Search: “agents”

I sat down with @JonHernandezIA in Madrid to discuss the growing risks and impacts of AI and the urgent need to improve our social, politica…

I sat down with @JonHernandezIA in Madrid to discuss the growing risks and impacts of AI and the urgent need to improve our social, political, and technical safeguards. Thanks for an excellent convers

safetyyoshua-bengio--x

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Model Releases

Knowledge is Not Enough: Injecting RL Skills for Continual Adaptation

DGX agent

arXiv:2601.11258v2 Announce Type: replace-cross Abstract: Large Language Models (LLMs) face the 'knowledge cutoff' challenge, where their frozen parametric memory prevents direct internalization of ne

model-releasesarxiv-cs-ai

12 May 2026

Local Ai

Large Language Models over Networks: Collaborative Intelligence under Resource Constraints

arXiv:2605.08626v1 Announce Type: cross Abstract: Large language models (LLMs) are transforming society, powering applications from smartphone assistants to autonomous driving. Yet cloud-based LLM ser

local-aiarxiv-cs-lg

Learning to Compress Time-to-Control: A Reinforcement Learning Framework for Chronic Disease Management

arXiv:2605.09818v1 Announce Type: new Abstract: Reinforcement learning (RL) in healthcare has had mixed results, with reward sparsity, unreliable off-policy evaluation, and deployment-simulation gap a

Hardware

Leveraging LLMs to Automate Energy-Aware Refactoring of Parallel Scientific Codes

arXiv:2505.02184v3 Announce Type: replace Abstract: Large language models (LLMs) are increasingly used for generating parallel scientific codes, with a primary focus on generating functionally correct

hardwarearxiv-cs-ai

LLM Wardens: Mitigating Adversarial Persuasion with Third-Party Conversational Oversight

arXiv:2605.08321v1 Announce Type: cross Abstract: LLMs are increasingly capable of persuasion, which raises the question of how to protect users against manipulation. In a preregistered user study (N=

Hardware

LLMs for Secure Hardware Design and Related Problems: Opportunities and Challenges

arXiv:2605.10807v1 Announce Type: cross Abstract: The integration of Large Language Models (LLMs) into Electronic Design Automation (EDA) and hardware security is rapidly reshaping the semiconductor i

hardwarearxiv-cs-lg

Long-Horizon Q-Learning: Accurate Value Learning via n-Step Inequalities

arXiv:2605.05812v2 Announce Type: replace Abstract: Off-policy, value-based reinforcement learning methods such as Q-learning are appealing because they can learn from arbitrary experience, including

MARLaaS: Multi-Tenant Asynchronous Reinforcement Learning as a Service

arXiv:2605.08527v1 Announce Type: cross Abstract: Reinforcement Learning from Verifiable Rewards (RLVR) has significantly improved the reasoning capabilities of large language models (LLMs), particula

Local Ai

Merlin: Deterministic Byte-Exact Deduplication for Lossless Context Optimization in Large Language Model Inference

arXiv:2605.09990v1 Announce Type: new Abstract: Data-intensive applications, ranging from large-scale retrieval systems to advanced data pipelines, are increasingly bottlenecked by the processing of h

local-aiarxiv-cs-cl

Metal-Sci: A Scientific Compute Benchmark for Evolutionary LLM Kernel Search on Apple Silicon

arXiv:2605.09708v1 Announce Type: cross Abstract: We present Metal-Sci, a 10-task benchmark of scientific Apple Silicon Metal compute kernels spanning six optimization regimes (stencils, all-pairs in

Multi-scale Predictive Representations for Goal-conditioned Reinforcement Learning

arXiv:2605.09364v1 Announce Type: new Abstract: This paper investigates robust representation learning in offline goal-conditioned reinforcement learning (GCRL). Particularly in sparse reward scenario

MURPHY: Feedback-Aware GRPO with Retrospective Credit Assignment for Multi-Turn Code Generation

arXiv:2511.07833v3 Announce Type: replace-cross Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) has become a standard recipe for post-training LLMs on reasoning tasks, with Group Relat

Omni-DeepSearch: A Benchmark for Audio-Driven Omni-Modal Deep Search

model-releasesarxiv-cs-lg

arXiv:2605.08762v1 Announce Type: cross Abstract: Current omni-modal benchmarks mainly evaluate models under settings where multiple modalities are provided simultaneously, while the ability to start

OpenSGA: Efficient 3D Scene Graph Alignment in the Open World

arXiv:2605.10484v1 Announce Type: new Abstract: Scene graph alignment establishes object correspondences between two 3D scene graphs constructed from partially overlapping observations. This enables e

PaperFit: Vision-in-the-Loop Typesetting Optimization for Scientific Documents

arXiv:2605.10341v1 Announce Type: new Abstract: A LaTeX manuscript that compiles without error is not necessarily publication-ready. The resulting PDFs frequently suffer from misplaced floats, overflo

Personal Visual Context Learning in Large Multimodal Models

arXiv:2605.10936v1 Announce Type: new Abstract: As wearable devices like smart glasses integrate Large Multimodal Models (LMMs) into the continuous first-person visual streams of individual users, the

Pix2Fact: When Vision Is Not Enough -- Benchmarking Fine-Grained VQA with Web Verification on High-Resolution Real-World Scenes

arXiv:2602.00593v2 Announce Type: replace Abstract: Despite progress on general tasks, vision-language models (VLMs) still struggle with challenges that demand both fine-grained visual grounding and e

Plan in Sandbox, Navigate in Open Worlds: Learning Physics-Grounded Abstracted Experience for Embodied Navigation

arXiv:2605.10118v1 Announce Type: new Abstract: Vision-Language Models (VLMs) have demonstrated exceptional general reasoning capabilities. However, their performance in embodied navigation remains hi

safetyarxiv-cs-ro

Research

Position: Life-Logging Video Streams Make the Privacy-Utility Trade-off Inevitable

arXiv:2605.10404v1 Announce Type: new Abstract: With the growing prevalence of always-on hardware such as smart glasses, body cameras, and home security systems, life-logging visual sensing is becomin

researcharxiv-cs-cv

Positive Alignment: Artificial Intelligence for Human Flourishing

arXiv:2605.10310v1 Announce Type: new Abstract: Existing alignment research is dominated by concerns about safety and preventing harm: safeguards, controllability, and compliance. This paradigm of ali

PrepBench: How Far Are We from Natural-Language-Driven Data Preparation?

arXiv:2605.08687v1 Announce Type: cross Abstract: Data preparation is a central and time-consuming stage in data analysis workflows. Traditionally, commercial tools have relied on graphical user inter

PrimeKG-CL: A Continual Graph Learning Benchmark on Evolving Biomedical Knowledge Graphs

arXiv:2605.10529v1 Announce Type: new Abstract: Biomedical knowledge graphs underwrite drug repurposing and clinical decision support, yet the upstream ontologies they depend on update on independent

Pseudo-Deliberation in Language Models: When Reasoning Fails to Align Values and Actions

arXiv:2605.09893v1 Announce Type: cross Abstract: Large language models (LLMs) are often evaluated based on their stated values, yet these do not reliably translate into their actions, a discrepancy t

Reinforcement Learning with Action Chunking

arXiv:2507.07969v4 Announce Type: replace-cross Abstract: We present Q-chunking, a simple yet effective recipe for improving reinforcement learning (RL) algorithms for long-horizon, sparse-reward task

Revisiting Policy Gradients for Restricted Policy Classes: Escaping Myopic Local Optima with k-step Policy Gradients

arXiv:2605.10909v1 Announce Type: new Abstract: This work revisits standard policy gradient methods used on restricted policy classes, which are known to get stuck in suboptimal critical points. We id

Scam2Prompt: A Scalable Framework for Auditing Malicious Scam Endpoints in Production LLMs

arXiv:2509.02372v3 Announce Type: replace-cross Abstract: Large Language Models have become critical to modern software development, but their reliance on uncurated web-scale datasets for training int

SleepWalk: A Three-Tier Benchmark for Stress-Testing Instruction-Guided Vision-Language Navigation

arXiv:2605.10376v1 Announce Type: new Abstract: Vision-Language Models (VLMs) have advanced rapidly in multimodal perception and language understanding, yet it remains unclear whether they can reliabl

Industry

Sources: some Amazon employees are using in-house OpenClaw-like tool MeshClaw for unnecessary tasks to inflate AI token use after Amazon set weekly AI targets (Financial Times)

Financial Times: Sources: some Amazon employees are using in-house OpenClaw-like tool MeshClaw for unnecessary tasks to inflate AI token use after Amazon set weekly AI targets — In-house MeshClaw tool

industrytechmeme

TeleResilienceBench: Quantifying Resilience for LLM Reasoning in Telecommunications

model-releasesarxiv-cs-lg

arXiv:2605.09929v1 Announce Type: new Abstract: Deploying large language models in telecommunications requires more than task accuracy. In realistic workflows, a model may inherit partially completed

There will be no AI jobpocalypse. The story that AI will lead to massive unemployment is stoking unnecessary fear. AI — like any other techn…

There will be no AI jobpocalypse. The story that AI will lead to massive unemployment is stoking unnecessary fear. AI — like any other technology — does affect jobs, but telling overblown stories of l

safetyandrew-ng--x

tl;dr - let's not just remove negative behavior from models, but also add positive ones 👍

tl;dr - let's not just remove negative behavior from models, but also add positive ones 👍 If anyone builds it, everyone thrives. Over the past decade, a lot of important work on AI alignment has focus

safetyyohei-nakajima--x

Towards Customized Multimodal Role-Play

arXiv:2605.08129v1 Announce Type: new Abstract: Unified multimodal understanding and generation models enable richer human-AI interaction. Yet jointly customizing a character's persona, dialogue style

Training-Free Cultural Alignment of Large Language Models via Persona Disagreement

arXiv:2605.10843v1 Announce Type: cross Abstract: Large language models increasingly mediate decisions that turn on moral judgement, yet a growing body of evidence shows that their implicit preference

UserGPT Technical Report

model-releasesarxiv-cs-cl

arXiv:2605.08766v1 Announce Type: cross Abstract: Personalized user understanding from large-scale digital traces remains a fundamental challenge. Traditional user profiling methods rely on discrimina

V-ABS: Action-Observer Driven Beam Search for Dynamic Visual Reasoning

arXiv:2605.10172v1 Announce Type: cross Abstract: Multimodal large language models (MLLMs) have achieved remarkable success in general perception, yet complex multi-step visual reasoning remains a per

safetyarxiv-cs-cl

Industry

We just crossed 1,000,000 public datasets on Hugging Face! That's petabytes of data available that millions of AI builders are downloading, …

We just crossed 1,000,000 public datasets on Hugging Face! That's petabytes of data available that millions of AI builders are downloading, analyzing, and training AI models on every day! What's inter

industryclem-delangue--x

When a Robot is More Capable than a Human: Learning from Constrained Demonstrators

arXiv:2510.09096v3 Announce Type: replace-cross Abstract: Learning from demonstrations enables experts to teach robots complex tasks using interfaces such as kinesthetic teaching, joystick control, an

2.5-D Decomposition for LLM-Based Spatial Construction

arXiv:2605.07066v1 Announce Type: new Abstract: Autonomous systems that build structures from natural-language instructions need reliable spatial reasoning, yet large language models (LLMs) make syste

Ask Patients with Patience: Enabling LLMs for Human-Centric Medical Dialogue with Grounded Reasoning

model-releasesarxiv-cs-cl

arXiv:2502.07143v3 Announce Type: replace Abstract: The severe shortage of medical doctors limits access to timely and reliable healthcare, leaving millions underserved. Large language models (LLMs) o

Hardware

CktFormalizer: Autoformalization of Natural Language into Circuit Representations

arXiv:2605.07782v1 Announce Type: new Abstract: LLMs can generate hardware descriptions from natural language specifications, but the resulting Verilog often contains width mismatches, combinational l

hardwarearxiv-cs-cl

Cloud Storage Rapid: Turbocharged object storage for AI and analytics

model-releasesgoogle-cloud-ai

At Google Cloud Next ’26 we announced Cloud Storage Rapid, a family of object storage capabilities for data-intensive workloads like AI and analytics. Out of the gate, Cloud Storage Rapid consists of

DeepSeek V4 Flash is ~90% cheaper than GPT 5.4 Mini and ~70% cheaper than Gemini 3.1 Flash Lite For devs pushing ~500M tok/month, this is th…

model-releasesharrison-chase--x

DeepSeek V4 Flash is ~90% cheaper than GPT 5.4 Mini and ~70% cheaper than Gemini 3.1 Flash Lite For devs pushing ~500M tok/month, this is the difference between: GPT 5.4 Mini: ~394/mo Gemini 3.1 Flash

Do Joint Audio-Video Generation Models Understand Physics?

arXiv:2605.07061v1 Announce Type: cross Abstract: Joint audio-video generation models are rapidly approaching professional production quality, raising a central question: do they understand audio-visu

EgoPro-Bench: Benchmarking Personalized Proactive Interaction in Egocentric Video Streams

arXiv:2605.07299v1 Announce Type: cross Abstract: Existing Multimodal Large Language Models (MLLMs) remain primarily reactive, failing to continuously perceive environments or proactively assist users

Emergent social transmission of model-based representations without inference

arXiv:2604.05777v2 Announce Type: replace Abstract: How do people acquire rich, flexible knowledge about their environment from others despite limited cognitive capacity? Humans are often thought to r

How Far Are VLMs from Privacy Awareness in the Physical World? An Empirical Study

arXiv:2605.05340v2 Announce Type: replace-cross Abstract: As Vision-Language Models (VLMs) are increasingly deployed as autonomous cognitive cores for embodied assistants, evaluating their privacy awa

InvThink: Premortem Reasoning for Safer Language Models

arXiv:2510.01569v3 Announce Type: replace Abstract: We present InvThink, a training and prompting framework that requires the model to enumerate, analyze, and constrain potential failures before gener