On evaluating multi-step scientific tool use in LLM agents. SciAgentGym provides an interactive environment with 1,780 specialized tools across 4 scientific disciplines. The core finding: even advanced models like GPT-5 see success rates drop sharply from 60.6% to 30.9% as
@dair_ai
-
Top AI Papers of the Week: February 9-15
By
–
The Top AI Papers of the Week (February 9-15) – Gaia2
– SkillRL
– LLaDA 2.1
– EchoJEPA
– InftyThink+
– AdaptEvolve
– AgentSkiller Read on for more: -
Improving Efficiency of Evolutionary AI Agents
By
–
// Improving Efficiency of Evolutionary AI Agents // Evolutionary AI agents are powerful but can be wasteful. Systems, inspired by AlphaEvolve and OpenEvolve, iteratively generate, mutate, and refine candidate solutions using LLMs. However, every refinement step invokes the
-
Shallow Exploration Trap in Test-Time Reasoning Models
By
–
Test-time reasoning models often converge too early. Achieving broader reasoning coverage requires longer sequences, yet the probability of sampling such sequences decays exponentially during autoregressive generation. The authors call this the "Shallow Exploration Trap." This
-
Build Apps with Claude Code: Complete Course Guide
By
–
This course is useful for both technical and non-technical builders. You also get to deploy your app and learn how to use Claude Code for it. Certificates of completion available. Enroll here: https://
academy.dair.ai/courses/build-
apps-with-claude-code
… -
Claude Code: Building AI Apps with On-Demand Course
By
–
Claude Code is one of the most powerful tools for AI builders. We are excited to introduce our new on-demand course on how to vibe code AI apps with Claude Code. We cover topics ranging from connecting services to memory to leveraging Skills to build full AI apps.
-
AgentSkiller: Automated Multi-Turn Agent Training Data Framework
By
–
// Scaling Generalist Agent Intelligence // This new research introduces AgentSkiller, a fully automated framework for synthesizing multi-turn interaction data across realistic, semantically linked domains. The architecture uses a Directed Acyclic Graph to ensure determinism and
-
Defining Data Agents: The Challenge of Ambiguous Terminology
By
–
Everyone is building "data agents" but nobody agrees on what that means. The term gets applied to everything from a simple SQL chatbot to a fully autonomous data scientist. This ambiguity makes it impossible for users and builders to reason about what a system can actually do.
-
Top AI Papers: Agents, Reinforcement Learning, and World Models
By
–
The Top AI Papers of the Week (January 26-February 1): – Kimi K2.5
– VibeTensor
– Insight Agents
– Self-Improving Pretraining
– Open-Source World Simulator
– How AI Impacts Skill Formation
– Orchestrating Reliable AI Agents
– Reinforcement Learning via Self-Distillation Read on -
World Models Replacing Video Generation in 2026
By
–
World models are going to take over in 2026. While video generation models are impressive dreamers, they aren't world simulators. Generative video models can produce stunning clips, but they hallucinate pixel transitions based on statistical correlations. This leads to