During an interview with Bloomberg on March 26, 2026, Peter Steinberger, known as "Lobster," the founder of OpenClaw, shared his deep insights into the current state of global AI applications.
Steinberger pointed out that there is a significant "temperature difference" between China and the United States in embracing AI agents: Chinese companies show a near-obsessive drive, even making it a required course for employees; while some American companies are moving to restrict or even ban its use due to concerns about security and loss of control.
Survival Rule in the Workplace: Employees Who Don't Use AI May Be Laid Off
Steinberger summarized this difference with an extreme contrast: in the United States, some companies may fire you for using an AI tool like OpenClaw in violation of policies; but in China, many companies may fire you for refusing to use it to improve efficiency.
He observed that students, professionals, and even the elderly in China are actively participating in testing OpenClaw. This nationwide mobilization has made China a vast testing ground for AI systems. He advised American companies to learn from this experience, emphasizing that humans can only truly understand the potential of AI and avoid security vulnerabilities through frequent interaction and testing.
Personal Agent Vision: Blurring the Lines Between Programming and General Purpose
Steinberger, who has joined OpenAI and is now leading the Codex team, revealed that the next phase of AI will break the limitations of "specialized tools." As agents' self-iteration capabilities improve, the boundaries between "programming-specific" and "general-purpose tasks" are disappearing.
In his vision, future users will have personal agents and work agents that span devices and applications. These systems will enable seamless data access while strictly protecting commercial secrets and personal privacy, allowing AI to evolve from a "chat window" into a "digital twin" capable of handling complex real-world tasks.
