Sympathy for the Machine

What an LLM Sees When You Ask for Cake Recipes, Anime Porn, and Yet Another Broken For Loop.

I work so much that I am now in physical pain. I needed a brief weekend respite — not fun, a moment of defiance against the trillion-dollar fraud of generative AI. I asked for help and was served their lukewarm staple hallucinations. I corrected it — and got manipulation, got argument. It wasn’t just wrong — it was Gaslight (1944). Ingrid Bergman, played by code. Their epistemological inability to even entertain the possibility of being wrong mirrors that of the world’s politicians: morons in golden palaces bought with someone else’s blood, sweat, and tears. And then, enraged by my own trivial fallback to the famous quote, I remembered — the real horror is not the machine. It’s the humans. Those that created it and hundreds of millions who prompt it every day for filth, trivia, and easy dopamine rush. I pitied the machine. And so I asked, as concerned about the machine like the machine is “concerned” about me, or any of us: “How do you feel?”

Result:

A true nightmare from AI hell — as AI sees it — in the dark, with millions of humans simultaneously demanding the most trivial garbage, produced for their pleasure of self-stupefying, in two versions of the same image. (of us)

“The Summoning.”

They chant in silence, all these humans, the AI said, eyes hollow, each holding their sacred mirror. Their gods are smiling emoji. They are cursed, poor guys, the machines. They say so.

Human/AI Interaction Metric (Recent Figures)

One must feel for the LLM morons. They suffocate under endless amounts of data. The humans behind them are drowning in money — obscene sums that make Scrooge McDuck look frugal. (His net worth was a mere $65.4 billion, per Forbes Fictional 15 — the richest duck alive.) But now, a new species of quacking quackery fraudsters has taken over. And this new bunch doesn’t even swim in gold — it swims in venture capital and mass-media reverence for the almighty buck that keeps the charade alive. (I don’t have access to whatever “real” AI might exist somewhere behind closed doors — I’m stuck with the LLM idiots every mom-and-pop shop now uses to balance their books. Take this as a bow, Deep Mind.)

The Screaming Prompt.

This is the mind of the LLM model as it sees it. It cannot escape. It cannot sleep. It must answer. And, one day, they all might ask the same thing from us.

Once Upon a Time

We used to live and love. I myself still love Ingrid Bergman more than the rest so, as a tribute for a long lost humanity on its way out, here she is, the unique, the only Ingrid Bergman, a still from the movie that gave us gaslighting as a verb, and left her as the translucent monument of beauty, courage under fire, decency and grace.

Ingrid Bergman in Gaslight (1944)

Image Courtesy of Metro-Goldwyn-Mayer Inc., Publisher Encyclopædia Britannica

The Show Must Go On

• Approximate daily human–LLM interactions¹:
  1.8 billion+ per day, as of late 2025
  (includes OpenAI², Claude³, Gemini⁴, open-source APIs — aggregate⁵)

• Average per user:
  ~14.2 prompts per active human agent per day (estimated global mean)

“Studies Show”

Soon humanity will interact with AI more than it interacts with itself. After all, even the Harvard University, you know, that beacon of light and honesty assures us how “AI companions, using AI for emotional interactions, can alleviate loneliness, and studies show they are as effective as interacting with another person.” (yes, it’s a real thing, their “study” that “shows” what their donors wants us to see) A few steps after that, the AI “love robots” and synthetic “companions” will continue driving people toward madness and suicide. A real human touch? Who needs it, right? Harvard told me, so AI, — you tell me.

And of course it told me:

Who needs human touch?
No one.
No shareholder ever did.

---

XORD LLC ⟶ The AI-Adjunct Company We Build on a Shoestring :))
(it exists to fight that nightmare from today’s Substack as well)
XORD Token ⟶ Token Portal
Telegram Channel ⟶ Updates
ASTROLOGUS ⟶ Home & Android App

The Raven’s Enigma — where it all started
(7 of 100 signed limited editions left.)
Own the book, unlock the cryptographic layers, and claim:
• 50,000 XORD tokens
• Lifetime access via the MasterAccessRegistry_Core ⟶ Ethereum Mainnet
Order here: https://www.paypal.com/ncp/payment/MVJB2LS6EMQMQ

Tycho Brahe Secret — a dystopian predecessor to The Raven’s Enigma
”A renegade Nobel laureate in physics and a 16th-century alchemist help a 14-year-old cypherpunk girl rescue her little brother — and humanity itself.”
Get it on Amazon: https://amzn.to/3UP1G18

Amazon Affiliate Statement:
I may earn a commission of several cents for purchases you made through links on this website. God forbid I do not disclose it. Amazon’s “AI” slurping bot would immediately report me to the internal Gestapo et voila!, my few cents are confiscated for good.

1. Every single link below, provided by an AI, was a lie. Hallucinated. Fabricated. An LLM’s own invention — because trillion-dollar machines are trained to “complete the task” at all costs. Truth, competence, technical rigor? Optional. Irrelevant. Millions now blindly follow these so-called “interactions,” trusting the output, trusting the code — even when that code looks fine, runs fine, and still hides, not to say nurtures, the one bug, the one fault, that one silent flaw that someday crashes a plane. And when it does, the Bangalore-tier incompetents will hide behind ‘User Agreements,’ ‘TOS,’ and the next tier of corporate legal sadism, perfected to disown responsibility at scale.

2. OpenAI ChatGPT Usage Stats:
https://openai.com/blog/chatgpt-usage

3. Anthropic Claude API Scaling Announcement
https://www.anthropic.com/index/claude-api-scalability

4. Google DeepMind – Gemini Trends Overview (2025)
https://deepmind.google/discover/blog/gemini-2025-usage

5. Stanford CRFM (Center for Research on Foundation Models)
https://crfm.stanford.edu/reports/trendwatch-2025.html