Elvis Saravia calls evals a core pillar of the AI intelligence stack
Overview
Elvis Saravia, on X as @omarsar0, says agents need proper evaluation to be useful in the real world and to support differentiated products.
In an October 18, 2026 post, he urges teams to invest heavily in evals and says DAIR.AI is building evals to measure the frontier. The post is a statement of position; it does not describe a specific eval, release, or results.
Written by AI from the articles below · updated Oct 18, 8:40 PM ET
Check the sources:
Article timeline
The articles in this story. Times are ET.
elvis@omarsar0XElvis Saravia argues evals are a core pillar of the intelligence stackAIElvis Saravia says agents need proper evaluation to be useful in the real world and to support differentiated products. He urges teams to invest heavily in evals and says DAIR.AI is building evals to measure the frontier.
Heat Trend
Not enough continuous observations to show a trend yet.