The video discusses the rapidly evolving landscape of AI app builders, highlighting companies like Lovable, Replet, and Verscel that aim to automate app creation using AI models such as OpenAI’s GPT and Anthropic’s Claude. Despite the excitement around these platforms, the speaker argues that their core offerings are largely similar and easily replicable, leading to a “middleware trap” where differentiation is minimal and moats are short-lived. The real challenge for builders is finding durable niches that AI model providers cannot easily disrupt, especially for those outside the major players like OpenAI or Google.
The speaker identifies five durable verticals of value on the future web that AI cannot replace: trust, context, distribution, taste, and liability. Trust is crucial as the web becomes flooded with AI-generated apps and content, many of which may be low quality or malicious. Companies that establish themselves as verification and accountability layers—such as Stripe, Shopify, and Apple’s App Store—will capture significant value by ensuring safety and reliability in an increasingly agent-driven economy. Trust becomes a fundamental gatekeeper for AI agents transacting autonomously on behalf of users.
Context refers to the unique, specific data and situations that make AI applications truly useful. Unlike general AI models, context is proprietary and tied to individual users or organizations, making companies that own and manage this data—like Notion, Salesforce, and Snowflake—key players in the AI ecosystem. These context layers serve as choke points through which AI agents must pass to access relevant information, creating a defensible position that AI model providers cannot easily replicate.
Distribution remains a critical bottleneck despite the ease of app creation. With an explosion of AI-generated products, standing out and reaching users is more challenging than ever. Established platforms like Google, Apple, TikTok, and YouTube dominate distribution channels, and new mechanisms for agent discovery will be essential as AI agents increasingly interact autonomously. The speaker emphasizes that successful businesses must think deeply about how agents will find and transact with them, as this will be a major determinant of viability in the agentic economy.
Finally, taste and liability are highlighted as uniquely human elements that AI cannot replace. Taste involves the human judgment and editorial decisions that shape what products and experiences resonate with users, while liability concerns who is accountable when AI-driven decisions cause harm or legal issues. Regulated industries like healthcare, finance, and law will especially rely on human accountability, creating opportunities for businesses that specialize in governance and risk management. Overall, the future web will be organized around these five verticals, with AI serving as a powerful tool but not a substitute for human-led value creation.
