I trained a GPT to think like Steve Jobs and help me run my company. AI is scary, but it's also my biggest tool.

The rise of artificial intelligence is both thrilling and intimidating, especially for those navigating the startup landscape. Transitioning from a team of more than 20 engineers to an AI-driven workforce of six has transformed the way I operate my business. My experience has taught me valuable lessons about integrating AI into the workplace, particularly for those engineers who are being encouraged to embrace this technology. As a co-founder who transitioned from Amazon to launch my own startup, I quickly discovered that the habits learned in big tech don’t always apply in a startup environment. Now, I use AI as a strategic intermediary, empowering my junior team members to produce work that rivals that of senior-level professionals. Renowned investor Ben Horowitz once suggested that effective teams should be no larger than a basketball squad, a philosophy I’ve adopted in my lean operation. To maximize the potential of AI, I realized I had to be strategic about the questions I asked. I developed a custom GPT modeled after Steve Jobs to serve as a mentor and guide in my decision-making process. This innovative tool allows me to concentrate on broader strategic objectives while my AI assistants handle specific tasks. Creating over 17 customized GPTs has been crucial for optimizing my operations. Initially, I aimed for a compact team of four AI agents, but I soon recognized that spreading tasks too thin resulted in diminished performance. Instead, I tailored each AI to focus on distinct responsibilities, which ultimately led to a more effective virtual team. The process of building these AI assistants is rapid; I can create a new GPT in just a few minutes. However, the real power lies in the extensive training each one undergoes. By developing standard operating procedures and client-specific training materials, I ensure that my AI agents are well-equipped to handle tasks that align with my methodology. This approach allows me to craft communications and content that resonate with my clients' unique voices, which feels like a labor-intensive process but is achieved with efficiency thanks to my AI team. My toolkit now includes specialized roles such as market researcher, sales analyst, proposal writer, and even an evaluator for LinkedIn profiles, all designed to free me up for more significant strategic planning. In my quest for guidance, I pondered which mentors—living or deceased—I would want by my side. Steve Jobs emerged as an ideal figure, known for his creativity and innovative thinking. I meticulously trained my custom GPT to emulate his approach, using video transcripts that capture his strategies and product launches. This involved around 40 hours of research to develop a comprehensive training framework that continues to evolve as I unearth new insights. While training AI can be a complex endeavor, I focus on crafting prompts that encourage honest feedback. Rather than simply seeking approval, I ask for evaluations on a scale, pushing the AI to draw on the extensive knowledge it has absorbed from Jobs. This iterative process often yields the most insightful and constructive critiques. The introduction of advanced AI tools, like NotebookLM, has stirred both excitement and apprehension in me. I often contemplate the far-reaching implications of AI—wondering how it might reshape society in the coming years. While I sometimes feel overwhelmed by these thoughts, I remind myself that AI, in its essence, is a powerful tool. Its true potential is unlocked when we harness our human expertise alongside it, transforming it from a mere replacement into a collaborative partner in innovation. As I navigate this new frontier, I recognize that the path forward is one of integration and growth, where AI acts as an extension of our abilities, enhancing rather than replacing the human touch.