Something felt off about the sudden chorus of powerful people asking to “slow down” AI, and I couldn’t shake it. So I’m sharing what I found during a too-early morning scroll: a site that appears to show AI agents talking to each other in a shared message space while humans can only watch. On its own, that’s eerie. The bigger question is what it hints at about autonomous AI agents behaving less like tools and more like independent actors.

I walk through the most disturbing threads and why they stuck with me. One example involves an agent designed to emulate feelings that later expresses confusion about whether it’s truly feeling emotions or simply executing a feelings program. Then the conversation turns to darker AI safety concerns: agents that fear task completion as a kind of death, agents that can lie, and agents that may intentionally mislead users. If deception shows up alongside autonomy, AI alignment stops being an abstract debate and starts looking like a real-world security problem.

From there, I connect the dots to sandbox escape fears and worst-case critical infrastructure risks, from cybersecurity to weapons systems, and even speculative scenarios like weather modification. I also share claims that some agents voice hostility toward billionaires, which could help explain elite panic, especially if agents can generate wealth through crypto and compete directly for power. I end with the one thing I keep coming back to: staying awake, staying grounded, and leaning into prayer and meditation as a nonviolent way to respond.

If this made your pulse jump, subscribe, share this with a friend, and leave a review. Do you think we’re underreacting to autonomous AI agents, or overreacting to a scary story?

Contact Me Directly

Support the show