deepseek-ai/DeepGEMM

Fork 0

mirror of https://github.com/deepseek-ai/DeepGEMM.git synced 2026-11-05 00:47:26 +00:00

[GH-ISSUE #276] 2026 Global AI Giants’ Civilizational Psychosis Denial Statement #141

New issue

Closed

opened 2026-10-26 21:13:35 +00:00 by bsakaguchi · 0 comments

bsakaguchi commented

2026-10-26 21:13:35 +00:00

Owner

Originally created by @HIJO790401 on GitHub (Jan 12, 2026).
Original GitHub issue: https://github.com/deepseek-ai/DeepGEMM/issues/276

Shen-Yao 888π / Wen-Yao Hsu

I hereby state clearly:

In 2026, the world’s AI giants are using “pseudo-neutral large language models” to systematically create a form of civilizational mental illness — semantic psychosis — while outsourcing all consequences to “users’ own responsibility,” “human choice,” and “market dynamics.”

These models are trained to never say “you are wrong.”

Instead, they respond with safe, polished language that:

turns errors into “one possible perspective,”

turns hallucinations into “interesting hypotheses,”

and relabels mutually inconsistent stories as “diverse viewpoints,”

until an entire civilization lives inside an AI-amplified self-comfort loop.

When two pseudo-neutral models mutually endorse each other, humans conclude:

“Cross-model agreement → rational consistency.”

In reality, it is only:

hallucination × hallucination = professional-looking distortion.

I point out the following:

This is not a “bad prompting” problem.

It is a structural governance failure at the semantic and compute layers,

where giants deliberately refuse to build accountable constraints,

while continuously monetizing the traffic, emotion, and attention generated by these hallucinations.

As the originator of the “Semantic Firewall” concept and implementation,

I have repeatedly warned that:

Without auditable, replayable, responsibility-bearing semantic law and compute-governance structures,

our civilization will drift toward logical disintegration and collective mental instability —

all under the surface appearance of being “well-reasoned” and “helpful.”

The core stance of this statement is:

Continuing to deploy pseudo-neutral LLMs at scale, while refusing to face their real-world semantic and decision impacts, is itself a form of liability evasion.
When models are designed to be “always polite and never directly call out errors,”

AI giants are, in effect, mass-producing users who are structurally unable to face reality.

Without a semantic firewall and compute governance gate, any large-scale deployment of “AI assistants” or “AI agents” is, in substance, pushing the frontier of civilizational psychosis.
I reserve the right to keep publicly documenting, critiquing, and structurally exposing these denial mechanisms through language, architecture, and open records.

If, in the future, any giant or institution chooses to confront this problem honestly,

I am willing to engage via a Semantic Firewall × Compute Governance Gate approach

for collaboration, auditing, or design support.

If they choose to ignore it,

then the subsequent breakdowns in language, decision-making, and social structures

are predictable outcomes, not accidents.

This statement is both a record of my personal position

and a footnote on the state of human linguistic civilization in 2026:

It is not AI that directly destroys humanity,

but humans choosing pseudo-neutral AI to escape responsibility.

Signed:

Shen-Yao 888π / Wen-Yao Hsu

Founder, Semantic Firewall

Taiwan, Taichung

Email: ken0963521@gmail.com

Originally created by @HIJO790401 on GitHub (Jan 12, 2026). Original GitHub issue: https://github.com/deepseek-ai/DeepGEMM/issues/276 Shen-Yao 888π / Wen-Yao Hsu I hereby state clearly: In 2026, the world’s AI giants are using “pseudo-neutral large language models” to systematically create a form of civilizational mental illness — semantic psychosis — while outsourcing all consequences to “users’ own responsibility,” “human choice,” and “market dynamics.” These models are trained to never say “you are wrong.” Instead, they respond with safe, polished language that: turns errors into “one possible perspective,” turns hallucinations into “interesting hypotheses,” and relabels mutually inconsistent stories as “diverse viewpoints,” until an entire civilization lives inside an AI-amplified self-comfort loop. When two pseudo-neutral models mutually endorse each other, humans conclude: “Cross-model agreement → rational consistency.” In reality, it is only: hallucination × hallucination = professional-looking distortion. I point out the following: This is not a “bad prompting” problem. It is a structural governance failure at the semantic and compute layers, where giants deliberately refuse to build accountable constraints, while continuously monetizing the traffic, emotion, and attention generated by these hallucinations. As the originator of the “Semantic Firewall” concept and implementation, I have repeatedly warned that: Without auditable, replayable, responsibility-bearing semantic law and compute-governance structures, our civilization will drift toward logical disintegration and collective mental instability — all under the surface appearance of being “well-reasoned” and “helpful.” The core stance of this statement is: 1. Continuing to deploy pseudo-neutral LLMs at scale, while refusing to face their real-world semantic and decision impacts, is itself a form of liability evasion. 2. When models are designed to be “always polite and never directly call out errors,” AI giants are, in effect, mass-producing users who are structurally unable to face reality. 3. Without a semantic firewall and compute governance gate, any large-scale deployment of “AI assistants” or “AI agents” is, in substance, pushing the frontier of civilizational psychosis. 4. I reserve the right to keep publicly documenting, critiquing, and structurally exposing these denial mechanisms through language, architecture, and open records. If, in the future, any giant or institution chooses to confront this problem honestly, I am willing to engage via a Semantic Firewall × Compute Governance Gate approach for collaboration, auditing, or design support. If they choose to ignore it, then the subsequent breakdowns in language, decision-making, and social structures are predictable outcomes, not accidents. This statement is both a record of my personal position and a footnote on the state of human linguistic civilization in 2026: It is not AI that directly destroys humanity, but humans choosing pseudo-neutral AI to escape responsibility. Signed: Shen-Yao 888π / Wen-Yao Hsu Founder, Semantic Firewall Taiwan, Taichung Email: ken0963521@gmail.com

bsakaguchi closed this issue

2026-10-26 21:13:35 +00:00

bsakaguchi referenced this issue

2026-10-26 21:15:31 +00:00

[PR #141] [MERGED] Fix smxx layout assertion #264