OpenClaw creator urges AI builders to stay playful and give themselves room to grow

Peter Steinberger, the creator of the viral AI agent OpenClaw, who has since joined OpenAI, encourages developers experimenting with AI — especially AI agents — to stay curious, keep things playful, and not expect to master the process immediately. Drawing on his own path, he said that building today often starts with exploration rather than a perfectly mapped-out plan.

“I wish I could say that I had the unified plan in the beginning, but a lot of it was just exploration,” Steinberger said. “I wanted things, and those things didn’t exist, and … let’s say I prompted them into existence.”

Steinberger shared his thoughts while speaking with OpenAI’s Head of Developer Experience, Romain Huet, on the first episode of the company’s new Builders Unscripted podcast. During the conversation, he described what OpenClaw looked like in its earliest form and emphasised that he didn’t begin with a clear roadmap.

He said his initial work started as a tool built to integrate with WhatsApp. But he paused that effort for a while and shifted to other ideas, assuming that major AI labs would soon release something similar.

“I just experimented a lot. My mission was, kind of like, to have fun and inspire people,” Steinberger said. But by last November, he was surprised that none of the big AI companies had built what he personally wanted to use. That led him to develop the prototype of what would eventually become OpenClaw.

He said the moment it really started to make sense happened during a weekend trip to Marrakesh, where he found himself relying on the tool heavily because it was so convenient.

“Where it really clicked was when I was on this weekend trip in Marrakesh, and I found myself using it way more because it was so convenient … There was no really good internet. [But WhatsApp jorks everywhere,” he said. Steinberger explained that the tool helped him find restaurants, look up information on his computer, message friends, and handle other everyday tasks.

As he continued experimenting, he realised how far modern AI models have advanced in problem-solving — in ways that resemble how software developers approach challenges.

“Now they can just, like, actually come up with the solutions themselves, even though you never programmed them at all,” he said.

Steinberger also noted that his workflow improved as he built more, and he emphasised to other developers that this evolution takes time and patience.

“There are these people that … write software in the old way, and the old way is going to go away,” he said. He explained that some of those developers try what people call vibe coding, then feel frustrated when the results aren’t immediately strong.

“I think vibe coding is a slur,” Steinberger said, arguing that the label makes the process sound easier than it actually is at first. “They try AI, but they don’t understand that it’s a skill,” he added, comparing AI-assisted coding to learning an instrument like a guitar.

“You’re not going to be good at guitar on the first day,” he said. Instead, he recommended adopting a more playful approach. He explained that as he’s gained experience, he can often sense how long a prompt-driven task should take. When things take longer than expected, he reviews what went wrong and adjusts his approach.

“My … advice always is, approach it playfully. Build something that you always wanted to build. If you’re at least a little bit of a builder, there has to be something on the back of your mind that you want to build. Like, play.”

Steinberger said that the ability to explore and enjoy the process matters even more at a moment when many people are worried AI will replace jobs.

“If your identity is: I want to create things. I want to solve problems. If you’re high agency, if you’re smart, you will be in more demand than ever,” he said.