The archive

The Research Digest

Execution harnesses, scoring rules and control interfaces explain consequential differences that model-level comparisons can miss.

The Research Digest

Across agents, robotics and model training, the strongest papers separate useful behavior from the metrics commonly used to predict it.

The Research Digest

Across agents, training and evaluation, the strongest evidence separates improving a proxy from improving the system that practitioners actually deploy.

The Research Digest

Across agents, evaluation, robotics and safety, progress came from supervising the right intermediate state rather than scaling blunt end metrics.

The Research Digest

Across agents, multimodal models and infrastructure, apparent model gains repeatedly vanished when researchers tested the full system.

The Research Digest

The strongest results move reliability from model reasoning into explicit state, memory, communication and audit mechanisms.

The Research Digest

The week’s clearest result is that reliable agents require runtime constraints, process-aware evaluation, and targeted correction—not stronger instructions alone.

The Research Digest

Across evaluation, training, memory, and security, the strongest gains came from controlling the system around the model.