Original Title: OpenClaw Creator Says US Can Learn From China's AI Adoption
Original Author: Shirin Ghaffary, Bloomberg
Translated by: Peggy, BlockBeats
Editor's Note: This article is translated from Bloomberg's interview with OpenClaw founder Peter Steinberger. After joining OpenAI, he is involved in advancing the next generation of AI agent technology. The direction of making AI not just a question-answering tool, but capable of invoking tools, collaborating across systems, and acting continuously in environments is becoming the new competitive core of the industry.
In this interview, he discusses several key questions: What do the different adoption paths of OpenClaw in China and the US mean? How can AI Agents become better? How can secure collaboration between personal and work Agents be achieved? How will OpenAI advance this technical direction?
Following is the original text:
The original design of OpenClaw is for automating tasks such as flight check-in and schedule management.
The creator of OpenClaw (who recently joined OpenAI) believes that more people should personally try using artificial intelligence to learn from it and help society better prepare for this technology. However, before that…
You first need to understand three things:
• OpenAI has ceased support for Sora and is gradually ending its collaboration with Disney.
• Apple plans to rework Siri with AI and introduce a new interface and "Ask Siri" button in iOS 27.
• Amazon acquires Fauna Robotics, entering the consumer humanoid robot market.
Embracing AI agents
After OpenClaw gained popularity for several months, the paths embraced by the US and China for cutting-edge artificial intelligence products have notably diverged, and this difference could have profound effects on the technological competition landscape between the two countries.
In China, more and more people, from students and professionals to the elderly, are trying OpenClaw, with some companies even mandating its use for employees. Although regulators have begun to restrict its application in state-owned enterprises and government agencies, China is becoming an extensive testing ground—allowing AI systems to gradually take over people's digital lives.
In contrast, in the US, OpenClaw (previously known as Moltbot and Clawdbot), despite sparking widespread interest among developers and early users, has yet to create a comparable wave of enthusiasm at the general public level. Some US companies are even starting to limit employees' use of such AI Agent tools due to security risk concerns.
This starkly different market response has also caught the attention of OpenClaw’s founder.
"In the US, I feel like in some companies, if you use OpenClaw, you might get fired," said the tool’s developer, Austrian software engineer Peter Steinberger. He has now joined OpenAI, working on AI Agent-related technology. "In China, many companies are quite the opposite—if you don’t use OpenClaw, you might get fired."
This month, at Baidu's OpenClaw "Lobster Market" event in Beijing, themed merchandise around lobsters was displayed.
Steinberger's product was once called "perhaps the most important software release in history" by Jensen Huang (CEO of Nvidia). However, he also admitted that neither the US nor the Chinese paths are perfect. Though the intention behind OpenClaw is to automate tasks such as flight check-in and schedule management, he pointed out that there are still potential security risks.
Note: Peter Steinberger is an Austrian software engineer and developer, gaining attention for creating the open-source AI Agent tool OpenClaw.
In his new role at OpenAI, Steinberger will participate in the development of Codex, a programming tool currently utilized by over two million users weekly. He is also aware that the market demands higher standards for product safety and stability on such a influential platform, needing to minimize errors as much as possible.
During our conversation, Steinberger talked about how to improve AI Agents, OpenAI's future plans for the technology, and why he will still retain OpenClaw as an open-source project with the support of his new employer, planning to hand it over to an upcoming foundation for management. The following interview content has been moderately edited and organized without changing its original meaning.
Original Interview
Bloomberg: Sam Altman once called you a "genius" and stated that you would drive the development of the next generation of personal AI Agents. What will this look like at OpenAI?
Steinberger: We are rapidly moving towards a future where everyone will have a personal Agent for private life and a work Agent for professional duties. Through OpenClaw, I am essentially building a "window to the future," showcasing what I envision my ideal world to be. Of course, I also understand that currently no company can truly push it to the masses because there are still some key issues that need resolving.
Bloomberg: What specific issues are those?
Steinberger: In that future, my Agent needs to be able to communicate with your Agent. For instance, I work at OpenAI and regularly use Codex for knowledge work, but sometimes I need to access data in my personal "claw." Therefore, there must be a mechanism that allows my work Agent to invoke my personal Agent. At the same time, I must ensure that my personal Agent does not leak any information I consider too private; while OpenAI must also ensure that internal company data isn’t brought back to my personal device.
Bloomberg: You have likely also noted issues arising from employees misusing Agent tools in companies like Meta Platforms, prompting some companies to enhance restrictions.
Steinberger: In the US, I feel like in some companies if you use OpenClaw, you might get fired; while in China, many companies are quite the opposite—if you don’t use OpenClaw, you may be at risk of being fired. They even showed me a form listing each employee's name, alongside a column "What was automated today?" Companies are actively prompting employees to think: how to improve efficiency tenfold.
Neither approach is perfect, but we can indeed learn from adopting newer technologies more swiftly and exploring different risk-taking methods. Because this technology is too fresh, we can only grasp it through constant attempts and trial-and-error.
Even at Meta, a security researcher faced a lot of ridicule on Twitter for publicly discussing related issues. I find that quite brave. If everyone laughs at these attempts, it will only deter more people from speaking out.
Bloomberg: What do you think about the craze for OpenClaw in China? Many are even queuing to experience it. Have you collaborated with Chinese companies?
Steinberger: At GTC, I have communicated with many companies, such as MiniMax, Kimi, Tencent. I can genuinely understand this "fervor" because I have gone through similar moments myself.
A year ago, when I first attempted programming Agents, they had about a 30% success rate, but as soon as you got even a little right, it gave a strong dopamine feedback. You could also realize that this would utterly change the industry, and this was their "worst time," with the future only set to improve. In that moment, I realized I could virtually build anything because everything became faster.
Now imagine if you’re not a tech person, but a small business owner, and suddenly discover: "It can read my emails, manage schedules, write Google Docs, connect my home devices, check WhatsApp, handle customer service requests..." You would undergo the same epiphany that engineers experienced over the past year.
During that time, I even had insomnia because this change was indeed disruptive. I am delighted to bridge AI closer to diverse individuals from various backgrounds.
Bloomberg: OpenAI's Codex has been growing rapidly. What are your thoughts on the integration of Codex and OpenClaw?
Steinberger: One core issue we face right now is: how to help users understand that a product named after "programming" is not just about programming.
If you look at it from a longer-term perspective, all prompts will become stronger due to programming capabilities. AI Agents are smart enough to recognize their shortcomings and will compensate for them by writing code.
Bloomberg: What if the Agent could access all your files and run continuously?
Steinberger: This is actually a question of "how to explain to the user." You can currently connect almost anything in ChatGPT's application ecosystem, such as Slack, Google Docs, Notion, health data, etc. But the current challenge lies in how to help users truly understand: these capabilities are already available.
Another challenge is that if you're working on an open-source project, it can progress very quickly because users are more forgiving, knowing it’s a preview version and won’t be used with work data. But once real work data is involved, the issues become entirely different and take much longer to refine.
I am looking forward to participating in solving these challenges.
Bloomberg: What is the progress of the OpenClaw foundation? Does OpenAI support it?
Steinberger: I try not to let OpenAI become too involved because this project needs to remain independent. Legal and organizational structure improvements will take a few weeks.
Currently, we have some good partners, like NVIDIA, and I have communicated with Microsoft, ByteDance has joined, and Tencent is also progressing. I hope to maintain a "Swiss-style neutrality."
Our goal is to spark more interest in AI and truly start thinking about problems with AI. The most critical aspect for the future is to get more people to spend more time understanding what AI can do, thereby preparing society as a whole. This is the best way to ensure a bright future.
免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。
