[DigitalToday reporter Yoonseo Lee] It has been introduced that adding short directives such as “explain like I’m 5” or “too long; didn’t read” to ChatGPT can noticeably change the way it responds. On April 1 local time, IT media outlet TechRadar reported that so-called “codes” shared online are used to adjust ChatGPT’s tone and flow and reduce unnecessary wording.

“Explain like I’m 5” is a request to break down an unfamiliar concept to a child’s level in simple terms. For example, if the directive is added before a question such as, “In a homeowners insurance document, what is the difference between actual cash value and replacement cost?”, it summarises the answer in plain language rather than legal phrasing.

“Too long; didn’t read” was presented as a useful directive for summarising long writing. It said that pasting long text after the phrase produces a summary without additional explanation. It can be used to condense cooking recipes with lengthy introductions, subscription cancellation policies, or extended group chat messages. It also described that when the directive was entered for an event notice, it compressed key points into 5 items and summarised them to the level of, “Arrive a bit early and pull up your ticket easily on your phone.”

“Jargon” is used to convert sentences into excessively complex office-style wording. By contrast, “humanise” is used to polish sentences that may sound stiff or aggressive into more natural phrasing. It works by helping avoid trite expressions or excessive formality in sensitive writing such as condolence messages.

Physicist Richard Feynman is also well known for the “Feynman technique,” which explains complex concepts simply and checks understanding. The idea that “if you can’t explain something simply, you don’t understand it properly” is also often mentioned with the technique.

Users can type “Teach this using the Feynman technique” and add the topic they want. ChatGPT then explains the concept simply, points out gaps in understanding and explains it again more easily. It said it can also quiz the user or ask them to explain incorrect parts again, and introduced it as useful for learning topics that require repeated checks such as music theory or economics.

Such directives are drawing attention because they can quickly adjust ChatGPT’s response format, difficulty and style without writing a long, complex prompt. But results can vary depending on the context of the question and the input information even with the same phrase, so using short directives together with specific questions may be a more effective way to get the answer you want.