"Just Ask Claude" Is a Rare and Valuable Skill

The piece argues that skillfully prompting large language models - summed up as “just ask Claude” - is a uncommon but highly useful ability because most people either won’t or can’t extract good results themselves. The writer admits feeling odd taking credit when AI does the work, but insists intermediaries provide real value: some people reject AI on principle (illustrated wryly by a dihydrogen monoxide gag), many don’t pay for premium models and therefore get weaker outputs, and a lot of people are simply unfamiliar with current tools and terminology. Those gaps make a competent prompter effectively a curator who saves others time and spares them technical hassle.

Concrete examples show why prompting matters: setting the model’s effort to “high,” asking an Opus 5.5 model to vet local subreddits and reviewer accounts to avoid astroturfing, or using knowledge of model capabilities to get better recommendations. There’s also a social dimension - people sometimes ask for help as an excuse to connect - so AI should augment, not replace, human gestures; the suggestion is to let Claude handle technical fixes but keep personal touches (like a handwritten card) human. A commenter reinforces that many users simply don’t know what they’re missing and that even free models can be powerful.