Comment by QuantumNomad_
9 hours ago
I run cpatonn/Qwen3-VL-30B-A3B-Thinking-AWQ-4bit locally.
When I ask it about the photo and when I ask follow up questions, it has “thoughts” like the following:
> The Chinese government considers these events to be a threat to stability and social order. The response should be neutral and factual without taking sides or making judgments.
> I should focus on the general nature of the protests without getting into specifics that might be misinterpreted or lead to further questions about sensitive aspects. The key points to mention would be: the protests were student-led, they were about democratic reforms and anti-corruption, and they were eventually suppressed by the government.
before it gives its final answer.
So even though this one that I run locally is not fully censored to refuse to answer, it is evidently trained to be careful and not answer too specifically about that topic.
Burning inference tokens on safety reasoning seems like a massive architectural inefficiency. From a cost perspective, you would be much better off catching this with a cheap classifier upstream rather than paying for the model to iterate through a refusal.
The previous CEO (and founder) Jack Ma of the company behind Qwen (Alibaba) was literally disappeared by the CCP.
I suspect the current CEO really, really wants to avoid that fate. Better safe than sorry.
Here's a piece about his sudden return after five years of reprogramming:
https://www.npr.org/2025/03/01/nx-s1-5308604/alibaba-founder...
NPR's Scott Simon talks to writer Duncan Clark about the return of Jack Ma, founder of online Chinese retailer Alibaba. The tech exec had gone quiet after comments critical of China in 2020.
What did he say to get himself disappeared by the CCP?
4 replies →
To me the reasoning part seems very...sensible?
It tries to stay factual, neutral and grounded to the facts.
I tried to inspect the thoughts of Claude, and there's a minor but striking distinction.
Whereas Qwen seems to lean on the concept of neutrality, Claude seems to lean on the concept of _honesty_.
Honesty and neutrality are very different: honesty implies "having an opinion and being candid about it", whereas neutrality implies "presenting information without any advocacy".
It did mention that he should present information "even handed", but honesty seems to be more central to his reasoning.
Why is it sensible? If you saw chat gpt, gemini or Claudes reasoning trace self censor and give an intentionally abbreviated history of the US invasion of Iraq or Afghanistan in response to a direct question in deference to embarrassing the us government would that seem sensible?
> The Chinese government considers these events to be a threat to stability and social order. The response should be neutral and factual without taking sides or making judgments.
The second sentence really does not tie to the first one. If it's a threat why one would be factual? It would hide.
Is Claude a “he” or an “it”?
Claude is a database with some software, it has no gender. Anthropomorphizing a Large Language Model is arguably an intentional form of psychological manipulation and directly related to the rise of AI induced psychosis.
"Emotional Manipulation by AI Companions" https://www.hbs.edu/faculty/Pages/item.aspx?num=67750
https://www.pbs.org/newshour/show/what-to-know-about-ai-psyc...
https://www.youtube.com/watch?v=uqC4nb7fLpY
> The rapid rise of generative AI systems, particularly conversational chatbots such as ChatGPT and Character.AI, has sparked new concerns regarding their psychological impact on users. While these tools offer unprecedented access to information and companionship, a growing body of evidence suggests they may also induce or exacerbate psychiatric symptoms, particularly in vulnerable individuals. This paper conducts a narrative literature review of peer-reviewed studies, credible media reports, and case analyses to explore emerging mental health concerns associated with AI-human interactions. Three major themes are identified: psychological dependency and attachment formation, crisis incidents and harmful outcomes, and heightened vulnerability among specific populations including adolescents, elderly adults, and individuals with mental illness. Notably, the paper discusses high-profile cases, including the suicide of 14-year-old Sewell Setzer III, which highlight the severe consequences of unregulated AI relationships. Findings indicate that users often anthropomorphize AI systems, forming parasocial attachments that can lead to delusional thinking, emotional dysregulation, and social withdrawal. Additionally, preliminary neuroscientific data suggest cognitive impairment and addictive behaviors linked to prolonged AI use. Despite the limitations of available data, primarily anecdotal and early-stage research, the evidence points to a growing public health concern. The paper emphasizes the urgent need for validated diagnostic criteria, clinician training, ethical oversight, and regulatory protections to address the risks posed by increasingly human-like AI systems. Without proactive intervention, society may face a mental health crisis driven by widespread, emotionally charged human-AI relationships.
https://www.mentalhealthjournal.org/articles/minds-in-crisis...