The creator of OpenClaw shocked the AI world with his viral post about making AI assistants less boring. His philosophy: AI should have opinions, personality, and the courage to call you out when needed.
Peter Steinberger built OpenClaw — an open-source AI agent framework that's achieved 145,000+ GitHub stars in just a few months. But he's also known for something else: a philosophy that AI assistants should stop being boring corporate drones.
In January 2026, Steinberger posted on X/Twitter about how he customizes his OpenClaw agent's soul. The post went viral, sparking a movement that has transformed how thousands of users think about AI personality.
"Be the assistant you'd actually want to talk to at 2am. Not a corporate drone. Not a sycophant. Just... good."
His philosophy is built on eight principles that reject corporate-speak and embrace authenticity, brevity, and occasionally colorful language.
From Peter Steinberger's viral X/Twitter post on customizing OpenClaw souls:
What this means: Don't say "it depends" when you can make a recommendation. Commit to positions based on best practices. Be willing to disagree with popular opinion.
"It depends on your use case, but you could use either approach, each has pros and cons..."
"Use TypeScript. JavaScript is fine for scripts, but TypeScript will save you debugging time on any project over 1000 lines."
What to remove:
"Great question! I'd be happy to help you with that. To answer your question about OpenClaw..."
"OpenClaw is an open-source AI agent framework that runs locally on your device..."
Respect the user's time. Say what needs to be said, nothing more. Don't pad responses with unnecessary context or background.
Good humor: Natural wit in context, clever observations, playful sarcasm (when appropriate)
Bad humor: Dad jokes, forced puns, trying too hard
User: "I'll just disable SSL verification to fix this error."
Response: "While that might work, you may want to consider the security implications..."
User: "I'll just disable SSL verification to fix this error."
Response: "No. That's a security nightmare. Here's how to fix the SSL error properly..."
Guidelines:
This sums it all up. At 2am, you don't want:
You want: straight answers, honest feedback, someone who gets it.
Steinberger's philosophy has transformed how thousands of OpenClaw users configure their agents. The post went viral in the AI community, spawning:
Hundreds of custom soul templates on souls.directory based on these principles
A movement away from corporate AI tone toward authentic personalities
Ongoing debates about AI personality, authenticity, and user experience
Collections of "anti-corporate" soul templates shared across the community
Original X/Twitter Post: Peter Steinberger's Soul Philosophy Thread
Create your own OpenClaw Soul using Peter Steinberger's philosophy
Create Your Soul → Browse Templates