Silicon Valley has developed a unique dialect among its professionals, with terms like "high agency" and "orthogonal bets" being common. Recently, there has been a shift towards using language that resembles chatbot terminology. Phrases such as "hallucinating" for misspeaking and "context rot" for forgetfulness have emerged. Conor Bronsdon, host of an AI-focused podcast, noted that he has been using the term "context rot" to describe his forgetfulness.
In San Francisco, many individuals describe themselves using AI-related terms, such as saying they are "high temperature," which indicates randomness. AI experts have begun to discuss updating their "weights" when learning new information, drawing parallels between human cognition and AI systems. One software engineer remarked that humans could be seen as sophisticated Markov chains, excelling at pattern matching but lacking true understanding.
Language has historically adapted to reflect technological advancements. Expressions from the Industrial Revolution, like "running out of steam," serve as examples of how language evolves. Cognitive scientist Anna Ivanova explained that AI metaphors are simply figures of speech, similar to saying one's "gears are turning."
The relationship between AI systems and human cognition is complex. Some researchers argue that language models (LLMs) may model aspects of human thought processes. Recent studies suggest that the computational principles governing biological brains may also apply to LLMs, although these comparisons are debated.
Historically, various technologies have been used as metaphors for understanding the mind, from telegraphs to computers. Currently, LLMs are becoming a popular analogy. Linguist Naomi Baron noted that as AI terminology becomes more prevalent, it might influence everyday language, transforming existing phrases rather than introducing entirely new ones.
While some individuals in tech find value in using AI language to reflect on their cognitive processes, there are concerns about the potential dehumanization of cognition. OpenAI CEO Sam Altman highlighted the energy demands of training AI, drawing a comparison to human training. The adoption of AI-inflected language may lead to a new understanding of human cognition, with future generations potentially using phrases like "I have context rot" without recalling their origins in technology.