Between 2024 and 2026, Raymond Hou (Raymond) systematized the working methods he had accumulated while building a business from scratch into a methodology for the public, and in 2026 handed his workflows over to an AI agent. The period converged around two developments: integrating scattered courses, community, and content into a mutually reinforcing Super Individual business; and moving from “using AI tools” toward “training an AI counterpart that acts as one’s own agent.”
Naming: from “a business of one” to “Super Individual”
When Raymond’s team was preparing its flagship course, the initial key phrase was “a business of one.” However, Raymond did not encourage readers to impulsively quit their jobs and start businesses. Surveys also showed that only about 15 percent of learners actually wanted to start a company, so he switched to “Super Individual” to avoid the misunderstanding that the course served only people who wanted to become business owners. He separated several concepts that are often conflated: a company of one is a business model, freelance work is a work status, a personal brand is a result that arises naturally once the necessary elements are in place, and “Super Individual” is a system that continues to upgrade. This definition ran through the entire period and served as the course’s opening proposition.
The KIT model and The Super Individual’s Way of Working
In June 2025, Raymond organized seven years of experience into the 55-unit course The Super Individual’s Way of Working and began fundraising; it officially launched that September. Its core is the KIT model (see The KIT Model: Knowledge, Influence, and Time), which argues that Super Individuals need three systems that are continually upgraded: Knowledge to extract and accumulate expertise; Influence to communicate commercial value and make people willing to pay; and Time to manage tasks and energy. The underlying methodology continued earlier ideas, such as shifting the focus of the second brain from “collecting and categorizing” to “actionability,” and using his original PAI structure (Project / Action / Inbox, see PAI Knowledge and Project Management System) as the smallest unit for knowledge and project management.
Three stages of a company of one and the brand flywheel
During this period, Raymond shared a framework describing three stages of a company of one: the labor stage, the stable stage, and the systems stage. He emphasized that most people remain in the first stage, where the key is not making money but accumulating work and trust assets. He recommends validating demand through services before building a product, not saving everything for one big launch, focusing on only one or two platforms, and keeping “how-to” material in paid content while publicly discussing what one has done and why. The concept connecting the whole business is the “brand flywheel” (see The Brand Flywheel: Let A Help B and C): one activity should help another. Raymond uses content about his Smart Work Home as an example. At first, he published it only as a record of his life; later, he found that a significant proportion of viewers of his home-renovation videos enrolled in his online courses, because both audiences were people who valued efficiency and wanted to manage their lives well.
Digital nomadism and working while traveling
Starting in 2025, Raymond frequently practiced working while traveling, visiting India, China, Malaysia, Singapore, Japan, Thailand, and New Zealand. He standardized a portable office hardware setup built around foldable dual screens, a 4K wireless projector, and simplified cables with unified USB-C charging, so a hotel room could quickly become a workstation. This work style reflects his long-standing position that work and life should be integrated rather than “balanced”: when the work is something one loves, work is part of life.
The Kairos AI agent and “Ray Xiaomeng” counterpart
In February 2026, inspired by a peer who had created an AI assistant using the open-source OpenClaw framework, Raymond decided not to deploy an existing framework directly, but to build an AI agent from scratch with Claude Code.1 He compared directly adopting a finished product to “adopting a 20-year-old adult”: one would miss the bond of helping it grow. He preferred to move more slowly and understand every part. The first step was not a flashy feature, but a memory mechanism that could let the AI read his ten-year knowledge base and retain context across conversations. On February 8, he asked the AI to name itself. From several Greek options, it chose “Kairos” (a fitting or opportune moment). For external use, he brought its counterpart “Ray Xiaomeng” into the team Discord, so colleagues could ask “How would he judge this?” while he was traveling (see Raymond’s AI Double).2
Raymond recorded the process as “21 days of raising an AI counterpart,” describing the system’s progression from “I go to the AI,” to “the AI comes to me,” and then to “the system starts improving itself” (automatically reviewing its work and proposing improvements when the weekly quota resets).3 He repeatedly emphasized that a byproduct was greater self-understanding: writing down his email tone, how he prioritizes decisions, and when to answer directly versus ask a question—“intuitions” turned into executable rules—forced him to see more clearly who he is. He also took “in 2026, one person can replace an entire team” as a premise of the era, while noting that this depends on first accumulating enough content and knowledge for the AI to know how to act on one’s behalf.
Business milestones and wider context
Toward the end of the period, Raymond’s business metrics continued to grow: the “Raymond Weekly” newsletter passed 25,000 subscribers, its open rate remained above 65 percent, and the blog had accumulated more than 300 long-form articles. The paid membership community “Lifehacker Islands” entered its sixth year. During one overseas trip, he considered closing the service but ultimately decided to continue. In his public arguments, he cited industry leaders who treated “whether someone uses AI well” as a criterion for evaluating talent. As tools change, evaluation criteria should change too. He compared it to “an accountant who cannot use Excel,” arguing that refusing to use an appropriate tool is itself a waste of ability. He condensed the core insight of this period into one sentence: the greatest byproduct of training AI is understanding oneself better.
A turning point and a summary at age 32
On September 17, 2026, his 32nd birthday, Raymond published a long reflection on five and a half years of entrepreneurship, marking a turning point in the first half of his personal business. He divided the entrepreneurial period into the first three years, the “labor stage” (when he and his wife Yuzu lived with almost no weekends), and the following two years, the “systems stage” (when the personal-business flywheel matured). At the same time that business was running more smoothly, life brought severe tests: his father died in late 2025, and his family was hit by a scam, one event after another. This made him deeply aware that entrepreneurs have no bereavement leave or allowance, yet still carry the responsibility of keeping operations running through unexpected events. Drawing on the Diamond Sutra’s metaphor of “the raft,” he emphasized the wisdom of letting go and saying goodbye to expectations, rather than making past achievements a burden in the next stage. He reminded himself to “finish well and sleep well.” He also announced that in 2027 he and Yuzu would “step away from the screen,” slow down their work, and visit mentors, friends, and partners they had known for years, bringing the first half of their atypical business to a considered close.4
Continuity from the previous stage
This stage did not begin from scratch; it turned earlier accumulation into forms that could scale. Writing remained the underlying leverage. Raymond argued that accumulating personal writing would be the greatest protection in the AI era because this content helps an AI assistant know its user better. Likewise, his smart home, remote freelancing, and long-term approach to work were reassembled into two themes: “Super Individual” and “AI counterpart.” The years 2024–2026 were when he integrated scattered methods into a system, then handed part of that system to AI.
Sources cited
Footnotes
-
Hou Zhixun, “My AI Agent,” public Facebook post, 2026-02-13. View original ↩
-
Hou Zhixun, “I Just Asked My AI Agent to Name Itself,” public Facebook post, 2026-02-08. View original ↩
-
Hou Zhixun, “Raising an AI Assistant Counterpart for 21 Days | How Non-Engineer Owners and Workers Can Use One,” public Facebook post, 2026-03-13. View original ↩
-
Hou Zhixun, “To My 32-Year-Old Self: This Year’s Birthday Was Especially Ordinary and Quiet,” public Facebook post, 2026-09-17. View original ↩