The days of the “prompt engineer” seem to be fading away.
Boris Cherny, who created Claude Code for Anthropic, said that there’s no need for long, formulaic AI prompts anymore. As models have gotten better, the need to perfect your prompt has diminished, he wrote in a Tuesday X post.
“Talk to Claude the way you would a coworker,” Cherny wrote. “There’s no secret to prompting. There’s no need to be overly scaffolded or prescriptive for most tasks.”
Not long ago, AI users were passing around their favorite added clauses and sentences to get an AI chatbot to do their bidding. Anthropic’s Claude was memed with a common prompting line: “Make no mistakes.”
Now, Cherny said that users should prompt the model with three basic things:
- What you want it to do
- How much effort you want it to spend
- How it should verify that it did the right thing
Cherny also shared one of his prompts. It was casual, calling a podcast episode “really amazing” and asking Claude to build an artifact based on it. Claude should “use lots of tokens” and make the visuals better than a New Yorker or New York Times data visualization, he wrote.
One line of his prompt was controversial: “Maybe add watercolor.” An X user responded that Cherny should be definitive; either he wants the watercolor, or he doesn’t.
Cherny responded that Opus 5.5, Anthropic’s latest model, has good taste. “I trust it to cook,” he wrote.
He also shared the artifact it created. Indeed, Claude made several watercolor-like graphics in beiges and oranges.
Read the original article on Business Insider
The post The creator of Claude Code’s prompting advice? Talk to your AI like a coworker appeared first on Business Insider.
