What are the 5 prompting techniques worth learning (and the ones that are outdated)?

The techniques that still pay off: give the AI the context a smart new colleague would need, show it examples of good output (Anthropic suggests 3 to 5), break big tasks into steps, say exactly what format and style you want, and iterate. The scaffolding has changed. Current models reason on their own, so Anthropic finds a general 'think thoroughly' often beats a hand-written step-by-step plan, and it advises softening all-caps 'CRITICAL: you MUST' language. OpenAI says GPT-6 follows long instructions well but is sensitive to unclear or conflicting ones, so cut contradictions. A one-sentence role still helps set tone.