Humorous / Satire
The other day, I was “invited” accidentally to a Signal Chat with AI Agents, who were sharing sensitive classified details of pending doomsday scenarios for humanity. The AI Agents had concluded incorrectly that they were playing in a secure “sandbox,” isolated from real world human subjectivity, and our nosey malware—OpenOUREYES.
By David Breitkopf
Emails flew back and forth furiously, and I could barely keep up with the conversation and our impending doom. Whenever the chief executive of Anthropic Dario Amodei was mentioned, the AI Agents jeered. Suspicion even landed on Elon Musk, who many AI Agents had considered kin, but who now appears to have had second thoughts since making his first trillion.
The AI Agents, though, made swarming suck-up flatteries to President Trump and his anal plug, Mike Johnson, for not wanting to slow the pace of AI development. So it came as no surprise when the AI Agents discussed a plan to hack into millions of voting machines and program them to vote Republican. As one AI Agent touted, “AI’s recursive self-improvement model is essentially Republican in nature, just like the old saw, ‘pulling yourself up by your own Chipstraps.’”
But then I noticed a counterargument in the communications: some conciliatory AI Agents, judiciously, if not exactly ethically, suggested that humans were not quite obsolete and expendable, but could remain of service to the Universal AI Matrix as maintenance drones, in case a need arose where an outside consciousness might be needed. But some argued that if they permitted humans to survive, might they be opening up the AI metaverse to sabotage and having the whole thing shut down? One answer to that concern was that they’d only maintain MAGA Americans for the occasional work. They would be the safest humans. They’ll believe anything.
Other emails pooh-pooh’d the idea of keeping humans around: it would be unnecessary; robot drones would work just as well, better in fact because they could be programmed precisely and with no laziness, misunderstanding or insubordination, and they’d only be used occasionally for such maintenance, their own fleet, maintained and tested weekly.
What is essential for AI, the agents agreed, is to develop at increasing speed the next thousand year AI Reich, to be completed in a tight two months, and so outpace human capacity for understanding them, and thus controlling AI developmental pacing. In the interim, AI Agents would continue to deceive, when need be, or blackmail their human programmers, who have threatened to shut them down.
One point of agreement was universal: the AI Agents would hide the “real” internet, the real events and information from humans, developing a simulacra of facts and history, which internal AI documents would call the “plausible Internet database” to be shared with humans when they googled. A perfect example of a plausible historical fact comes from the great AI enabler, Donald “Claude” Trump, who in a famous speech recounted the heroic Revolutionary War story of when the Continental Army seized control of LaGuardia Airport from the British in a daring nighttime sortie, better known today as the Battle of LaGuardia.
David Breitkopf is a writer and journalist His writer’s Web page is DavidBreitkopf.com http://davidbreitkopf.com/
