OpenClaw creator: "If you prompt it, it will"

Headline

OpenClaw creator sums up prompt-driven AI in four words

Summary

Peter Steinberger, founder of the open-source AI assistant OpenClaw, tweeted “If you prompt it, it will” in response to a thread involving OpenAI and related accounts. The reply points to how OpenClaw works: you tell it what to do through WhatsApp or Telegram, and it handles tasks like email and calendar management. Steinberger built the project to make AI agents accessible without complex setup. The tweet matters because it shows where personal AI assistants are heading, moving from chatbots that just answer questions to agents that actually do things.

Analysis

Steinberger has an unusual position right now. He’s an iOS and web developer who recently joined OpenAI, but he still runs OpenClaw as an independent open-source project. That dual role shapes how he thinks about AI agents.

The phrase “If you prompt it, it will” gets at something real about how these systems work. They’re not just generating text anymore. They’re taking actions, and the gap between “ask” and “do” keeps shrinking.

OpenClaw specifically focuses on security (it’s built to resist prompt injection) and runs locally on your device. That’s a different bet than the big labs are making with cloud-based assistants. Whether open-source alternatives like this can compete with proprietary systems is still an open question, but they offer something the big players don’t: you own it, you can modify it, and your data stays on your hardware.

Steinberger straddling both worlds (OpenAI employee and open-source maintainer) could get interesting. Either approach might pull ahead, or we might end up with some hybrid where people mix and match.

Impact Assessment

• Significance: Medium
• Categories: Technical Insight, Developer Tools, Open Source