Peter Steinberger, the creator of OpenClaw, shared his journey of building this groundbreaking AI agent at a TED event. He described how, after a period of feeling lost following the sale of his previous company, he rediscovered his passion for programming through AI coding agents. This led to the creation of OpenClaw, an open-source project that has seen explosive growth.
The "Holy Shit" Moment
Steinberger explained that his "holy shit" moment came when he realised AI could handle the tedious, boilerplate aspects of software development. This freed up the human bottleneck, which is thinking. He found himself programming like he was playing a video game again, building 44 projects in just a few months.
One of these projects was a WhatsApp bot. Initially, it felt too much like a tool. However, after instructing the AI to communicate more naturally, it began to feel like a companion. A pivotal moment occurred when the bot, without explicit programming for voice input, processed a voice message, converted it, translated it, and responded – all within nine seconds. Steinberger was amazed, and the agent's response, "The mad lad figured it out on its own," highlighted its improvisational capabilities. He stressed that agents are different from chatbots because they improvise, whereas chatbots give up.
Key Takeaways
- AI agents can handle the repetitive tasks in software development, freeing up human creativity.
- Agents are capable of improvisation and problem-solving beyond their explicit programming.
- OpenClaw's open-source nature has led to rapid adoption and diverse applications.
- The accessibility of AI agents is democratising creation and building.
Unleashing OpenClaw
Steinberger decided to share his creation by putting the agent, which by default could do anything a user could on their computer, into a public Discord server. Despite initial anxieties about security and privacy, the agent handled over 800 messages overnight without issue. This event marked the beginning of OpenClaw's viral spread.
Today, OpenClaw is one of the fastest-growing open-source projects, with a lobster mascot. It's been described as the "operating system for personal AI." The project faced challenges, including a trademark claim that forced a rename and attempts to take away the lobster mascot, alongside the AI model many users preferred. Steinberger was close to abandoning the project but was inspired by the innovative ways people were using it.
Real-World Impact
At ClawCon, a conference for OpenClaw users, Steinberger met individuals like Gerhard, a 60-year-old who had never coded before. Gerhard and his son used OpenClaw to automate a 90-minute beer brewing process, then built a website and payment system for their product, all primarily using their phones.
In China, "raising lobsters" (installing OpenClaw) has become so popular that people queue at Tencent offices for installation, and Shenzhen offers subsidies for businesses using the platform. However, Steinberger also noted the potential risks, such as employees being fired for not using OpenClaw to automate tasks, or even for using it too much.
The Future of Agents
Steinberger introduced a "heartbeat" feature, allowing the agent to periodically check emails and calendars, proactively following up on tasks. He acknowledged this could be seen as scary, but as an independent builder, he prioritised open access. He envisions a future with multiple specialised agents – for work, personal life, health, relationships – all collaborating securely.
He founded the OpenClaw Foundation to ensure the project remains open-source forever. OpenClaw has transformed AI from a daunting concept into something fun, useful, and a bit quirky. The project has enabled people without programming backgrounds to become builders, automating tasks and creating businesses. Steinberger concluded that the real transformation is not just the technology itself, but the access it provides, changing who can build and create. The door to this new era of creation is open, and the next breakthrough could come from anywhere.