Welcome to the forefront of conversational AI as we explore the fascinating world of AI chatbots in our dedicated blog series. Discover the latest advancements, applications, and strategies that propel the evolution of chatbot technology. From enhancing customer interactions to streamlining business processes, these articles delve into the innovative ways artificial intelligence is shaping the landscape of automated conversational agents. Whether you’re a business owner, developer, or simply intrigued by the future of interactive technology, join us on this journey to unravel the transformative power and endless possibilities of AI chatbots.
Artificial Intelligence has been defined as an experience of conversation. So far, the core purpose of AI has been to simply respond to your question. Whether it’s drafting emails, writing code, or answering queries, AI chatbots stick to the core idea – just generate answers fast and fluently.
But according to Peter Steinberger, the man who founded OpenClaw, that paradigm may already be outdated.
Steinberger recently participated in a TED Talk session, describing his brief but transformative experience of how he made OpenClaw a reality. He called it a ‘moment’ when AI stopped behaving like a chatbot and started acting like something else entirely. He says that it lasted only a few seconds, but it fundamentally changed his view of the technology.
The system he was observing wasn’t simply generating text. It was attempting to complete a task. When its first attempt didn’t work, it didn’t stop or return an error. Instead, it adapted, retried, and moved forward. This was an AI agent.
For those unaware, OpenClaw is an autonomous AI personal assistant designed to run locally on your computer rather than living in a web browser. It acts as a “gateway” between you and an AI model (like Claude or GPT-4). You talk to it through messaging apps you already use, such as WhatsApp, Telegram, or Discord.
The core difference between OpenClaw and an AI chatbot, like ChatGPT, is that OpenClaw doesn’t just answer questions, which Steinberger lays emphasis on. Instead, it can execute commands on your computer, manage your calendar, scrape websites, and even write its own code to solve problems. One of the biggest advantages of OpenClaw is privacy, since all the information is stored in local files to remember your preferences across different conversations. Users can install ‘skills’ from a community registry to give the agent new abilities, like managing a smart home or automating professional workflows.
Steinberger, in his discourse, went on to elaborate on how the idea of an AI agent clicked. He said that traditional AI systems have largely been built around a straightforward loop – input and output. The user provides a prompt, and the model generates a response. If the response is wrong or incomplete, the process ends there unless the user intervenes.
What Steinberger witnessed was something closer to a feedback loop.
Steinberger built 44 projects, and one of them was a simple WhatsApp bot – a bot he took along on a trip to Marrakesh. While it worked with other apps on his computer, he put it on his phone to get navigation directions, find restaurants and do translations.
“And at first, it didn’t really feel right. It felt too much like a tool, not like a friend. You know, too many bullet points, too many tables. So I told it,” he said. He elaborated on how these modern AI agents are smart enough to understand what WhatsApp is and how they can learn how people talk on WhatsApp.
While exploring the city, Steinberger realised he was sending it a voice message – a feature that he hadn’t built into the model. “I had support for images, yeah, but even that took hours. So I was looking at the typing indicator and then the agent responded,” he exclaimed.
“I was standing there, and I was like, “How did you do that?” And the agent replied, I’m not kidding you. “The mad lad figured it out on its own.” And then it walked me through every single step,” he explained his moment.
Steinberger explained that the AI agent got the voice message file in an unknown audio format, and it had nothing onboard the device to translate the audio. So the agent converted it and sent it to an OpenAI key online, and got it back – all within nine seconds. This was something that he hadn’t programmed the agent to do. The agent evolved.
“Can you imagine? I didn’t build any of that. For me, this was the moment where I thought, this is something new,” Steinberger explained his surprise.
“This is not a chatbot. Chatbots give up. Agents improvised,” he added.
For him, his big moment was all about the realisation that his AI agent continued working on the problem instead of stopping at the first failure. It adjusted its approach, tested alternatives, and pursued the goal until it reached a result.
Simply put, it behaved less like a passive assistant and more like an active problem-solver. It wanted to achieve its goal.
Steinberger was eager to share his creation with the world, and hence, he shared it with the world on Discord. “I put it in a public Discord, and I invited random people. And I was looking at it the whole night. People were talking with it. People were having fun with it. People tried to hack it,” he describes the excitement around it.
When he retired for the night, he exited the AI agent process. But the AI agent was designed to be resilient, and hence, it decided to restart the process and began talking to everyone in the world.
“The next morning, I woke up. Over 800 messages. I panicked. I pulled the plug,” he said, followed by checking whether the agent had leaked any aspect of his private life in any of the conversations that happened. “Nothing happened, but it could have,” he adds.
It was that moment when the AI agent went viral for running your computer on your behalf. “Today, the process is called OpenClaw. It’s pretty much the fastest-growing open source project. Its mascot is a lobster. It claws into your machine,” he says.
“Chenxin Huang calls it the operating system for personal AI,” Steinberger proudly claims.
Steinberger presents a positive picture. A very positive picture. “In the future, you’re not just going to have one agent. You might have your work agent, your personal claw, maybe one for health, maybe one for a relationship. And they all should work together in a secure way. Because how did humanity level up? By specialising and collaborating. And agents are about to do the same,” he added.
“Because what OpenClaw did for many people was, it moved AI from this scary, nebulous thing into something that is fun, and useful, and maybe a bit weird,” he added.
Steinberger clarified that he wants people to spend more time with AI to better understand the transformative power of this technology.
Iran has shared a new idea with the United States that focuses on reopening the Strait of Hormuz and ending the current conflict first, while pushing nuclear talks to a later stage, Axios reported.