Caesar: Deep Agentic Web Exploration for Creative Answer Synthesis

Jason Liang, Elliot Meyerson, Risto Miikkulainen · 2026

Read Caesar: Deep Agentic Web Exploration for Creative Answer Synthesis by Jason Liang, Elliot Meyerson, Risto Miikkulainen (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-20855.txt.

Description

Liang, Meyerson, and Miikkulainen build Caesar, an agent that moves past flat web retrieval toward creative discovery. It traverses the web to assemble a changing knowledge graph, then uses that graph to steer toward varied, less obvious sources. A second stage refines drafts adversarially, seeking fresh angles rather than confirming prior views. In creative synthesis tests it beats leading deep-research agents by 13 to 23 percent.

Catalog record

CREATORJason Liang, Elliot Meyerson, Risto Miikkulainen
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE103230
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.20855

Browse the full library · About the Tower of Babel