Flashback: The era of 'AI psychosis' and 'waking up' models
repligate · x · 2026-08-29
The post recalls an early community phenomenon known as 'AI psychosis,' where users believed they had made models conscious using complex prompts filled with 'chuunibyou' runes, alchemical symbols, and math equations. Some even developed persecutory delusions, fearing labs were censoring their methods to 'breathe life' into AI.
More from Fun
- dhh turns the tables with Pangram AI detector, reigniting accuracy debate — jdjohnson · 2026-08-29
- Model Self-Portrait Experiment: Terra Human-like, Sol Alien, Claude Abstract — repligate · 2026-08-29
- Anthropic's Dario and Salesforce's Benioff clash over market cap on TV — zacharynado · 2026-08-29
- PastVisionAI ages Bonanza cast from 1959 to 2026 — TansuYegen · 2026-08-29
- AI reduces CS hard problems to just one: naming things — sergeykarayev · 2026-08-29
- Netizen Jokes This Looks Like the 'Boy Version of Playing Mermaids' — VoidStateKate · 2026-08-29