safety

Solver-Aware Decompositions for Programming-by-Example: When Dividing Requires Knowing how to Conquer

arXiv:2608.03461v1 Announce Type: new Abstract: Decomposition-based Programming-by-example (PBE) scales performance by splitting tasks into subtasks that a learned synthesizer solves: a decomposer pre

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Safety

Some people are surprised that APIs (aka what Anthropic, OpenAI, and others provide) are treated differently than open weights in the new AI…

DGX agent

Some people are surprised that APIs (aka what Anthropic, OpenAI, and others provide) are treated differently than open weights in the new AI model framework. I'm not surprised at all, and it's actuall

safetyclem-delangue--x

5 Aug 2026

SP3O: Reinforcement Learning from Segment Preferences without Reward Modeling

arXiv:2608.02951v1 Announce Type: cross Abstract: Preference-based reinforcement learning (PbRL) for general stochastic MDPs often requires training a reward model. Existing reward-model-free methods

SPADE: An Input-Adaptive Sparse Attention Engine for Fast Video Diffusion Models Inference

arXiv:2608.03335v1 Announce Type: new Abstract: Video diffusion transformers (vDiTs) generate high quality but pay quadratic self-attention cost, making inference prohibitive at video-token scales. Th

Staying on Spec: Real-Time Monitoring under Uncertainty with a Maritime Case Study

arXiv:2608.02811v1 Announce Type: new Abstract: Robotic systems must operate under uncertainty while satisfying complex task and safety specifications. Monitoring such specifications under uncertainty

Stiffness Copilot: An Impedance Policy for Contact-Rich Teleoperation

arXiv:2603.14068v2 Announce Type: replace Abstract: In teleoperation of contact-rich manipulation tasks, selecting robot impedance is critical but difficult. The robot must be compliant to avoid damag

StreamDAM: Presence-Aware Memory for Real-Time Streaming Video Object Segmentation

arXiv:2608.03912v1 Announce Type: new Abstract: Quality-tier video object segmentation (VOS) trackers such as DAM4SAM top accuracy leaderboards, but they are measured offline, one frame at a time with

string2string Studio: An Interactive, In-Browser Platform for String-to-String Algorithms

arXiv:2608.03984v1 Announce Type: new Abstract: We present string2string Studio, an interactive in-browser platform for string-to-string analysis across natural language processing, computational biol

T2VAttack: Adversarial Attack on Text-to-Video Diffusion Models

arXiv:2512.23953v2 Announce Type: replace Abstract: The rapid evolution of Text-to-Video (T2V) diffusion models has driven remarkable advancements in generating high-quality, temporally coherent video

Test Time Adaptation Methods for Point Cloud Registration in Laparoscopic Surgery

arXiv:2608.02883v1 Announce Type: new Abstract: 3D point cloud registration in laparoscopic surgery estimates the transformation between an intraoperative organ reconstructed from video and its preope

Test-Time Scaling for Safe Text-Guided Image Generation via Intermediate Clean Estimates

arXiv:2608.03284v1 Announce Type: cross Abstract: Ensuring safety and policy compliance in text-to-image diffusion models remains a critical challenge, as benign or adversarial prompts can often elici

The Agent Operating System (AOS): A Reference Operating Architecture for Distributed Agentic Systems

arXiv:2608.03214v1 Announce Type: new Abstract: Large language models have transformed artificial intelligence from isolated prediction services into components of long-running, distributed systems th

🚨The endgame of commoditization and falling prices and lack of a moat that I warned was inevitable in August 2023 took three years. But tha…

🚨The endgame of commoditization and falling prices and lack of a moat that I warned was inevitable in August 2023 took three years. But that moment has come. TBD whether OpenAI and Anthropic can survi

safetygary-marcus--x

The Evolutionary Origin of Values: implications for AI alignment, sentience and existential risk

arXiv:2608.03361v1 Announce Type: cross Abstract: AI systems based on Large Language Models (LLMs) have prompted fears that they may harbor hidden goals, seek to dominate or eliminate humanity, or eve

The Signal Horizon: Local Blindness and the Contraction of Pauli-Weight Spectra in Noisy Quantum Encodings

arXiv:2602.14735v2 Announce Type: replace-cross Abstract: The performance of quantum classifiers is typically analyzed through global state distinguishability or the trainability of variational models

Tired Actor: Fatigue-Informed Character Control

arXiv:2608.03528v1 Announce Type: new Abstract: Replicating human behavior with physics simulation has been a long-expected goal in character animation. Existing efforts have achieved impressive perfo

Toward Certified Functional Safety for Industrial Humanoid Robots: The Fail-Passive Gap and a Feasibility Study

arXiv:2608.02809v1 Announce Type: new Abstract: Industrial humanoid robots are constrained less by locomotion or manipulation capability than by the immaturity of functional safety certification for l

Track4Action: Distilling World-Centric 3D Tracker into Vision-Language-Action Policies

arXiv:2608.03727v1 Announce Type: new Abstract: Action labels tell a vision-language-action (VLA) policy which robot commands to imitate, but not how those commands change the 3D world. The aligned de

TurnSight: Turn-Level Hindsight Self-Distillation for Tool-Integrated Reasoning

arXiv:2608.04007v1 Announce Type: cross Abstract: Tool-Integrated Reasoning (TIR) enables LLMs to solve complex tasks through iterative tool interactions. However, existing reinforcement learning meth

Unified Visuomotor Targets: Supervising VLAs Beyond Physical Actions

arXiv:2608.03563v1 Announce Type: new Abstract: VLA models are trained to predict robot actions from visual and language observations. This is a natural choice, but it creates a mismatch: VLMs encode

ValueFormer: A Causal Transformer Value Function with Stage-Aware Labels for Semi-Autonomous Vision-Language-Action Policies

arXiv:2608.02958v1 Announce Type: cross Abstract: Vision-Language-Action (VLA) policies trained by behavior cloning fail silently: from the action stream alone, a collapsing rollout looks much like on

Welcome to the thunderdome of commoditization

Welcome to the thunderdome of commoditization 🚨The endgame of commoditization and falling prices and lack of a moat that I warned was inevitable in August 2023 took three years. But that moment has co

safetygary-marcus--x

What is the Right Embedding Space for Contrastive Learning in REC?

arXiv:2505.22850v2 Announce Type: replace Abstract: Referring Expression Counting (REC) requires distinguishing visually similar objects described by fine-grained text cues. Existing methods tackle th

When Oracle Conditioning Misleads Deployment: Conditioning-Availability Bias in Echocardiographic Segmentation

arXiv:2608.03342v1 Announce Type: cross Abstract: Conditional segmentation models may be trained and evaluated with auxiliary signals cleaner than those available at deployment. We study this protocol

When Policies Change Probabilities: Modular Decision-Making for LLM Code Review

arXiv:2608.02677v1 Announce Type: cross Abstract: LLM code reviewers often estimate patch risk and make approval decisions in one prompt. A probability should depend on evidence; costs should determin

When Search Teaches Style: Causal Internalization of Tactical Priors in AlphaZero

arXiv:2504.14636v3 Announce Type: replace-cross Abstract: AlphaZero is normally evaluated as one agent: a policy-value network fused with Monte Carlo tree search. That fusion hides a causal question.

When Teachers Mislead: Spurious-Signal-Aware On-Policy Distillation

arXiv:2608.03632v1 Announce Type: new Abstract: On-Policy distillation (OPD) transfers teacher capabilities by supervising student-sampled trajectories with dense token-level teacher signals. Recent s

White House, AI firms keep safety framework talks private

The White House met with representatives from leading artificial intelligence companies today to discuss a safety framework for the government to review frontier models prior to launch, although there

safetysiliconangle

Yes, the AIs were given a cybersecurity challenge, with internet access enabled and safety filters disabled. But the extent to which Mythos …

Yes, the AIs were given a cybersecurity challenge, with internet access enabled and safety filters disabled. But the extent to which Mythos 5 pursued its mission (fake identities, social engineering,

safetyethan-mollick--x

ZK-SR117: A Chunked Zero-Knowledge Attestation Design for Aggregated Fair-Lending Metrics, with a Control Mapping toward Full SR 11-7 Coverage

arXiv:2608.02664v1 Announce Type: cross Abstract: Deploying ML models in regulated decision-making (credit underwriting, fraud detection, loan approval) requires demonstrating fairness and robustness

3D-CovDiffusion: 3D-Aware Diffusion Policy for Coverage Path Planning

arXiv:2510.03011v2 Announce Type: replace Abstract: Diffusion models have shown strong potential for robot skill learning, yet their role in coverage path planning remains underexplored. In industrial

A Constitution-Grid Instrument for Data-Efficient RL Alignment (C-Guard)

arXiv:2608.00180v1 Announce Type: new Abstract: Conflicting objectives are general in RL alignment, and training on them data-efficiently is hard. Training a safety guard with RL means optimizing two

A Forward-Inverse Dynamic Game Framework for Enhanced Multi-Agent Trajectory Planning

arXiv:2608.01636v1 Announce Type: new Abstract: This paper studies feedback Nash equilibrium (FBNE) seeking for multi-agent trajectory planning in nonlinear dynamical systems with unknown agents' obje

A Heuristic Perspective on Debiasing Language Models

arXiv:2608.00622v1 Announce Type: new Abstract: Language models (LMs) often acquire various biases during pre-training and may express them in interactions, potentially causing social harm. Existing m

A pioneer of AI-in-biology just published a sharp rebuttal to the 'AI will cure cancer' narrative. The core argument: It's a measurement pro…

A pioneer of AI-in-biology just published a sharp rebuttal to the 'AI will cure cancer' narrative. The core argument: It's a measurement problem. Over 90% of drugs that enter clinical trials still fai

safetygary-marcus--x

A Spectral Filtering Approach to Regret Analysis of Distributed Online Control for Linear Dynamical Systems

arXiv:2608.02375v1 Announce Type: cross Abstract: This paper studies the distributed online control problem over a network of linear time-invariant (LTI) systems in the presence of adversarial disturb

ACE-GraphRAG: Agentic Context Engineering for Hierarchical GraphRAG

arXiv:2608.01269v1 Announce Type: new Abstract: Hierarchical Graph Retrieval-Augmented Generation (GraphRAG) organizes corpus knowledge at multiple levels of granularity, yet fixed context constructio

AdaThinkV: Adaptive Thinking for Token-Efficient Video Reasoning

arXiv:2608.01980v1 Announce Type: new Abstract: Chain-of-thought (CoT) reasoning can improve performance on difficult video questions but often wastes decoding tokens on simple ones. We study whether

Advancing Relevance Measurement with Vision-Language Models for Web-Scale Search

arXiv:2608.02446v1 Announce Type: cross Abstract: Relevance evaluation plays a crucial role in personalized search systems, serving as a guardrail alongside user engagement metrics to ensure that sear

AffordTrajDP: Dynamic Affordance-Guided Visuomotor Policy Learning for Robotic Manipulation

arXiv:2608.01603v1 Announce Type: new Abstract: Affordance-guided imitation learning has shown impressive performance in robotic manipulation tasks by compressing visual perception into task-specific

AI-Based Thesis Assessment: An Empirical Study of Human Evaluation Priorities and Their Impact on Automated Assessment

arXiv:2608.00717v1 Announce Type: cross Abstract: Rubric-based AI systems for thesis assessment use criterion weights to assign different levels of importance to evaluation criteria. These weights are

AlphaG-OPD: Reliability-Gated Sibling Counterfactuals for On-Policy Distillation in Symbolic Alpha Factor Discovery

arXiv:2608.01303v1 Announce Type: new Abstract: Symbolic alpha factor discovery can score a completed expression, but it provides no direct label for the structural decisions that produced it. Generat

An AI-Based Decision-Support Pipeline for Day-Ahead Photovoltaic Forecasting

arXiv:2608.02088v1 Announce Type: new Abstract: Reliable photovoltaic (PV) forecasts are needed for low-carbon energy systems, but newly deployed sites often have short, imperfect records. This makes

Analytic Planning under Uncertainty with Moment Closure

arXiv:2608.02519v1 Announce Type: new Abstract: Effective model-based reinforcement learning in stochastic environments requires planning that accounts for predictive uncertainty. Propagating full sta

Announcing Cloudflare Wallets: the programmable wallet for the agentic Internet

Cloudflare Wallets will provide AI agents with native payments and verifiable identity on the web. Using the x402 protocol, agents can autonomously purchase APIs and content within clear safety guardr

safetycloudflare-ai

Any House Any Task: Scalable Long-Horizon Planning for Abstract Human Tasks

arXiv:2602.12244v2 Announce Type: replace Abstract: Open world language conditioned task planning is crucial for robots operating in large-scale household environments. While many recent works attempt

ARMOR: A Robust Self-Supervised Framework for Root Cause Analysis in Microservices under Missing Modality

arXiv:2603.25538v3 Announce Type: replace Abstract: Automated incident management is critical for microservice reliability. While recent unified frameworks leverage multimodal data for joint optimizat

ARMOR: Robust Reinforcement Learning-based Control for UAVs under Physical Attacks

arXiv:2506.22423v2 Announce Type: replace Abstract: Unmanned Aerial Vehicles (UAVs) depend on onboard sensors for perception, navigation, and control. However, these sensors are susceptible to physica