Labelbox CEO explains how AI shifted from building models to renting intelligence

Labelbox CEO Manu Sharma joined Andreessen Horowitz partner Matt Bornstein on the AI + a16z podcast to discuss the evolution of data labeling and evaluation in artificial intelligence. The conversation highlighted how the industry has shifted from pre-training to post-training optimization, with companies now building global networks of domain experts to fine-tune AI systems and align outputs with user expectations.

What you should know: The AI industry has fundamentally transformed from building custom models to renting base intelligence and enhancing it for specific use cases.

• Labelbox originally focused on computer vision but pivoted as foundation models and generative AI changed the competitive landscape.
• The company now operates a global network of experts from fields like coding, healthcare, and customer service who label and evaluate data for AI fine-tuning.
• Modern AI systems are increasingly trained not just to answer questions but to assess the quality of their own responses.

The big picture: Data labeling has evolved from simple supervised learning tasks to sophisticated reinforcement learning loops that require domain expertise.

• Value creation has shifted from the pre-training phase to post-training optimization and evaluation.
• Companies are no longer primarily building their own models but instead renting foundational intelligence and customizing it.
• The focus has moved toward aligning AI outputs with human expectations through expert evaluation.

What they’re saying: Sharma explained the strategic pivot Labelbox made to adapt to the changing AI landscape.

• “It took us some time to really understand like that the world is shifting from building AI models to renting AI intelligence. A vast number of enterprises around the world are no longer building their own models; they’re actually renting base intelligence and adding on top of it to make that work for their company.”
• “The even bigger opportunity was the hyperscalers and the AI labs that are spending billions of dollars of capital developing these models and data sets. We really ought to go and figure out and innovate for them.”

Why this matters: The conversation underscored how critical high-quality data and specialized talent have become in the race toward artificial general intelligence.

• Meta’s acquisition of Scale AI demonstrates the strategic importance of data labeling companies.
• Expert evaluation networks are becoming essential infrastructure for AI development.
• The shift from software-tools mindset to service-oriented approaches reflects broader industry transformation.

Key challenge: Sharma acknowledged the difficulty of transitioning from a traditional software company approach to serving AI labs and hyperscalers.

• “The hardest part for many of us, at that time, was to just make the decision that we’re going just go try it and do it. And nothing is better than that: ‘Let’s just go build an MVP and see what happens.'”