Peter Steinberger, the mind behind the widely popular AI agent OpenClaw and now a member of OpenAI’s team, offers guidance for anyone diving into AI development and agents. Reflecting on his journey, he believes the most effective approach today is to stay curious, embrace experimentation, and avoid putting pressure on yourself to master everything immediately. “I wish I could say that I had the unified plan in the beginning, but a lot of it was just exploration,” Steinberger shared. “I wanted things, and those things didn’t exist, and…let’s say, I prompted them into existence.”

During a conversation with Romain Huet, OpenAI’s Head of Developer Experience, on the debut episode of the Builders Unscripted podcast, Steinberger revisited OpenClaw’s origins and his initial lack of a fixed roadmap. He started by developing a tool integrated with WhatsApp, then paused to focus on other projects, assuming leading AI labs would soon release something similar. “I just experimented a lot. My mission was, kind of like, to have fun and inspire people,” he recalled.

By November of last year, however, he was surprised that no AI lab had built the tool he envisioned, prompting him to create the first prototype of what became OpenClaw. “Where it really clicked was where I was at 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 just works everywhere,” he explained. The agent helped him effortlessly locate restaurants, search for information on his computer, message friends, and handle various tasks.

As he continued experimenting, Steinberger recognized how advanced modern AI models have grown in problem-solving, resembling the way skilled coders operate. “Now they can just, like, actually come up with the solutions themselves, even though you never programmed them at all,” he observed.

Throughout his development process, Steinberger noticed his own workflow steadily improving—a point he emphasizes to fellow developers, noting that such progress takes patience and persistence. “There’s these people that…write software in the old way, and the old way is going to go away,” he remarked. When they attempt so-called “vibe coding” but are dissatisfied with the outcomes, he cautions against underestimating the learning curve. “I think vibe-coding is a slur,” Steinberger said, implying the term oversimplifies a more nuanced skill. “They try AI, but they don’t understand that it’s a skill,” he added, comparing AI-assisted coding to learning an instrument like guitar. “You’re not going to be good at guitar on the first day.”

Instead, he encourages a more playful, exploratory mindset. These days, when he crafts a prompt, he can intuitively gauge how long it should take; if it takes longer, he reviews what might have gone wrong and adjusts. “My… advice always is, approach it in a playful way. 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, just play.”

This spirit of experimentation and enjoyment is especially crucial now, Steinberger notes, when many fear AI might replace their roles. “If your identity is: I want to create things. I want to solve problems. If you’re a high agency, if you’re smart, you will be in more demand than ever,” he affirmed.