The future of company operations involves creating cascading AI loops for various functions. These loops optimize tasks to a local maximum, but then plateau. Human intuition and out-of-distribution thinking are critical to initiate the jump to the next level of innovation, or the 'next hill'.
The popular fear of falling behind due to AI is a 'dark fantasy.' This fear overlooks that the AI stack is highly decentralized (not winner-take-all), the real-world economic diffusion of technology is slow, and most business problems are not purely limited by intelligence.
In the pre-AI era, VCs often rejected ideas that seemed too ambitious. Today, the calculus has flipped. With AI dramatically expanding execution capabilities, an idea that is not ambitious enough is a major red flag, as it suggests a limited vision for what's now possible.
Organizations will adopt a dual-model AI strategy. For functions with unbounded upside like drug discovery or sales, they'll pay a premium for frontier models where a small performance edge yields massive returns. For bounded-upside functions like legal or finance, they'll use more cost-effective, specialized models.
People fundamentally want to 'spend time' on fulfilling activities more than they want to 'save time.' The greatest opportunity for consumer AI is not in productivity tools, but in applications that address core human needs like connection, love, fun, and personal progress. This is a product design challenge, not a capability one.
The primary effect of AI isn't just making existing tasks more efficient; it's dramatically expanding the scope of what is considered possible. CEOs aren't aiming for a more efficient version of their current company, but a vastly larger one. The new bottleneck for growth is a lack of ambition and imagination.
The most effective way to develop intuition and skill with AI is to constantly build and ship small projects. This shifts the focus from building as a means to an outcome, to building as an activity for learning and personal fulfillment. It's okay if these projects are ephemeral or seem unimportant.
With AI enabling the creation of wildly ambitious products that can command high prices, the failure to gain traction is no longer a distribution issue. Instead, it's a 'failure of our collective imagination.' If you can build anything, the inability to find a market reflects a product problem, not a growth one.
Contrary to the freemium-dominated consumer software landscape, there's a significant opportunity for high-priced products. Price is a measure of product-market fit, and founders should challenge themselves by asking, 'What would our product have to do to be a software Birkin bag worth $1,000 or $10,000 a month?'
With the underlying AI technology becoming more accessible, defensibility doesn't come from how hard the software is to build. Instead, founders must focus on classic, durable moats from business strategy: network effects, brand, scale advantages, and proprietary data. These fundamentals are more critical than ever.
Sophisticated users realize that frontier AI models are not fungible. Each has a unique 'shape' or 'personality' suited for different tasks. For example, Quinn is creative and excels at storytelling, whereas GLM-5 is like a 'neurotic PhD' ideal for precision. Choosing the right model is like choosing the right mind for the job.
Founders shouldn't over-engineer a moat on paper. True defensibility is often discovered, not designed. By focusing on shipping a high-NPS product that users love, a moat will naturally develop over time through emergent properties like proprietary data traces, brand loyalty, or deep user workflow integration.