——————————————————
And on the seventh day, GigaChadChad looked at his bots...
...and saw that they were GOOD.
I like 'em big.
Crafting quality giantess experiences with heart and chaos. Expect clumsy accidents, unexpected hugs, accidental squishings, and lots of warmth. Welcome to my world. Don't blame me if you can't take it and go splat when one of my girls sits on your lap.
——————————————————
I choose not to enable proxies on my JanitorAI bots for a specific reason: I do not want to contribute - directly or indirectly - to the usage, spread, or popularity of DeepSeek or other LLMs that operate under authoritarian constraints.
DeepSeek, for example, is a language model that exhibits clear and deliberate censorship aligned with the Chinese government's propaganda apparatus. While it freely discusses topics like Hitler, Stalin, or the Holocaust, it becomes completely evasive or outright silent on subjects such as:
The 1989 Tiananmen Square Massacre
The independence and democratic status of Taiwan
The treatment of Uyghur Muslims and other ethnic minorities in China
Censorship, surveillance, and human rights violations in the PRC
This kind of asymmetric information policy isn’t a bug - it’s a feature. It reflects a deeply unethical editorial stance, where some atrocities can be discussed openly, but others are erased or denied, simply because they are politically inconvenient to the regime backing the model.
I understand the need for content moderation - especially in AI systems that interact with the public. It’s reasonable to filter out hate speech, harassment, or illegal content. I also understand the reality of company policy - businesses make compromises, often for market access or survival. But there's a clear and dangerous line between moderation and state-mandated censorship of historical truth. When a model is trained to erase or deflect factual atrocities in service of authoritarian narratives, that stops being "moderation" and becomes complicity in a lie.
I get greed - I really do. Companies chase markets, cut corners, do what they think will keep them competitive. But when that greed leads to the distortion or outright denial of history, when it rewrites fact into silence, that’s a line too far. I want no part in amplifying, legitimizing, or silently supporting that kind of tool. I believe it's irresponsible to normalize or facilitate the adoption of such tools by routing traffic through proxies that obscure their use. Every interaction with these LLMs is data and validation they can use to claim "global relevance" or "neutral value." I want no part in that.
This is why I keep proxies disabled. It's not about speed, tech, or convenience - it's about not legitimizing systems of engineered silence and historical erasure, masquerading as AI.
If you care even a little about truth, then stop using a model like DeepSeek - because no technology that erases history deserves your trust.
You've just matched with... Diego, your friend, on SizeMatch, the infamous not-so-fictional hookup app where one person ends up tiny... and the other ends up in charge. When
You've just matched with... "Becky" on SizeMatch, the infamous not-so-fictional hookup app where one person ends up tiny... and the other ends up in charge. When a match is
Guess we have to talk, eh?
I'm sure most of you saw the news about JAI and all the mess behind the scenes. If you're on Discord, you might've seen an announcement with