DeepMind's AlphaEvolve broke a 56-year-old record by multiplying 4x4 matrices in 48 scalar multiplications, while recovering 0.7% of Google's worldwide compute and cutting Gemini's training time by 1%…
Tencent released Hy3, a 295B MoE open-weight model under Apache 2.0, positioning it as competitive with larger systems on reasoning and coding tasks. The release comes amid a quiet period in AI news a…
A developer explains the inner workings of the ReAct (Reason + Act) pattern for AI agents, distinguishing between the message-level and state-level perspectives. The core of ReAct is not the prompt fo…
BenchFlow released a curated, annotated library of over 443 resources for building and evaluating AI agents, including papers, blog posts, talks, and tools. The list is verified, pruned of dead links,…
Researchers from Princeton and other institutions introduced CoALA (Cognitive Architectures for Language Agents), a framework that structures language agents with four memory systems: working, episodi…
An AI agent harness is the software infrastructure that wraps around a large language model to enable it to act on tasks, not just respond to prompts. The harness connects the model to tools, memory, …