Within online drug forums, some highly knowledgeable users—termed 'e-psychonauts'—provide reliable information on psychoactive substances. A netnographic study observed cyber drug communities in early 2014, analyzing posts with phenomenological methods. E-psychonauts typically described themselves as 'psychedelic researchers,' 'new Shamans,' 'philosophers,' or 'alchemists.' They were mainly young, male, unmarried, Caucasian, and in good or excellent employment. They possessed strong IT skills, verbal fluency, and deep chemical/pharmacological knowledge, often testing or synthesizing drugs to explore altered consciousness. The findings suggest a need to improve professional awareness of this group and develop targeted prevention approaches.
Hypnotized minds and large language models (LLMs) both produce sophisticated, context-appropriate behavior through automatic pattern-completion with limited executive oversight. This review identifies three shared principles: automaticity (responses arise from associative rather than deliberative processes), suppressed monitoring (leading to confabulation in hypnosis and hallucination in LLMs), and heightened contextual dependency (immediate cues override stable knowledge). Both systems produce coherent but ungrounded outputs requiring an external interpreter to supply meaning. They also exhibit functional agency—complex, goal-directed behavior—without subjective agency or conscious awareness. The review argues that hypnosis offers an experimental model for understanding how intention can dissociate from conscious deliberation, and suggests that reliable artificial intelligence may require hybrid architectures integrating generative fluency with executive monitoring.