Instead of competing with labs on model training, the defensible strategy is to build the ideal environment or 'habitat' for an LLM in a specific vertical. Replit did this for programming by adapting its editor, cloud infrastructure, and deployment tools to serve the AI, not just the human.

The notion of building a business as a 'thin wrapper' around a foundational model like GPT is flawed. Truly defensible AI products, like Cursor, build numerous specific, fine-tuned models to deeply understand a user's domain. This creates a data and performance moat that a generic model cannot easily replicate, much like Salesforce was more than just a 'thin wrapper' on a database.

The pace of AI development means a startup's competitive advantage can be erased overnight by the next model release from a major lab like Google or Anthropic. Dr. el Kaliouby stresses that true defensibility now requires more than just a proprietary algorithm; it demands unique data, distribution, or IP that cannot be easily replicated.

Competitors trying to distill a specific OpenAI model miss the real advantage. The durable moat is the entire "machine that makes the models"—the infrastructure, data, and talent. By the time a competitor copies one model, OpenAI's factory is already building the next, better one.

As foundational AI models become commoditized, the competitive advantage is no longer raw intelligence. Lasting value comes from building a reliable ecosystem around the AI, focusing on deep workflow integration, governance, user trust, and flawless operational execution. This is the true defensible moat.

Greg Brockman reveals that OpenAI's core defensibility isn't any single model, which can be imitated. Instead, their strategic advantage is the end-to-end, repeatable system—a combination of people, processes, and infrastructure—that consistently produces next-generation models.

In the AI era, defensibility comes from building a complex system of record, not just a thin wrapper on an LLM. Companies with a 'thick application layer' that offers standalone value are unattractive for model providers to replicate, whereas thin wrappers risk being absorbed by the platform they are built on.

Powerful AI products are built with LLMs as a core architectural primitive, not as a retrofitted feature. This "native AI" approach creates a deep technical moat that is difficult for incumbents with legacy architectures to replicate, similar to the on-prem to cloud-native shift.

A complex "applied AI layer" is emerging as the source of durable value in enterprise AI. This goes beyond simple API calls to include model routing, bespoke workflow integration, and unique human-in-the-loop interfaces. Companies building this complex layer gain a defensible moat that thin wrappers on LLMs cannot replicate.