3,690 items tracked: 31 reported by two or more outlets, the rest by a single source. Most-corroborated first.
Small AI models may be just months away from spreading on their own across poorly secured consumer devices like smart TVs and robot vacuums, according to a new Anthropic study.
We explain how Ai2’s new GPU scheduler uses time budgets, fair-share allocation, and time-slicing to prioritize high-impact research, shorten queue waits, and keep GPUs busy.
arXiv:2610.11012v1 Announce Type: new Abstract: Powerful misaligned AI models might recognize alignment evaluations and strategically behave well on them, rendering direct audits uninformative.
Three fired OpenAI safety researchers dispute allegations of mishandling sensitive information, warning in an open letter that their dismissals are creating a chilling effect on the company’s AI safety culture.
arXiv:2610.16042v1 Announce Type: cross Abstract: The increasing use of artificial intelligence (AI) in mental health applications raises concerns about its potential to internalize and perpetuate societal biases.
arXiv:2610.17090v1 Announce Type: cross Abstract: On 6 October 2026 OpenAI published an initial 722 AI-written mathematical manuscripts, organized into 372 problem families, in the public repository github.com/openai/mat…
arXiv:2605.16428v4 Announce Type: replace-cross Abstract: Search engines traditionally complement online content platforms by directing users seeking information to external websites.
Researchers say ‘AI worms’ could soon infiltrate internet-connected devices amid calls for increased guardrails Anthropic is warning that some of its artificial intelligence models could soon become “self-replicating” an…
Recent video super-resolution (VSR) models achieve efficient one-step generation through diffusion distillation.
For video generation in dynamic world modeling, observed static and dynamic content should be efficiently reused to provide conditions for repeated generation at different times and viewpoints.
Test-time scaling has enabled LLMs to solve increasingly complex reasoning, discovery, and optimization tasks.
Deep research (DR) agents address open-ended information needs through multi-step reasoning, long-horizon web search, and evidence synthesis.
A robot following a multi-step instruction needs to act on the right objects in the right order, even as the scene changes.
System One models choose an agent's next action by scoring a supplied set of options in a single pass.
Controlling an acting LLM agent means deciding, at each step of a long-horizon task, whether to trust its next action, abstain, or hand off to a human.
System one models enable direct, low-latency decision-making.
Recent advances in vision-language models (VLMs) have substantially improved the capabilities of GUI agents, while memory augmentation has emerged as a promising approach for reusing knowledge from past interactions.
City-scale 3D grounding is commonly formulated as candidate ranking, which requires precomputed instance proposals and becomes inefficient in large outdoor scenes.
Terminal agents have shown strong abilities in complex coding tasks.
Reasoning models achieve strong performance through long chains of thought (CoT), but often spend more tokens than needed to solve a problem.
Long-horizon agents have drawn growing attention for their ability to work autonomously -- from hours to days -- on economically valuable tasks, a capability essential to automated discovery and recursive self-improvemen…
Generating information-rich visual artifacts such as slides, posters, and dashboards as code is appealing because the outputs are structured, editable, and verifiable.
Textual reasoning before acting has improved vision-language-action (VLA) models, but at the cost of latency: every action conditioned on the reasoning text must wait for autoregressive decoding.
Autonomous research agents must acquire and reason over evidence within a limited budget.
Reinforcement learning with verifiable rewards (RLVR) for multi-turn agents relies on sparse terminal rewards that say little about individual steps.
Scaling interactive environments is important for training long-horizon agents through reinforcement learning.
StudentSim is a research system for evaluating and improving AI tutoring tools before they are tested with learners.
A personal eval I run on new AI models is “design a novel game, with mechanics that have never been seen before.” I do this because this is a domain I have enough taste in that I think I can tell if they are doing someth…
Omni-modal large language models (Omni-LLMs) extend LLMs to process heterogeneous modalities (e.g., vision, audio, and IMU signals) alongside language, enabling a broader range of multimodal tasks.
Recent multimodal large language models (MLLMs) achieve strong results on video question answering, yet often fail to recover the temporal structure that makes a video coherent.
Processing multilingual text requires both identifying a broad range of languages and distinguishing the language of each token in inputs containing multiple languages.
Vision-language-action (VLA) models show strong potential for end-to-end autonomous driving by leveraging world knowledge and understanding to excel in long-tail scenarios.
Large language model agents are emerging as an important new class of users of desktop applications.
Large language model (LLM) agents are increasingly deployed as teams, yet they fail in characteristically social ways: information holders stay silent, dominant speakers anchor groups on wrong answers, and norm violation…
AI coding agents now write much of the code that ships, yet whether that code is secure remains unclear.
Language models are increasingly used to create synthetic survey samples, also called silicon samples.
Large language models are increasingly deployed in applications that act on users' long-term behavior histories, from personal assistants to recommendation and advertising systems.
Recent bidirectional encoders have been modernized with innovations from causal large language models (LLMs).
Foundation models need to support diverse inference budgets, yet training a separate model for each budget duplicates pretraining effort.
Deep research requires connecting multiple steps of search, reasoning, and computation, with intermediate findings guiding what to investigate next.
arXiv:2610.11355v2 Announce Type: replace Abstract: Researchers increasingly use ChatGPT to revise their papers, and when asked to make a paper hold up under review, recent GPT versions often narrow or even retract the a…
arXiv:2610.11167v1 Announce Type: cross Abstract: Vertical-domain few-shot classification remains challenging for small language models, as limited supervision makes it difficult to acquire domain-specific decision knowl…
Pixel-space world models have recently emerged as a promising approach for GUI agents, predicting future screens prior to execution and subsequently re-perceiving them for action selection.
Large language model (LLM) agents are becoming direct users of the open web.
The rapid advancement of Large Language Models (LLMs) has driven their deployment toward general-purpose harness environments, where LLM-based agents operate in sandbox workspaces and interact directly with file systems…
Video diffusion models increasingly serve as priors for 4D generation, but the generated videos often lack geometric consistency.
On-policy distillation (OPD) provides teacher supervision on student-generated trajectories, but a single trajectory realizes only one continuation at each decision, limiting supervision coverage under a finite sampling…
LLM agents' capabilities depend not only on their underlying models but also on their execution harnesses.
Large language models have achieved remarkable advances in complex reasoning.
As the production of software passes into the hands of software itself, coding agents are increasingly commissioned to build agents fitted to their users' own workflows.