Articles & Guides

Long-horizon agents need to judge not only what a tool call does now, but whether it will help complete the task later. The proposed CITA framework trains a model to compare candidate calls before execution.

A new study uses provenance-rich synthetic tasks to test whether training-data attribution identifies examples that truly shape a tabular foundation model. Its results show that intervention-based faithfulness and provenance similarity are related, but not interchangeable.

ArrivalBench shows that passing a one-shot snapshot test does not prove an agent-generated data pipeline is reliable. Replay under late, duplicated, out-of-order, and retried deliveries exposes failures that ordinary grading misses.

Inherit-MAS introduces inheritance at two levels: it preserves useful workflow structure while reusing past executions only when requests and contexts fully match. The authors report stronger benchmark results and lower worker-token usage than several evolving multi-agent baselines.

Google has paused its Open Source Software Vulnerability Rewards Program after a sharp increase in automated submissions. The company says most of the new reports were invalid.

President Donald Trump has announced a new Super Intelligence Force to coordinate the federal government’s AI efforts. The group is expected to produce a report on AI risks and opportunities within 120 days.

Strata is an open-source deployment project for running Qwen3.8-Flash-Next, a roughly 125-billion-parameter model, on a personal computer with at least 12GB of VRAM. Its published tests show tens of tokens per second on systems using GPUs such as the RTX 5070 and RX 9070 XT.

Microsoft and Hugging Face’s ThinkingBox evaluates agents by the backend state and side effects they leave behind, then repeats each task 20 times to measure reliability.

Capcom says it is exploring a gradual evolution of the RE Engine toward AI-supported game development. For now, the company frames AI primarily as a way to improve workflows rather than generate in-game assets.

David Robinson, who helped write safety reports for major OpenAI launches, has resigned and criticized the company’s culture. He argues that frontier AI labs need high-reliability safety practices rather than relying mainly on iterative deployment.

David Robinson, who helped write safety reports for major OpenAI model releases, has resigned and publicly criticized the company’s culture. He argues that frontier AI labs need the layered redundancy and deliberate planning associated with nuclear plants and busy airports.

OneStreamer is a 4B streaming video model designed to connect real-time perception, reusable memory, and proactive responses. It turns observed evidence into timestamped captions and event summaries so information can remain available after the original frames leave the visual window.

Remembering interaction history does not guarantee that an agent understands the current world or makes progress toward its goal. PoS maintains explicit beliefs at inference time and detects, diagnoses, and recovers from stagnation, cycles, and drift.

AutoGUIWorld combines an image generator with a task planner to synthesize GUI interaction trajectories without deploying the underlying software. The resulting data improves computer-use agents across several desktop and browser environments.

A new study suggests that many pretrained Transformers do not lack the capacity for long reference chains; they simply stop using it early. A rank-8 LoRA placed at one early layer can make frozen middle layers carry the computation much farther.

X-Tree mines recurring, success-related action spans from agent trajectories and organizes them into a reusable experience hierarchy. The structure is then used in offline RL, online RLVR, and on-policy self-distillation without LLM calls.

PhysVista introduces a benchmark that evaluates whether vision-language models understand physical consistency rather than merely recognizing visual content. It combines physical state perception, dynamics reasoning, and plausibility assessment across real-world and AI-generated videos.

A new study turns intent drift into a measurable failure mode: superseded or withdrawn user requirements can continue to shape an agent’s answer or tool action. Its IntentFlux benchmark and StateForge method show that explicit state maintenance helps, but does not eliminate the gap between evolving dialogues and direct single-turn tasks.

RouteFM treats LLM routing as a reusable capability rather than a workload-specific fitting problem. By reading limited behavioral evidence from anonymous candidate models, a frozen router can adapt to new tasks, model pools, modalities, and context budgets.

OpenTumorBoard turns public multidisciplinary tumor board recordings into a benchmark for evaluating models on specialist answers and full clinical discussions. Its results show that even advanced general and medical models still struggle to reproduce expert responses and board-level consensus.

Audio LLMs may recognize a command correctly and still act when they should remain silent. VGBench evaluates whether voice agents can determine who is being addressed before invoking tools or responding.