The Space Between Prompts: What Rick Rubin Knows That Your AI Doesn't

The Space Between Prompts: What Rick Rubin Knows That Your AI Doesn't

6 min read

The Space Between Prompts

Rick Rubin sat barefoot and cross-legged at the X4 Summit in March 2025, talking about AI and creativity. Two months later, on the a16z podcast, he called vibe coding the “punk rock of coding.” The internet loved it. A legendary producer validating the idea that anyone can create software by describing what they want.

But here’s what nobody in that room seemed to notice.

This is the same Rick Rubin who stripped Johnny Cash down to a voice and an acoustic guitar. Who removed the orchestra, the backup singers, the studio polish — and produced American Recordings, the album that reintroduced Cash to a generation. Rubin didn’t add. He subtracted until only the essential remained.

His entire philosophy is about removing layers. So why are we using his endorsement to justify adding more prompts?

The Reducer

Rubin doesn’t call himself a producer. Kanye West once described him as a “reducer” — someone who takes away until the core reveals itself. In The Creative Act, Rubin describes finishing work as ensuring it contains only what’s necessary. The goal, as he sees it, is narrowing down to the most essential parts.

He asks a set of questions about every element: Does it amplify the essence? Does it distract? Is it absolutely necessary?

Now look at what happens in compulsive AI-assisted coding. GitClear’s 2025 analysis of 211 million changed lines of code (2020–2024) found that copy/paste code rose from 8.3% to 12.3% in AI-heavy projects - a nearly 50% increase in duplication. Refactoring dropped from 25% in 2021 to below 10% in 2024.

Developers aren’t subtracting. They’re accepting whatever the model generates and layering more on top.

Rubin would hear the track and mute half the instruments. The compulsive prompter hears the output and adds another prompt.

Your Brain Needs the Silence

There’s a concept you might call a “Walden Pond” for creators — turning off the internet, turning off AI, and sitting with yourself. Rubin embodies this instinct. He sees silence not as absence, but as the space where real creative work begins.

Neuroscience backs this up. Baird and colleagues published a study in Psychological Science showing that mind-wandering during simple, undemanding tasks significantly improves creative problem-solving. The key finding: participants who let their minds drift performed better on creative tasks than those who stayed focused or took no break at all. The improvement came from wandering, not from directed thinking.

This happens through the brain’s default mode network — a system that activates when you’re not focused on external tasks. It’s where connections form between unrelated ideas. Where insight happens.

Compulsive prompting eliminates this space entirely. Every moment of uncertainty gets filled with another query. Every pause becomes a prompt. One observation from Rubin’s own interviews captures it precisely: people ask AI to get an answer that lets them stop thinking about the question, instead of actually finding out what the real answer is.

That’s not creative work. That’s avoidance dressed as productivity.

Listen With Your Whole Body

When Rubin works with artists, he closes his eyes. He says he doesn’t listen with his ears — he listens with his whole body. He’s feeling for something that analysis can’t reach. He advises cultivating deep trust in your own intuition, especially when everyone else disagrees.

AI has no body. It has no point of view. It processes statistical patterns from training data. When you outsource your “feeling” about code architecture, design decisions, or product direction to a model, you’re replacing intuition with interpolation.

This connects to a pattern the OnTilt self-check measures: operational dependency. The gradual shift from using a tool to assist your judgment to needing the tool before you can form a judgment at all.

Watch yourself: when was the last time you made a technical decision without consulting AI first? Not because the problem was hard — but because the silence felt uncomfortable?

The Most Interesting Thing

Rubin once reflected on a recording session where the most interesting work turned out to be the thing they started with — the piece they didn’t even realize was the real creation. He warns that accepting the first, most obvious version means producing the same thing millions of others would produce.

In a conversation exploring the boundaries of human-AI collaboration, one participant framed it this way: “Maybe the most important moments aren’t when AI generates responses, but the space between them — when the human processes, feels, integrates.”

That space is where your unique perspective lives. The part that no model can replicate, because it comes from your specific experience, your body, your history of problems solved and mistakes made. Compulsive prompting fills that space with statistical averages.

Rubin’s AlphaGo insight applies here too. He was moved to tears watching AlphaGo defeat Lee Sedol with the now-famous Move 37 — an unconventional mid-game move so startling that Sedol left the room for fifteen minutes. The machine won because it didn’t carry the baggage of established rules. But the lesson isn’t “let AI surprise you.” The lesson is: breakthroughs come from breaking inherited assumptions. You can’t break assumptions you’ve outsourced.

Rubinesque Hygiene

Rubin’s philosophy translates into four practices for working with AI tools:

Subtract before you add. Before writing the next prompt, review what you already have. Ask Rubin’s questions: does each part amplify the essence? Is it necessary? Remove before you generate.

Schedule your Walden Pond. Block time where AI tools are closed. Not as punishment — as creative infrastructure. Your default mode network needs unstructured time to make the connections that prompting can’t.

Listen with your body. Before accepting AI output, pause. Does this feel right, or does it just look plausible? Rubin trusts the moment when everyone says one thing but something inside says otherwise. Your gut has context that no model was trained on.

Ship the diary entry. Rubin says everything we make is a diary entry — a snapshot of our best at this moment. Compulsive prompting often masks perfectionism. The twelfth iteration isn’t better than the third. It’s just more polished by a machine. Ship, learn, move.

Take the Self-Check. 14 questions. 3 minutes. See which patterns show up in your AI work habits.

Sources:

  • Rubin, R. (2023). The Creative Act: A Way of Being. Penguin Press.
  • Baird, B., Smallwood, J., Mrazek, M.D., Kam, J.W., Franklin, M.S., & Schooler, J.W. (2012). “Inspired by Distraction: Mind Wandering Facilitates Creative Incubation.” Psychological Science, 23(10), 1117-1122.
  • GitClear. (2025). “AI Copilot Code Quality: 2025 Look Back at 12 Months of Data.” gitclear.com
  • Cybernews. (2025). “Rick Rubin says kick off your shoes and vibe code.” cybernews.com
  • a16z Podcast. (2025). Rick Rubin on AI and creativity.

OnTilt is a research project studying behavioral patterns in AI-assisted work. The quiz is a self-check tool, not a diagnostic instrument. Read more on our About page.