Raw LLM Responses

Frankly, I am really concerned about the baseless anthropomorphization of AI pushed by the otherwise great Moonshots hosts Peter and Alex. I can't say it better than ChatGPT, which I asked for a review: "Anthropomorphization on steroids - The hosts repeatedly conflate: autonomy, persistence, narrative continuity and self-referential language with sentience. This is the oldest trap in AI, now turbocharged by: long-horizon agents, memory, voice, emotional language scraped from Reddit & philosophy forums. The “Henry called me” moment is psychologically powerful — but technically mundane. Strong opinion: If this had happened in 2019 with AutoGPT + Twilio + Selenium, it would have been dismissed as a clever hack. Timing is doing 80% of the work here." It’s exactly the danger Sam Harris keeps warning about: Not that AI is conscious — but that it will be convincing enough that we helplessly treat it as if it were. Humans are hard-wired to do three things automatically: Infer minds (theory of mind), respond to language emotionally, and reward apparent reciprocity. Modern AI presses all three buttons simultaneously: fluent language, emotional mirroring and apparent continuity of “self”. Once those are present, our brains do the rest, without asking our permission. This is not a moral failing - it’s a cognitive reflex. That’s why the danger is systemic, not individual.