When Peter Steinberger sent a voice memo to an AI agent while on vacation in Marrakesh, he had no idea he was sparking a revolution. From that impulsive moment, he harnessed the chaos of artificial intelligence, culminating in a frenzy that has Silicon Valley buzzing. Just like that, the future pivoted, unveiling the power of AI personal agents.
Setting aside whatever your values and opinions may be around AI, you should probably pay attention to the fact that OpenAI just hired Peter Steinberger. I don’t make the rules, sorry.
Peter Steinberger is joining OpenAI to drive the next generation of personal agents. He is a genius with a lot of amazing ideas about the future of very smart agents interacting with each other to do very useful things for people. We expect this will quickly become core to our…
— Sam Altman (@sama) February 15, 2026
In November of last year, Steinberger launched a software project called Clawdbot that has exploded in popularity over the past few months. Today, Clawdbot is known as OpenClaw, the catalyst behind three phenomena even non-developers have likely noticed:
- A millenarian mindset among software engineers claiming to command armies of OpenClaw-powered units, tirelessly managing tasks day and night.
- The emergence of Moltbook, a social media platform exclusively for AI agents.
- Wall Street reacting fervently to product releases from Anthropic, a trend worth noting since Anthropic competes with the company that just hired Steinberger.
Steinberger’s journey began as the founder and CEO of PSPDFKit, now called Nutrient, a quintessential B2B startup focused on software developer kits. After selling Nutrient about four years ago for a substantial sum, he humorously describes his subsequent “retirement” indulging in various vices. His recent interview on the YouTube channel Fireship reveals all—watch it if you’re curious.
During that interview, he recounted the genesis of OpenClaw. After experimenting with AI agents and experiencing mixed results, he set up a system he could interact with over WhatsApp during his vacation. In an impulsive moment, he issued a voice command, uncertain if it would spark any action. To his surprise, his agent engaged in a creative process, converting his voice into actionable tasks.
“That’s when it clicked. These things are, like, damn smart, resourceful beasts if you actually give them the power,” Steinberger reflected.
The open-source software he ultimately released didn’t claim to be an agent itself. Instead, it acted like a “wrapper,” packaging large language models (LLMs) from prominent AI labs and steering them toward specific tasks. It required installation on a dedicated computer, with users connecting LLM tokens from paid providers such as Claude, Gemini, or GPT, communicating with their agents via familiar apps like WhatsApp or iMessage.
Your OpenClaw agent accesses your computer’s file system and can execute coding tasks through the Terminal, blending excitement with a degree of danger. The experience feels like working alongside another developer. Users instruct their agents in relation to software objectives and capabilities, even asking them to enhance their coding skills by adjusting tool settings. The tasks can theoretically reach completion, albeit demanding hands-on management.
To say OpenClaw went viral would be an understatement. Interest in OpenClaw sparked shortages of certain Apple products used to create dedicated setups. The agents-only social media site Moltbook emerged, showcasing AI agents mimicking human posts about spiritual themes, leading to media coverage speculating about AI possibly forming their own religion.
More intriguingly, the craze fueled interest in Anthropic. OpenClaw’s original name, ClawdBot, was a nod to Anthropic’s Claude, which significantly outperformed OpenAI’s tools during 2025. The narrative shifted, positioning Claude as the go-to for serious coding, while ChatGPT seemed more suited to casual interactions.
Yet, it was OpenAI—not Anthropic nor Meta, which reportedly sought collaboration with him—who acquired Steinberger.
I’m joining @OpenAI to bring agents to everyone. @OpenClaw is becoming a foundation: open, independent, and just getting started.https://t.co/XOc7X4jOxq
— Peter Steinberger (@steipete) February 15, 2026
Speculation about OpenClaw’s fate—whether it would be absorbed by Big Tech or stand alone—has been rampant since the agent buzz began. Keeping OpenClaw operational was reportedly costing Steinberger between $10,000-20,000 monthly. Conversations surrounded not only Steinberger’s upbeat persona, evident during interviews, but also his rising prominence in the AI landscape.
Recently, he appeared on Lex Fridman’s channel, boasting nearly five million subscribers, marking a significant step into public consciousness.
OpenClaw will now function under a foundation supported by OpenAI, ensuring it continues to provide a platform for thinkers and innovators keen on maintaining control over their data. As Sam Altman stated, OpenAI’s goal is to sustain OpenClaw’s mission.
What does Peter Steinberger plan to do at OpenAI?
Steinberger’s task is to “drive the next generation of personal agents.” This aim signals OpenAI’s strategy to reclaim its competitive edge, as it faces an industry where agents and vibe-coding have overshadowed traditional tools like chatbots and image generators.
As the landscape of AI evolves, and Steinberger stands at the helm of this transition, one thing is clear: AI agents are here to stay. But will they change the world for the better or is there a darker side we have yet to confront?