Box CEO Aaron Levie stated that advancements in artificial intelligence fundamentally depend on how well systems can be evaluated and measured. In a post on X, Levie argued that almost all AI model and agent progress is downstream from evals, adding that improvements in open-weight models, post-training for specific domains, and agentic AI systems all rely on rigorous evaluation methods. He emphasized that agent improvements in the applied AI layer and enterprise AI deployments capable of augmenting work are all about evals. His comments echoed a broader industry discussion highlighted by Garrett Lord, who noted that companies struggle to move AI beyond pilot programs because they lack consistent ways to measure performance.
Box IncCEO's comments on AI evaluation importance are relevant to Box's enterprise AI strategy but no concrete impact.
Meta Platforms Inc.
NVIDIA Corporation