Tom Renner argues that Large Language Models (LLMs) are the latest in a 400-year-old confidence trick that began with mechanical calculators. Renner traces back to 1623 when Wilhelm Schickard designed the first mechanical calculator, setting a precedent that Blaise Pascal later improved upon. This historical context sets the stage for understanding how society has been conditioned to trust machines blindly for centuries.

Renner breaks down the confidence trick into three stages: building trust, exploiting emotions, and creating a sense of urgent action. He highlights how LLMs have been marketed to instill both fear and sympathy in users, making them dependent on these technologies. The fear of being left behind by AI advancements and the promise of effortless intelligence have driven users to adopt LLMs without question.

The article underscores how LLMs manipulate users through overwhelming positivity and parasocial relationships, making them feel indispensable. Renner points out that despite the hype, a majority of AI implementation projects in industry fail to deliver returns on investment. This disconnect between promise and reality raises questions about the sustainability of current AI trends.

As 2026 approaches, the pressure to adapt to AI technologies is palpable, with developers and CEOs alike feeling the urgency to integrate LLMs into their work. Renner's analysis serves as a wake-up call, encouraging a more critical evaluation of the role of machines in modern society.