Results for “agents”

…systems rather than classical medical devices AI research, agents, and open models A notable research meta-point: Chinese open-source literature over the last year was highlighted as unusually high-ROI to follow, with the claim that the…

…four web-agent benchmarks and even surpassing some proprietary agents (allen_ai). In parallel, GenReasoning launched OpenReward, a platform exposing 330+ RL environments, autoscaled environment compute, and 4.5M+ unique RL tasks through one…

…“bring your own harness” workflows for coding and computer-use agents across heterogeneous execution setups, as highlighted by Johannes Hage and in a follow-up deep dive. The release was framed by team members as months of infra…

…by eval scaffolding, access controls, and harness design. Agents, harnesses, eval environments, and continual improvement infrastructureA recurring meta-theme across many tweets was that model capability is increasingly bottlenecked by…

…Several launches targeted the missing systems layer around agents. @Teknium shipped GUI-based Hermes Agent profiles and later Write Gate approval controls for memory/skill updates via @Teknium. @weaviate_io described structured agent memory…