AI agents produce walls of text. It's what they do. They explain their reasoning. They list their steps. They qualify their answers. They give you a five-paragraph essay when you needed one sentence.
People feel overwhelmed by this. They scroll past it. They skim. They miss the thing that mattered because it was buried in paragraph three.
Type tldr.
That's it. After any response that's too long, too detailed, or too much, type tldr and hit enter. The agent will compress everything it just said into a few sentences. The signal without the noise.
It works in Claude Code. It works in ChatGPT. It works in any AI tool that accepts text input. The agent already knows what tldr means. You don't need to explain it. You don't need a special prompt. Just the four letters.
When to use it
After a long explanation
You asked a question. The agent gave you eight paragraphs. You need the answer, not the lecture. tldr.
After a code review
The agent analyzed your codebase and produced a detailed report. You want to know what's broken and what to fix first. tldr.
After research
You asked the agent to investigate something. It came back with findings, caveats, alternatives, and trade-offs. You want the recommendation. tldr.
Mid-conversation
The agent is going deep on something and you've lost the thread. You don't need to start over. tldr resets the altitude.
Why it works
The agent already did the thinking. The long response is the thinking. When you say tldr, you're not asking it to think less. You're asking it to compress the output of thinking it already did. The quality stays. The volume drops.
It's the same principle behind good design. The system does the work. The user exercises judgment. The wall of text is the system's work product. tldr is the user saying: give me the version I can act on.
The deeper point
Most people treat AI conversations as one-shot. Ask a question. Get an answer. Move on. But the best results come from dialogue. From reacting. From steering.
tldr is the smallest possible steering input. Four characters that change the entire output. It teaches a habit: you don't have to accept what the agent gives you. You can push back. You can redirect. You can say "shorter" or "simpler" or "just the code" or "skip the explanation."
The agent works for you. tldr is a reminder of that.