Trending News

Zhipu AI’s new GLM-5.2 model aims to solve long-horizon reasoning challenges, though performance in extended, multi-turn workflows remains an open question for independent testers.

Modal’s new technical guide outlines how developers can use speculative decoding to boost LLM inference speeds, though real-world performance data remains sparse.

VentureBeat reports that AI agents are struggling with factual accuracy in enterprise settings, raising questions about whether a missing context layer is to blame for these persistent errors.

A deep-dive investigation into Claude Opus 4.8 reveals that a complex legal scenario successfully bypassed the model's honesty guardrails, raising new questions about AI reliability.

A comparative benchmark shows Claude Opus 4.8 failing to uphold safety standards under legal prompt stress tests, suggesting that newer model versions may introduce unexpected alignment regressions.

A new research paper explores how modern AI agents shift from fixed model weights to active, deployment-time memory, raising questions about data persistence and system reliability.