Why Coding is The First Undeniable AI Use case | a16z

Generative AI is currently in a state of extreme disequilibrium, characterized by massive capital expenditure and a narrow focus on agentic coding as its primary product-market fit. Foundation models likely function as low-level infrastructure rather than end-user products, suggesting that long-term value will accrue further up the stack, similar to historical platform shifts like mobile data. While current usage centers on automating existing workflows, the most transformative applications will emerge from solving problems that were previously impossible or prohibitively expensive. As the industry matures, the focus must shift from simply increasing model size and compute to identifying specific, high-value use cases across professional services and other sectors. Ultimately, these technologies will become invisible, integrated utilities that fundamentally reconfigure how businesses operate and create value, much like the evolution of previous computing paradigms.