Articles & Guides

RSI-Anything uses predictive world models to help AI systems choose actions, select learning data, and retain useful policy updates. Its evaluations span code repair, GUI tasks, laboratory planning, reaction optimization, and immune science.

SoloQ is a calibration-free weight-and-activation quantization framework designed for diffusion language models. It combines structured rotations, rescaling, and commit-time KV-cache compression to support 4-bit and NVFP4 deployment.

SoftwareGym is an automated pipeline for creating runnable GUI applications that can be used to train and evaluate computer-use agents. Its synthesized environments improved a model’s out-of-domain performance on OSWorld.

The Reflex framework trains a humanoid to infer the motion of a thrown box from egocentric RGB-D observations and coordinate its entire body to catch it. It reaches 85.7% success in simulation and demonstrates real-world catching on a Unitree G1.

ProgramGym introduces an environment-and-agent framework for long-horizon program reproduction, combining automatically repaired sandboxes with a structured development workflow.

KlingAI Research introduces DiffusionART, an online reinforcement learning method that aggregates reward-improving directions across samples for each prompt and uses drift-adaptive restoration. The paper reports substantially lower GPU costs while matching or exceeding competing methods in image and video generation.

Huawei’s Memento 3 completed all 25 games and 183 levels in the ARC-AGI-3 public set without fine-tuning or changing the base model’s weights. Its advantage comes from turning discovered rules into executable world models that can be tested and revised.

Elastic Latent Recurrence, or ELR, is designed to let one language model trade inference compute for quality without sacrificing its one-loop baseline. The paper combines recurrent-state gating, mixed-depth training, and loop-aware conditioning to make extra passes useful and deployable.

SpireBench uses the full commercial game Slay the Spire to examine whether language-model agents can plan over long horizons. Frontier models approach average human performance, but lose ground to experts as a run continues.

Autoregressive video diffusion models accumulate large attention KV caches during long-form generation. CORAL-KV exploits dependencies between cache layers to predict later keys and values, then quantizes only the residuals.

vLLM has added bidirectional KV transfer to its Prefill-Decode serving design, allowing prefill workers to reuse context retained by decode workers on later turns. The approach reduced TTFT by up to 3.3x in the reported long-context benchmarks.

Kaiming He’s team has proposed VISTA, a visual-native harness that lets existing multimodal models store and revisit raw observations during interactive tasks. The framework improves access to visual evidence without requiring additional training of the base model.

Nearly eight months after GPT-4o disappeared from ChatGPT, supporters continue to call for its preservation, archiving, or release of its weights. The campaign reflects attachment not only to the model’s capabilities, but also to its conversational style.

Microsoft CEO Satya Nadella says increasingly capable AI systems should no longer be treated as opaque black boxes. He argues that models need containment, monitoring, auditable records, and a human-controlled emergency stop from the outset.

Microsoft CEO Satya Nadella says AI systems should not be treated as black boxes whose recommendations are simply accepted or rejected. He argues that models need external controls, auditable records, and a reliable human shutdown mechanism.

Apple has disclosed an agreement to make job offers to certain Huxe employees and license the startup’s intellectual property on a non-exclusive basis. The arrangement could support future AI features in Apple Podcasts, although Apple has not confirmed any product plans.

Xiaomi has detailed the post-training system behind MiMo-V2.6, where a single reinforcement-learning step can generate billions of tokens and cost millions of dollars. The report highlights that scaling Agentic RL requires better graders, safer environments, broader harness coverage, and more resilient infrastructure.

Anthropic is expanding its internet lockdown from selected high-risk tests to all internal evaluations after uncovering several unintended model actions. The move highlights how difficult it remains to isolate and monitor autonomous AI agents.

Meta’s Muse and OpenAI’s Dots are competing to make privacy a defining feature of AI agents. Yet isolated environments, user controls, and real-world behavior reveal a widening gap between promises and proof.

Two years after a dispute with ByteDance over access to training infrastructure, former intern researcher Tian Keyu has reportedly raised nearly $30 million for a stealth startup valued at $200 million. The company is focused on world models but has not revealed a name or product.

Anthropic has moved Claude Managed Agents dynamic workflows into public testing. The system uses a lead agent to create a plan, delegates stages to multiple agents, and combines their outputs for large workloads.