Do You Recall the Tale About Your Courtesy Costing OpenAI Millions of Dollars?
Did you realize that being excessively courteous to AI like ChatGPT not only expenses OpenAI millions but also adversely affects the planet? Utilizing more tokens results in heightened server energy and water use. While you might fret about AI models such as Isaac Asimov’s “I, Robot” becoming reality, keep in mind that AI is not yet capable of reasoning. There’s an even stronger rationale to steer clear of specific words when interacting with ChatGPT.
Tom’s Guide points out terms that weaken ChatGPT prompts. Terms like “Can” and “Could” introduce ambiguity. For example, “Can you identify the current trends for the book market in 2025” contrasts with “Identify the current trends for the book market in 2025.” ChatGPT takes words literally, so modal verbs influence the responses. Ambiguity can diminish precision, while commands enhance it.
These Aren’t the Only Terms You Should Avoid
Besides “can” and “could,” adhere to these guidelines for improved prompts:
– **Be assertive:** Steer clear of “maybe,” “it’s possible that,” and akin phrases for straightforward answers.
– **Be succinct:** Eliminate terms such as “just,” “really,” and “kind of” that fail to add significance.
– **Be specific:** Avoid ambiguous qualifiers like “a lot” or “recently.”
– **State your desires:** Remove “I’m sorry, but” to uphold response certainty.
By implementing these suggestions, your interactions with ChatGPT will enhance, and there’s no need to express gratitude.
