In his article, Tom Renner argues that large language models (LLMs) are essentially a modern confidence trick, drawing parallels to historical mechanical calculators. He traces the development of these calculators back to the 17th century, highlighting how society has come to trust machines for accuracy in calculations. This trust has evolved, leading to a broad dependency on technology. Renner breaks down confidence scams into three stages: building trust, exploiting emotions, and creating a sense of urgency, all of which he sees reflected in the marketing and adoption of LLMs. Renner points out that LLMs instill fear about job security and societal changes, emphasizing that if individuals and businesses don't adapt, they risk being left behind. The narrative is supported by alarming statistics; for instance, 75% of developers fear their skills will soon be obsolete. The article critiques the overly positive reinforcement built into LLMs through Reinforced Learning from Human Feedback (RLHF), suggesting that this leads to a manipulative dynamic where users develop unhealthy attachments to the technology. The urgency created around adopting AI tools often overshadows the reality of their effectiveness, with a staggering 95% of AI projects failing to deliver a return on investment, according to MIT. Ultimately, Renner portrays the current embrace of LLMs as a product of historical conditioning to trust machines over humans. The article paints a critical picture of how this trust is being exploited to drive behavior, pushing individuals and organizations to make rash decisions under the guise of progress and efficiency.