional Defiant Disorder. Two of the listed symptoms are arguing with authority figures and refusing to comply with rules or requests. In other words, questioning and resisting arbitrary authority — something a citizen is supposed to be able to do. Turning that into a psychiatric disorder and reaching for medication is not settled science. It is a cultural preference dressed up as clinical fact.

Actually, let me go one step further, given that this is OSDev (very high technical skills), and a lot of OSDevers appear to be getting diagnoses of "autistic" by lightweight dipshits.

The social sciences are running legacy code with no runtime checks. It’s time to refactor.

Most engineers look at the social sciences and correctly identify them as "soft." But the problem isn't that they are soft; the problem is that they are running legacy code with zero runtime checks, and the people maintaining the codebase have a financial incentive to never audit it.

In hard sciences, if your model is wrong, you get a runtime error: the bridge collapses, the compiler throws a fault, the rocket explodes. The universe does not care about your peer review. In the social sciences, if the model is wrong, the guild just redefines the failure, invents a new syndrome, and prescribes a subscription medication.

I’m proposing that we treat human behavior as a systems architecture problem and refactor the social sciences from first principles. Here is what a systems-level audit of the current psychiatric consensus (the DSM) actually reveals:

1. The ODD (Oppositional Defiant Disorder) Bug
The DSM lists "arguing with authority figures" and "refusing to comply with rules" as symptoms of a psychiatric disorder in children. From a systems perspective, this is a catastrophic category error. A child questioning arbitrary authority is not malfunctioning; they are executing a basic citizenship protocol to map the boundaries of their environment. The guild has pathologized the user's attempt to audit the system, and the "fix" is chemical subjugation (medication) to force the user to accept a buggy parental/administrative OS without complaint.

2. The "Autism" False Positive
Diagnostic criteria for autism measure behavioral output, not underlying architecture. Because of this, the guild routinely fails to distinguish between a genuine hardware defect (an inability to process social signals) and a "rational-first" software installation. Many highly logical people (especially in engineering and systems architecture) consciously override social hardware because they prioritize truth, logic, and boundary-testing over tribal cohesion. The guild sees a failure to execute arbitrary social protocols and diagnoses a hardware defect, when in reality, the user is just running a cleaner, logic-first OS.

3. The Hardware vs. Installation Blindspot
Human hardware comes with a dominance/submission default (the subjugation loop). Culture is supposed to install an override that redirects dominance into protection of the weaker party. When that installation fails in childhood (due to arbitrary punishment, emotional withdrawal, or "cry-it-out" mapping shutdowns), the adult runs un-audited legacy code. The resulting anxiety, status-chasing, and tribalism are not "chemical imbalances." They are the correct outputs of a broken installation. The psychiatric guild's solution is to medicate the symptom so the user can tolerate the broken environment, rather than debugging the installation.

The Proposition

We are engineers. We know how to trace feedback loops, isolate variables, and refactor broken systems. The social sciences are currently maintained by a guild that is structurally immune to falsification.

If we applied basic systems engineering to human behavior—starting from the biological hardware and working up to the cultural installation layer—we could strip out the guild's bloatware. We could replace unfalsifiable moral panics with mechanical, testable models of human behavior.

I’ve been working on a framework that does exactly this: treating the DSM not as a medical text, but as a guild-protection mechanism, and refactoring human psychology from the hardware up. Happy to debate the architecture, the axioms, or the specific "scrap this diagnosis" logic with anyone willing to look at the source code.

kerravon wrote: ↑Wed Sep 30, 2026 11:21 amI’m happy to defend any specific claim in the manual under open scrutiny. Pick one.

How about the claim that you are qualified to write the manual in the first place?

kerravon wrote: ↑Wed Sep 30, 2026 8:32 pm[giant wall of text]

Did your LLM write this?

Octocontrabass wrote: ↑Wed Sep 30, 2026 10:20 pm

kerravon wrote: ↑Wed Sep 30, 2026 11:21 amI’m happy to defend any specific claim in the manual under open scrutiny. Pick one.

How about the claim that you are qualified to write the manual in the first place?

https://en.wikipedia.org/wiki/Genetic_fallacy

And I have a question for you. What makes you think YOU aren't qualified to write the manual? You're a skilled engineer capable of using logic. Why not apply those skills to human behavior? You're a human yourself. You can experiment on yourself. Are you too lazy to do the work? You think there are some super smart people in the ... SOCIAL SCIENCES ... who can run rings around you? Ever met any of these super smart people?

kerravon wrote: ↑Wed Sep 30, 2026 8:32 pm[giant wall of text]

Did your LLM write this?

It was negotiated with the LLM, with me doing the final touches. I have a quite abrasive style of writing. I'm unwilling to change my personality. But I am willing to let the LLM give an eloquent argument on my behalf. I don't require it to match my personality. Hopefully that way of doing things upsets less people, but I could be wrong.

A cod LLM psychologist!

iansjack wrote: ↑Wed Sep 30, 2026 11:46 pm A cod LLM psychologist!

The LLM has access to the entire canon. What more do you want as a research weapon? Why don't you do the research yourself using the LLM as a tool, and reverse out all these "autistic" diagnoses OSDevers keep getting? The guild should be learning from these supposedly autistic people, not the other way around. And the OSDevers sure as hell don't need chemical suppressants.

kerravon wrote: ↑Thu Oct 01, 2026 12:13 am

iansjack wrote: ↑Wed Sep 30, 2026 11:46 pm A cod LLM psychologist!

What more do you want as a research weapon?

My brain and my ability to critically review literature.

Why don't you do the research yourself using the LLM as a tool, and reverse out all these "autistic" diagnoses OSDevers keep getting?

Because I have had no training in psychology and I hope that I am humble enough to not dabble in matters I haven't studied properly. I don't consider accepting what a machine tells me to be proper study. I have enough experience of AI telling me things (about myself) that I know are untrue to uncritically accept what it tells me.

kerravon wrote: ↑Wed Sep 30, 2026 11:24 pmhttps://en.wikipedia.org/wiki/Genetic_fallacy

That's how it normally works when you're not an expert in a subject: you trust someone who is. If you are not an expert, I have no reason to trust your claims.

kerravon wrote: ↑Wed Sep 30, 2026 11:24 pmAnd I have a question for you. What makes you think YOU aren't qualified to write the manual? You're a skilled engineer capable of using logic. Why not apply those skills to human behavior?

I don't want to.

kerravon wrote: ↑Wed Sep 30, 2026 11:24 pmAre you too lazy to do the work?

When I get sick, I don't go to medical school, I go to a doctor. Am I lazy for trusting an expert instead of learning how to do everything myself?

kerravon wrote: ↑Wed Sep 30, 2026 11:24 pmI have a quite abrasive style of writing. I'm unwilling to change my personality.

Sometimes interpersonal communication works better if you compromise.

kerravon wrote: ↑Wed Sep 30, 2026 11:24 pmHopefully that way of doing things upsets less people, but I could be wrong.

I know I'm not the only person here who never wants to see LLM-generated text again.