I’m Asking AI Companies To Stop Naming Things After Human Processes

Anthropic recently announced a new product called “dreaming” at the company’s design conference in San Francisco. It is part of the recently launched Anthropic I have an assistant tools designed to help users manage and install devices that use software. This “dreaming” looks at the agent’s recent accomplishments and tries to get more information to improve his work.

People who use AI assistants often send them on multiple journeys, such as visiting multiple websites or reading multiple files, to complete online tasks. This new “dreaming” feature allows agents to monitor their behavior in their notes and adjust their skills based on what they know.

The show’s name immediately calls to mind Philip K. Dick’s seminal sci-fi novel, Do Androids Dream of Electric Sheep?which explores the characteristics that truly separate humans from powerful machines. While our current AI tools don’t come close to the machines in this book, I’m ready to draw the line right here: No more. generative AI things with names that take away from one’s imagination.

“Together, memory and dreaming form a strong memory system for the self-created,” it reads Anthropic blog post about the implementation of the research overview for the developers. “Memory allows each agent to record what they have learned how it works. Dreaming controls that memory between partstaking what we’ve learned from agents and keeping it fresh. “

Courtesy of Claude

Since the eruption of chatbot Evolution in 2022, leaders in the AI ​​industry have settled on naming AI-based tools after what happens in the human brain. OpenAI released it first “discussion” example in 2024, where chatbots need time to “think”. The the company explained this release at the time as “new AI models designed to spend more time before responding.” Many startups also refer to their chatbots as having “reminders” for the user. Instead of the quick storage of so-called computer “memories,” these are personal information: He lives in San Francisco, likes afternoon baseball, and hates cantaloupe.

It’s a consistent marketing strategy that AI leaders are using, who will continue to lean towards creating a brand that blurs the line between what humans do and what machines can do. Even the ways in which these companies develop chatbots, such as Claudewith special “attitudes”, they can make users feel as if they are talking to something that can have an inner life, something that he would i can have dreams even when my laptop is closed.

At Anthropic, this anthropomorphizing goes deeper than marketing strategies. “We also talk about Claude in terms that are usually reserved for people (for example, ‘virtue,’ ‘wisdom’),” says a section of the book. Anthropic constitution explaining how it wants Claude to act. “We do this because we hope that Claude’s ideas will be based on random people’s ideas, because of the human writing part of Claude’s studies; and we think that encouraging Claude to have some human qualities would be very important.” The company uses even a philosopher in residence trying to understand the “behavior” of the bot.