AsianFin— Palantir Technologies Inc. executives downplayed the standalone capabilities of large language models (LLMs), characterizing them as “jagged intelligence” during the company’s second-quarter earnings call, even as the data analytics firm reported its first-ever billion-dollar quarter.

“LLMs, on their own, are at best a jagged intelligence divorced from even basic understanding,” Ryan Taylor, Palantir’s Chief Revenue Officer and Chief Legal Officer, told investors Monday. “One moment, they may outperform humans in solving complex problems; the next, they commit catastrophic errors no human would make.”

The Denver-based company, a key player in government and enterprise AI software, said its next phase in the AI race will focus on integrating LLMs into its platforms in a way that mitigates their weaknesses while leveraging Palantir’s deep domain expertise and data orchestration capabilities.