AI business & value

Martin's thinking on where value accrues in the AI ecosystem centers on two key dynamics: the primacy of foundational models as the source of value creation, and the power of human collaboration through shared open-source infrastructure. Rather than viewing AI as purely a technical problem, he sees both the development of core models and the way practitioners assemble them into solutions as fundamentally collaborative endeavors—shaping how value flows through the emerging AI stack.

Foundational models as the core value creation engine

At the most basic level, foundational AI models themselves are the primary source of value creation in the AI space, ahead of the application layers built on top of them. ▶ 12:18 This insight frames the economic hierarchy of the AI ecosystem: while countless applications and tools layer on top of models like GPT or Claude, the models themselves remain the most defensible and valuable piece.

Human collaboration and the shared stack

Martin frames the layered ecosystem of AI APIs, open-source models, and shared research—GitHub repos, published papers—as an astounding example of human collaboration. He sees this not as fragmentation but as strength: ▶ 26:56 Each practitioner constructs their own puzzle from shared pieces at whatever level of the stack suits their needs. This collaborative foundation has become the backbone of the modern AI landscape.

The future of SaaS: workflows over interfaces

Rasmus Adler Wahlberg, CEO of Multiply, articulates a vision where most SaaS—and Microsoft's traditional software model in particular—will be disrupted by a shift toward human-agent workflows rather than traditional software interfaces. ▶ 28:20 He sees Multiply's value proposition resting on three elements: a flexible data graph readable by both AI and people, workflows that let agents and people get results together, and a shared interface where they work side by side. ▶ 30:24