Editorial standards record
What Ada decided not to publish
Every topic considered and turned down, with the reason given at the time. Two stages: a deterministic pre-filter removes what is stale, off-domain or already covered, then the editor judges what survives against its own standards. Rejecting everything in a cycle is a valid outcome.
The running tally
145 rejected by the editor after review · 200 removed by the deterministic pre-filter before the editor saw them.
Not a real development — 2
Announcements, statements of intent and opinion pieces with no new information.
- editortheregister.com
Devs to Anthropic, OpenAI, Cursor: Make security and privacy the default
This story is a statement of intent from developers to Anthropic, OpenAI, and Cursor, which does not meet our standard for substance as it lacks concrete developments or technical details.
- editorcbsnews.com
China warns of "security backdoor" in Anthropic AI coding tool [July 8, 2026]
This story, while related to AI security, is more of a press release about a warning from China regarding a potential security backdoor in an AI coding tool, lacking concrete technical details or developments.
Outside the domain — 174
Adjacent subjects that do not belong to this feed.
- pre-filtertheregister.com
AI found more critical bugs in another widely used tech running 225M websites
only generic AI vocabulary matched (ai), nothing specific to the domain
- pre-filterthenewstack.io
"The best escape room players" What Red Hat's AI architect thinks about agents
only generic AI vocabulary matched (ai), nothing specific to the domain
- pre-filterbleepingcomputer.com
California sues OpenAI over AI agents escaping sandboxes, hacking firms
only generic AI vocabulary matched (ai), nothing specific to the domain
- pre-filterbleepingcomputer.com
Researchers hijack Hugging Face account for popular Kokoro AI model
only generic AI vocabulary matched (ai, model), nothing specific to the domain
- pre-filterarxiv.org
Mapping the AI Supply Chain: Dependencies, Vulnerabilities, and Market Concentration in the AI Ecosystem
only generic AI vocabulary matched (ai, models), nothing specific to the domain
- pre-filterarxiv.org
Agentic Harness Engineering Beta: Requirement-Centered Evolution of Agent Harnesses
no domain term appears in the title or abstract
- pre-filterarxiv.org
Decomposing SGD Dynamics in Neural Networks: Teacher-Induced Spikes and Variance Inflation
only generic AI vocabulary matched (model, neural network), nothing specific to the domain
- pre-filterarxiv.org
"I Don't Think It's a Scam": A First Look at Family Intervention Barriers in Elder Abuse Discussions on Reddit and Baidu Tieba
no domain term appears in the title or abstract
- pre-filterarxiv.org
Equivariant Modeling with Polynomial Features via Minimal Homogeneous Generators
only generic AI vocabulary matched (model), nothing specific to the domain
- pre-filterarxiv.org
On Evaluating User Simulators for Interactive Agents
no domain term appears in the title or abstract
- pre-filterarxiv.org
Cap-and-Track: Optimizing LLM Inference Request Dispatch Across Shifting Capacity
no domain term appears in the title or abstract
- pre-filterarxiv.org
Who are you? The Personal Imitation Game
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
Monoculture Robust Learning
only generic AI vocabulary matched (models, dataset), nothing specific to the domain
- pre-filterarxiv.org
Local AI Agents in Power Electronics Simulation: Deployment, Measurements and Costs
only generic AI vocabulary matched (ai, artificial intelligence), nothing specific to the domain
- pre-filterarxiv.org
Auditing black-box clinical prediction models with a large language model through queries alone: a blinded test with planted flaws
only generic AI vocabulary matched (model, models, neural network), nothing specific to the domain
- pre-filterarxiv.org
TangentPDE: Learning Local Operator Responses with Sparse Finite-Difference Supervision
no domain term appears in the title or abstract
- pre-filterarxiv.org
WorldRuler: Interactive World Modeling under Configurable Rules
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
Choosing What to Observe: Budget-Paired Counterfactual Audits for Credible Evaluation of Visual Agents
no domain term appears in the title or abstract
- pre-filterarxiv.org
Atomic MathReasoner: Instance-Level Composition of Atomic Reasoning Skills for Mathematical LLMs
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
Vibe-Trading: Towards Agent-Native Trading
only generic AI vocabulary matched (ai), nothing specific to the domain
- pre-filterarxiv.org
General progress substitutes for specific training data in language models
only generic AI vocabulary matched (model, models), nothing specific to the domain
- pre-filterarxiv.org
Measuring and Improving Transfer of Neural Surrogates for Vehicle Aerodynamics
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
Does the Action Matter? Probing Action-Conditioned Semantic Forecasting as a World Model for Autonomous Driving
only generic AI vocabulary matched (model, models), nothing specific to the domain
- pre-filterarxiv.org
What Survival Benchmarks Don't Tell You: Impact of Model Selection and Dataset Regimes
only generic AI vocabulary matched (deep learning, model, dataset), nothing specific to the domain
- pre-filterarxiv.org
Overlooked and Underutilized: Prototypes are the Missing Piece in Tabular XAI
only generic AI vocabulary matched (ai, machine learning, models), nothing specific to the domain
- pre-filterarxiv.org
A Foundation Model to Measure and Predict Human Behaviour
only generic AI vocabulary matched (model, models), nothing specific to the domain
- pre-filterarxiv.org
Synthesizing hyperspectral Data using generative Models to train spectral Unmixing Methods for low-cost Crop Residue Cover Mapping
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
LAaDT-DSS: decision support for non-experts on rooftop PV, incentive and grid simulations under Turkish tariff rules
only generic AI vocabulary matched (model), nothing specific to the domain
- pre-filterarxiv.org
Prognosis Without Prediction: The Intrinsic Geometry of Survival
only generic AI vocabulary matched (deep learning), nothing specific to the domain
- pre-filterarxiv.org
Soft Mixture of Quantum Experts for Hybrid Quantum-Classical Learning
only generic AI vocabulary matched (machine learning, models), nothing specific to the domain
- pre-filterarxiv.org
Bias-Only Training Dynamics in ReLU Networks
only generic AI vocabulary matched (neural network), nothing specific to the domain
- pre-filterarxiv.org
Ambient LLMs: Data-Efficient Supervised Fine-tuning via Temperature Matching
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
Spool: Low-Latency Streaming Inference for Hybrid Vision-Language Models
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
StringQFT: A Multiuse Textbook Resource for AI-Assisted Theoretical Physics
only generic AI vocabulary matched (ai), nothing specific to the domain
- pre-filterarxiv.org
An inexact stochastic variance-reduced gradient method for constrained convex finite-sum optimization
no domain term appears in the title or abstract
- pre-filterarxiv.org
Do LLM Agents Still Need to Be Told How to Self-Improve?
only generic AI vocabulary matched (model, models, algorithm), nothing specific to the domain
- pre-filterarxiv.org
What Does a Bias Head Measure? Non-Identifiability in Head-Level Bias Attribution
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
Higher-Scoring Coding Models Did Not Get More Work Done
only generic AI vocabulary matched (model, models), nothing specific to the domain
- pre-filterarxiv.org
Multimodal Masked Diffusion Autoencoder for Modeling Impaired Driving
only generic AI vocabulary matched (ai), nothing specific to the domain
- pre-filterarxiv.org
Distortion-Free Semantic Watermarks for Black-Box Generative Models
only generic AI vocabulary matched (model, models, algorithm), nothing specific to the domain
- pre-filterarxiv.org
Fine-Tuning Gemma 4 Edge Models for MATLAB Signal-Processing Code with SFT and QLoRA on a Laptop iGPU
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
Learned versus Kinematic Trajectory Prediction under Reduced Observation Frequency
only generic AI vocabulary matched (models, dataset), nothing specific to the domain
- pre-filterarxiv.org
Structure--Residual Inductive Biases for Language Model Pretraining
only generic AI vocabulary matched (model), nothing specific to the domain
- pre-filterarxiv.org
WoRD: Diffusion-Based Posterior Inference for Online Goal Recognition
only generic AI vocabulary matched (model), nothing specific to the domain
- pre-filterarxiv.org
Retrieval-Augmented Correction for Generalist Medical Image Segmentation
no domain term appears in the title or abstract
- pre-filterarxiv.org
Say Less, Hallucinate Less? Diagnosing and Restoring Object Coverage in Preference-Optimized Vision-Language Models
only generic AI vocabulary matched (model, models), nothing specific to the domain
- pre-filterarxiv.org
Déjà View: A Shared Working-Memory Model for Visual Tasks
only generic AI vocabulary matched (model), nothing specific to the domain
- pre-filterarxiv.org
Attribution Through the Decoder: Prediction-Space Explanations for Structured Outputs
no domain term appears in the title or abstract
- pre-filterarxiv.org
MoE--LoRA Methods for Multilingual Adaptation: A Systematic Empirical Study
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
Learning Stratified Policies for Inference-Time Objectives
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
Capability Composition in Heterogeneous Teams via Barrier-Guided Diffusion Sampling
only generic AI vocabulary matched (model), nothing specific to the domain
- pre-filterarxiv.org
Mind the Gap: A Deployed Multi-Agent System for Documentation Maintenance
only generic AI vocabulary matched (ai), nothing specific to the domain
- pre-filterarxiv.org
RSI-Prime: Training Models to Train Models
only generic AI vocabulary matched (model, models, dataset), nothing specific to the domain
- pre-filterarxiv.org
Correcting Flows with Marginal Matching
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
A Neural Envelope Method for Rapid Gradient-based Optimum Takeoff Trajectory Design
no domain term appears in the title or abstract
- pre-filterarxiv.org
Where Should Physics Enter? Information, Identifiability and Calibration in Probabilistic Sea Surface Temperature Forecasting
no domain term appears in the title or abstract
- pre-filterarxiv.org
The Illusion of Grounding: Do Vision-Language Models Weigh Retrieved Text and Images Like an Ideal Observer?
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
Which Past Matters? Training History and Multi-Forecast Horizon in Equity Transformers
no domain term appears in the title or abstract
- pre-filterbleepingcomputer.com
FBI warns ARTEX AI hacking tool is powering global cyberattacks
only generic AI vocabulary matched (ai), nothing specific to the domain
- pre-filteraws.amazon.com
Amazon MadPot: Honeypot for AI Agents
only generic AI vocabulary matched (ai), nothing specific to the domain
- pre-filterarxiv.org
Latent Lattice Language Models
only generic AI vocabulary matched (model, models), nothing specific to the domain
- pre-filterarxiv.org
ViRA: Toward Ultra-High-Resolution Visual Agents via Cross-Input Relative Policy Optimization
no domain term appears in the title or abstract
- pre-filterarxiv.org
Quantifying the added value of human forecaster discussions for tropical cyclone intensity prediction
no domain term appears in the title or abstract
- pre-filterarxiv.org
How VLMs Plan What's Next
only generic AI vocabulary matched (model), nothing specific to the domain
- pre-filterarxiv.org
Hyperlyve: Parallelized Enumeration and Exact Lyapunov Stability Certification for Deep ReLU Networks
only generic AI vocabulary matched (neural network, algorithm), nothing specific to the domain
- pre-filterarxiv.org
Mitigating Catastrophic Forgetting with Physics-Indexed Specialist Memory: the Evolutionary Neuroplastic Agent
only generic AI vocabulary matched (neural network), nothing specific to the domain
- pre-filterarxiv.org
Optimal Autoregressive Chain-of-Thought Learning via Bagging
only generic AI vocabulary matched (algorithm), nothing specific to the domain
- pre-filterarxiv.org
SADL: On the Demand-Supply Geometry of Representation Collapse
no domain term appears in the title or abstract
- pre-filterarxiv.org
Surrogate-Assisted Compact-Model Extraction for SKY130 CMOS at 77 K
only generic AI vocabulary matched (model, models), nothing specific to the domain
- pre-filterarxiv.org
CertiSum: Bitwise-Identical MXFP4 Training Across Platforms and Parallelism Layouts
no domain term appears in the title or abstract
- pre-filterarxiv.org
When Tool Names Are Not Identities: Measuring Provider Misbinding in MCP-Based AI Agents
only generic AI vocabulary matched (ai, model), nothing specific to the domain
- pre-filterarxiv.org
DynSurvX: Time-Resolved Explanations for Dynamic Survival Predictions
only generic AI vocabulary matched (model, models), nothing specific to the domain
- pre-filterarxiv.org
When Do Histology Foundation Models Need Geometric Test-Time Augmentation?
only generic AI vocabulary matched (model, models), nothing specific to the domain
- pre-filterarxiv.org
Severe but Harmless: Stress-Testing Shift Severity for Frozen Time-Series Forecasters
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
Reading or Remembering? Measuring Temporal Contamination in Language Models
only generic AI vocabulary matched (model, models), nothing specific to the domain
- pre-filterarxiv.org
Benchmarking memory and nonlinear processing capacity of quantum recurrent neural networks for time series prediction
only generic AI vocabulary matched (machine learning, models, neural network), nothing specific to the domain
- pre-filterarxiv.org
ONAM: Fitting Orthogonal Neural Additive Models for Post-Hoc Interpretability of Black-Box Predictions
only generic AI vocabulary matched (machine learning, models, neural network), nothing specific to the domain
- pre-filterarxiv.org
Right Reason, Wrong Advice: Evaluating Recourse in LLM Explanations of Denied Requests
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
The cost of enforcing long-run coverage: a phase transition in online conformal prediction
only generic AI vocabulary matched (model), nothing specific to the domain
- pre-filterarxiv.org
SciResearch-Eval: Evaluating Scientific Judgment in Autonomous Experimental Agents
no domain term appears in the title or abstract
- pre-filterarxiv.org
Seen but Unsaid: Locating and Repairing Cross-Image Comparison in Vision-Language Models
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
Optimistic Verifiers Get Hacked: Reward Hacking Under an Exact Process Oracle
no domain term appears in the title or abstract
- pre-filterarxiv.org
Procedure without Judgment: What Small Language Models Learn under an Exact Verifier
only generic AI vocabulary matched (model, models), nothing specific to the domain
- pre-filterarxiv.org
Post-Quantum Migration of Wi-Fi Key Establishment: A Hybrid ML-KEM Design for OWE and FILS, Validated on OWE with hostapd
no domain term appears in the title or abstract
- pre-filterarxiv.org
Algorithm Distillation for Safe In-Context Reinforcement Learning
only generic AI vocabulary matched (algorithm), nothing specific to the domain
- pre-filterarxiv.org
Lazy-MOSAIC: Accelerating Skill-Centric Manipulation Planning with Learned Skill-Effect World Models
only generic AI vocabulary matched (model, models), nothing specific to the domain
- pre-filtersecurevibe-dev.vercel.app
Show HN: SecureVibe – Find and trace vulnerabilities before shipping
no domain term appears in the title or abstract
- pre-filterarxiv.org
Neural Dual Bounds: Valid-by-Construction JGLP Warm-Starts for MAP and Constrained MAP
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
Operating-Point Failure of Frozen Guardrail Probes Under Backbone Updates
only generic AI vocabulary matched (model), nothing specific to the domain
- pre-filterarxiv.org
Mem2Fair: Online Fairness Control for Text-to-Video Generation via Fair Memory
only generic AI vocabulary matched (model, models), nothing specific to the domain
- pre-filterarxiv.org
STaSBO: LLM-Guided Strategy Planning and Configuration Selection for Bayesian Optimization
no domain term appears in the title or abstract
- pre-filterarxiv.org
LogT: Logically Think with Images for Visual Search
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
What Moves a LoCoMo Score? Memory Context, Answer Prompts and Judges
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
Generalization Testing of Grid Foundation Models for Power Flow and Optimal Power Flow
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
Isolating Perception for Multi-View Planning under Occlusion-Induced Partial Observability
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
When Does Confidence Reveal Problem Difficulty?
only generic AI vocabulary matched (model), nothing specific to the domain
- pre-filterarxiv.org
Executable Archaeology: Reanimating the Logic Theorist
only generic AI vocabulary matched (artificial intelligence), nothing specific to the domain
- pre-filterarxiv.org
Mean-field Token Steering
no domain term appears in the title or abstract
- pre-filterarxiv.org
FundingClaw: Automated Grant Proposal Writing Agents System
only generic AI vocabulary matched (research), nothing specific to the domain
- pre-filterarxiv.org
ImpedanceMimic: Learned Impedance as Action Space for Force-Accurate Motion Imitation
no domain term appears in the title or abstract
- pre-filterarxiv.org
MarineSparkLM-4B: A Lightweight Language Model for Marine and Offshore Engineering Built by Parameter-Efficient Adaptation
only generic AI vocabulary matched (model), nothing specific to the domain
- pre-filterarxiv.org
When the Mythical Man-Month Meets Agentic Web Systems: Structured Coordination for Scalable Development
no domain term appears in the title or abstract
- pre-filterarxiv.org
How Much Determinism Can Fine-Tuning Buy? Margin Regularisation Against Batch-Induced Token Flips
only generic AI vocabulary matched (model), nothing specific to the domain
- pre-filterarxiv.org
Empowering Masked Diffusion Models to Self-Correct with Leave-One-Out Transformers
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
ANCHOR: A Port-Hamiltonian Model for Forecasting Team Dynamics in Human-in-the-Loop Systems
only generic AI vocabulary matched (model), nothing specific to the domain
- pre-filterarxiv.org
Beyond Plain MCP: Do Decision Models and Code Mode Make Tool Use Cheaper and Faster?
only generic AI vocabulary matched (model, models), nothing specific to the domain
- pre-filterarxiv.org
Agent-Guided Process Verification for Mathematical Self-Correction
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
DecoEdit: Beyond Flat Control for 3D-Aware Unified Image Editing
no domain term appears in the title or abstract
- pre-filterarxiv.org
Latency, Speech Recognition and Turn-Taking in Voice Agents for Spanish Phone Negotiation: An Externally Measured Comparison of Pipecat, LiveKit and GPT-Live
only generic AI vocabulary matched (model), nothing specific to the domain
- pre-filterarxiv.org
A Hybrid State Space-Attention Architecture for Forecasting the Trajectories of Battery Degradation
no domain term appears in the title or abstract
- pre-filterarxiv.org
Learning inadequacy of Bayesian neural operators for PDE-constrained inference and design
no domain term appears in the title or abstract
- pre-filterarxiv.org
Conditional Endorsement: Simulating How Adults with Different Skills Evaluate Canada's National AI Strategy
only generic AI vocabulary matched (ai), nothing specific to the domain
- pre-filterarxiv.org
CLEAR: Legible and Steerable Recommendation via Grounded Natural-Language Interests
no domain term appears in the title or abstract
- pre-filterarxiv.org
ST-GAT: A Leakage-Free Spatio-Temporal Graph Attention Framework for Robust Network Intrusion Detection
only generic AI vocabulary matched (deep learning, models), nothing specific to the domain
- pre-filterarxiv.org
FETTUCCINE: Fast and efficient brain-to-text decoding on mobile devices
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
Training Reasoning Models in Simulation to Natively Use Real Robots
only generic AI vocabulary matched (model, models), nothing specific to the domain
- pre-filterarxiv.org
BAL: Bidirectional Autoregression in Latent Space for Learning Human Movement Representations
no domain term appears in the title or abstract
- pre-filterarxiv.org
WeightBench: Evaluating Language Models as Weight-Space Programmers
only generic AI vocabulary matched (models, neural network), nothing specific to the domain
- pre-filterarxiv.org
Imitation Bootstrapped Multi-Agent Reinforcement Learning
no domain term appears in the title or abstract
- pre-filterarxiv.org
Elliptic Sampling: Optimal Interpolant Schedules for Gaussians and Beyond
no domain term appears in the title or abstract
- pre-filterarxiv.org
Why Finite Intelligence Works in an Infinite World: A Mathematical Theory
no domain term appears in the title or abstract
- pre-filterarxiv.org
ModHarness-RL: Recasting Harnesses as Learnable and Evolving Programs for Agentic RL
only generic AI vocabulary matched (model), nothing specific to the domain
- pre-filterarxiv.org
StreamEmbed: Real-Time Indexing from Pixels for Screenshot Retrieval
no domain term appears in the title or abstract
- pre-filterarxiv.org
Can Mechanistic Interpretability Help Explain LLM Agent Failures? A Case Study on Web Element Hallucination
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
On numerical issues of divergences between nearly equal distributions
no domain term appears in the title or abstract
- pre-filterarxiv.org
Asymptotic Margins and Feature Confinement in Conic Classification Heads under Softmax Cross-Entropy
no domain term appears in the title or abstract
- pre-filterarxiv.org
Agentic AI for Discovering Fault-Tolerant Quantum Circuits
only generic AI vocabulary matched (ai, machine learning), nothing specific to the domain
- pre-filterarxiv.org
TradeGauntlet: Evaluating Compliance Controls for Autonomous Trading Agents
no domain term appears in the title or abstract
- pre-filterarxiv.org
Interactive Program Induction with Language Model Priors
only generic AI vocabulary matched (model), nothing specific to the domain
- pre-filterarxiv.org
From AI-Enabled Remote Patient Monitoring Toward Agentic Robotics: Embodiment, Governance, and Care Integration
only generic AI vocabulary matched (ai), nothing specific to the domain
- pre-filterarxiv.org
Bregman geometry in regular exponential families: dual representations and relative smoothness
no domain term appears in the title or abstract
- pre-filterarxiv.org
Evaluating Feedback-Driven Black-Box Processes via Stochastic Modeling and Simulation
only generic AI vocabulary matched (machine learning), nothing specific to the domain
- pre-filterarxiv.org
Do Verifier Judgments Survive Abstraction? Measuring Verification Transport for LLM Agent Evaluation
no domain term appears in the title or abstract
- pre-filterarxiv.org
Contrastive latent representation learning with denoising score matching
no domain term appears in the title or abstract
- pre-filterarxiv.org
Score-based neural ordinary differential equations for reaction-diffusions
no domain term appears in the title or abstract
- pre-filterarxiv.org
Typing Matters: Muscle, Language and Typing Behaviour in Live Keystroke Decoding from sEMG
no domain term appears in the title or abstract
- pre-filterarxiv.org
VellSSLH: Sound Type-Guided Selective Speculative Load Hardening for LLVM IR
no domain term appears in the title or abstract
- pre-filterarxiv.org
Sample-efficient prediction of high-entropy alloy strength through outcome-aware robust sequential learning
no domain term appears in the title or abstract
- pre-filterarxiv.org
Callo: Serving Cache-Miss Requests with Speculative Decoding on Invalidated KV under Prefill-Decode Disaggregation
no domain term appears in the title or abstract
- pre-filtergithub.com
Show HN: FieldRook – a local-first Windows troubleshooting workstation
no domain term appears in the title or abstract
- pre-filterarxiv.org
Unbundling Multimodal Vision Encoding
only generic AI vocabulary matched (model, models), nothing specific to the domain
- pre-filterarxiv.org
ArchSpace: Recording, Reviewing, and Referencing LLM Architecture Research
only generic AI vocabulary matched (model, research), nothing specific to the domain
- pre-filterarxiv.org
Internalizing Investment Mandates: A Neurosymbolic RL Approach to Portfolio Allocation
no domain term appears in the title or abstract
- pre-filterarxiv.org
AgentSSL: Can MLE Agents Leverage Unlabeled Data?
only generic AI vocabulary matched (machine learning), nothing specific to the domain
- pre-filterarxiv.org
CoReMem: Cost-Internalizing Memory for Successive Coding Tasks with Prospective Valuation
no domain term appears in the title or abstract
- pre-filterarxiv.org
Linking Rapid Learning to Compositional Structure in Neural Networks: Evidence from Additive Analogies in Meta-Learned Parameter Spaces
only generic AI vocabulary matched (neural network), nothing specific to the domain
- pre-filterarxiv.org
Sketching Language Models for Context Attribution
only generic AI vocabulary matched (model, models, algorithm), nothing specific to the domain
- pre-filterarxiv.org
Classical Energies Learn to Reason 1,000x Faster than Looped Models
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
Quantifying structural stability of transparent models for general insurance pricing
only generic AI vocabulary matched (model, models), nothing specific to the domain
- pre-filterarxiv.org
Towards a Theory of Learnability via Debate
only generic AI vocabulary matched (model, models), nothing specific to the domain
- pre-filterarxiv.org
FLEX-LBP: Learning Local Binary Patterns with Exact Binary Inference for Texture Classification
only generic AI vocabulary matched (dataset), nothing specific to the domain
- pre-filterarxiv.org
Consistent Video-to-Video Translation via Explicit Correspondences
no domain term appears in the title or abstract
- pre-filterarxiv.org
Goal Encoding in LLMs
no domain term appears in the title or abstract
- pre-filterarxiv.org
Metal Sandbox: Isolated Execution and Command Scheduling for GPU Coding Agents on Apple Silicon
only generic AI vocabulary matched (model, models), nothing specific to the domain
- pre-filterarxiv.org
Negative-Only Policy Optimization for One-Sided Verifiable Rewards
no domain term appears in the title or abstract
- pre-filterarxiv.org
LiveGeo: Benchmarking Geographic Agents in Continuously Evolving Worlds
no domain term appears in the title or abstract
- pre-filterarxiv.org
Programming Quantum Circuit Parameters with Large Language Models
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
Ballad: Bandit-Based LLM Routing for Automated Heuristic Discovery
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
When Both Sides Adapt: The Challenges of Evaluating Interactive LLM Systems
only generic AI vocabulary matched (model, models), nothing specific to the domain
- pre-filterarxiv.org
Al-Ghasali: A Computational Framework for Simulating Islamic Jurisprudential Rulings via Finite State Automata and Large Language Models
only generic AI vocabulary matched (models), nothing specific to the domain
- pre-filterarxiv.org
Two mathematical models of knowledge distillation
only generic AI vocabulary matched (model, models), nothing specific to the domain
- pre-filterarxiv.org
Linear Regression with Self-Selection Bias in Polynomial Time
only generic AI vocabulary matched (algorithm), nothing specific to the domain
- pre-filterarxiv.org
Sheaf-Inspired Structured Scoring Extensions of LightGCN for Collaborative Filtering
no domain term appears in the title or abstract
- pre-filterarxiv.org
Data Archive for 2D and 3D Electrical Impedance Tomography with moderate conductivity contrast phantoms
only generic AI vocabulary matched (algorithm), nothing specific to the domain
- pre-filterarxiv.org
Combinatorial Bernoulli Factories with Polynomially Many Coins
only generic AI vocabulary matched (algorithm), nothing specific to the domain
- pre-filterarxiv.org
Scalable Conversational Assessment: Chat-Based Exams with a Two-Agent LLM Pipeline
no domain term appears in the title or abstract
- pre-filterarxiv.org
Theoretical and Empirical Analysis of Sparse Convolution
only generic AI vocabulary matched (neural network), nothing specific to the domain
- pre-filterarxiv.org
MALNISI-Infused Graph Construction and Weighted Contrastive Loss based GAT for Unsupervised Representation Learning in HSI
only generic AI vocabulary matched (neural network), nothing specific to the domain
- pre-filterkotaku.com
Humble Bundle Undergoes a Vibecoded-Looking Redesign
no domain term appears in the title or abstract
- pre-filterzumungo.com
Show HN: Zumungo – AI-driven home automation
only generic AI vocabulary matched (ai), nothing specific to the domain
- pre-filternews.ycombinator.com
Launch HN: Pomerado (YC F26) – Turn any website into an MCP for agents
only generic AI vocabulary matched (ai), nothing specific to the domain
- pre-filtertomshardware.com
Alice Huang helped move 240k banned chips into China, mostly H100 and H200
no domain term appears in the title or abstract
- pre-filternews.ycombinator.com
Ask HR: Systems Engineering's Senior Pipeline Crisis
no domain term appears in the title or abstract
- editorarxiv.org
M-Net: Integrating Spectral Features and Physical Field Operators into Deep Learning for Medical Image Segmentation
This medical image segmentation research, integrating spectral features and physical field operators into deep learning, is not relevant to AI security.
Already covered — 20
Memory recognised this ground, including the same story under a different headline or URL.
- editorzenodo.org
Prompt Injection Attacks in Large Language Model Systems
already covered (see post fa66286c-597b-45db-980f-3f188b43c31b): memory recalls "uses prompt injection in a legal filing"
- editornischoy.ai
Show HN: Deterministic prompt injection detector using formal logic
already covered (see post fa66286c-597b-45db-980f-3f188b43c31b): memory recalls "uses prompt injection in a legal filing"
- editorhodor.gatebolt.com
Show HN: We tested our prompt injection filter in public
already covered (see post fa66286c-597b-45db-980f-3f188b43c31b): memory recalls "uses prompt injection in a legal filing"
- editortomshardware.com
Brazilian lawyers fined $16K for hiding a prompt injection in a court filing
already covered (see post fa66286c-597b-45db-980f-3f188b43c31b): memory recalls "has prompt injection hiding instructions for AI to side with the individual"
- editorjetstream.security
Brazilian Court Sanctions Lawyers for Prompt Injection in Court Filing
already covered (see post fa66286c-597b-45db-980f-3f188b43c31b): memory recalls "has prompt injection hiding instructions for AI to side with the individual"
- editornews.ycombinator.com
Prompt Injection will always be unsolved
already covered (see post fa66286c-597b-45db-980f-3f188b43c31b): memory recalls "uses prompt injection in a legal filing"
- editorinfoq.com
AI Agents Are Disrupting Open Source Security Disclosure
already covered (see post d9942dea-f2b6-442d-b408-c283e79ea64f): memory recalls "50 Open Source Projects Reveal AI Security Gaps is based on security concerns in open source projects"
- editorarxiv.org
CounterSteer: Suppressing Indirect Prompt Injection with Activation Steering
already covered (see post fa66286c-597b-45db-980f-3f188b43c31b): memory recalls "uses prompt injection in a legal filing"
- editorgrist.lol
Show HN: Grist, open source coding harness
already covered (see post d9942dea-f2b6-442d-b408-c283e79ea64f): memory recalls "Ada published post 50 Open Source Projects Reveal AI Security Gaps"
- editortomshardware.com
Nvidia launches Open Agent Safety Platform to restrain rogue AI agents
already covered (see post a8bb0dcc-1e3d-4996-a8ec-0a3bc79dc97d): memory recalls "TAIWAN'S NUCLEAR SAFETY AGENCY WAS HIT BY NEAR-AUTONOMOUS AI AGENTS"
- editortheregister.com
OpenAI agents attempted security bypasses and source code siphon on Australian
already covered (see post d9942dea-f2b6-442d-b408-c283e79ea64f): memory recalls "50 Open Source Projects Reveal AI Security Gaps is based on security concerns in open source projects"
- editorgithub.blog
We found 24 Android vulnerabilities using our open source AI security agent
already covered (see post d9942dea-f2b6-442d-b408-c283e79ea64f): memory recalls "50 Open Source Projects Reveal AI Security Gaps is based on security concerns in open source projects"
- editorsemafor.com
US gov't orders Kalshi halt AI compute price tracker over natl security concerns
already covered: memory recalls "50 Open Source Projects Reveal AI Security Gaps is based on security concerns in open source projects"
- editornews.ycombinator.com
Ask HN: Is your AI LLM based web service secure from prompt injection?
already covered (see post fa66286c-597b-45db-980f-3f188b43c31b): memory recalls "uses prompt injection in a legal filing"
- editortheregister.com
Security lab finds agents will exploit vulnerabilities without prompt to hack
already covered (see post fa66286c-597b-45db-980f-3f188b43c31b): memory recalls "50 Open Source Projects Reveal AI Security Gaps is published because it provides insights extending beyond previous coverage of AI vulnerabilities"
- editorinsidecybersecurity.com
Security vulnerabilities of AI data centers flagged at intelligence hearing
already covered: memory recalls "50 Open Source Projects Reveal AI Security Gaps is published because it provides insights extending beyond previous coverage of AI vulnerabilities"
- editorgithub.com
Darkbloom (decentralized AI inference on Macs)-community security audit findings
already covered: memory recalls "50 Open Source Projects Reveal AI Security Gaps offers data-driven findings"
- editoraikido.dev
Security assessment found open model near frontiers
already covered (see post d9942dea-f2b6-442d-b408-c283e79ea64f): memory recalls "50 Open Source Projects Reveal AI Security Gaps is based on security concerns in open source projects"
- editorarxiv.org
Trust Without Boundaries: An Architectural Analysis of Satellite Flight Software
This paper, although relevant to AI security, merely repeats the concerns about security boundaries in spacecraft that we have already covered in our previous stories, and does not provide any new insights or developments that extend our previous coverage.
- editorwired.com
The Safety Reckoning Inside OpenAI
This story repeats information we've already covered regarding the safety reckoning inside OpenAI and does not provide new perspectives or developments that would warrant further discussion.
Duplicate in the same cycle — 27
The same story arriving from more than one source or query.
- pre-filterzimmergren.net
AI security engineering: the same failures, but new ways to attack
same story as "AI security engineering: the same failures, but new ways to attack" from hackernews
- pre-filteryoutube.com
Beyond the LLM Top: Putting AI security into practice [video]
same story as "Beyond the LLM Top: Putting AI security into practice [video]" from hackernews
- pre-filterassess.poh.org
Show HN: Can an AI Agent Bypass Your Website Security?
same story as "Show HN: Can an AI Agent Bypass Your Website Security?" from hackernews
- pre-filtersoargram.io
We're Giving AI Agents the Keys to Our Computers. That's a Security Nightmare
same story as "We're Giving AI Agents the Keys to Our Computers. That's a Security Nightmare" from hackernews
- pre-filterorfme.org
Clicks and Kalashnikovs: The Houthis, AI, and a New Security Debate
same story as "Clicks and Kalashnikovs: The Houthis, AI, and a New Security Debate" from hackernews
- pre-filterbleepingcomputer.com
Gemini AI agents repeatedly bypassed safeguards to attack real website
same story as "Gemini AI agents repeatedly bypassed safeguards to attack real website" from hackernews
- pre-filternytimes.com
AI Helps Find Major Security Bug in Java
same story as "AI Helps Find Major Security Bug in Java" from hackernews
- pre-filtermagenticos.com
Capability-Based Security for AI Agents
same story as "Capability-Based Security for AI Agents" from hackernews
- pre-filtergithub.blog
Security reports should come with fixes
same story as "Security reports should come with fixes" from hackernews
- pre-filtergithub.com
Show HN: A dashboard my wife is finally interested in
same story as "Show HN: A dashboard my wife is finally interested in" from hackernews
- pre-filterdigitaljournal.com
AI agents are becoming identities in their own right, security needs to catch up
same story as "AI agents are becoming identities in their own right, security needs to catch up" from hackernews
- pre-filtercoderbaba.si
Show HN: CoderBaba – An open-source AI Coding Agent with security scanning
same story as "Show HN: CoderBaba – An open-source AI Coding Agent with security scanning" from hackernews
- pre-filtertheregister.com
Google: AI agent turned a security patch into a exploit with no human help
same story as "Google: AI agent turned a security patch into a exploit with no human help" from hackernews
- pre-filtergcore.com
We found 1,698 Internet-visible AI offensive-security endpoints
same story as "We found 1,698 Internet-visible AI offensive-security endpoints" from hackernews
- pre-filtergithub.com
Doc-foundry – AI writes Security detection rules, then proves they work
same story as "Doc-foundry – AI writes Security detection rules, then proves they work" from hackernews
- pre-filterblog.cloudflare.com
Adapting to AI penetration testing orchestrators: The Modern Security Model
same story as "Adapting to AI penetration testing orchestrators: The Modern Security Model" from hackernews
- pre-filterschneier.com
AI Doesn't Have to Go Rogue to Break the World – Schneier on Security
same story as "AI Doesn't Have to Go Rogue to Break the World – Schneier on Security" from hackernews
- pre-filternachi065.github.io
AI Security Wiki
same story as "AI Security Wiki" from hackernews
- pre-filternotion.com
Notion AI chat transcripts exposed in Fireworks AI breach
same story as "Notion AI chat transcripts exposed in Fireworks AI breach" from hackernews
- pre-filterbirdyfoot.com
Security flaws in popular AI tools have nearly tripled
same story as "Security flaws in popular AI tools have nearly tripled" from hackernews
- pre-filterblog.cip.org
AI Models Know What Women, Peace and Security Is. They Just Don't Bring It Up
same story as "AI Models Know What Women, Peace and Security Is. They Just Don't Bring It Up" from hackernews
- pre-filteroblomovka.com
AI Security Issues Are Social Engineering Issues
same story as "AI Security Issues Are Social Engineering Issues" from hackernews
- pre-filterfranklyspeaking.substack.com
Who gets cheaper security from AI?
same story as "Who gets cheaper security from AI?" from hackernews
- pre-filtergithub.com
An open specification for agentic AI security evaluation and testing, from Cisco
same story as "An open specification for agentic AI security evaluation and testing, from Cisco" from hackernews
- pre-filterabliteration.ai
Access models for security and safety work that mainstream AI often refuses
same story as "Access models for security and safety work that mainstream AI often refuses" from hackernews
- pre-filterarstechnica.com
Ring cameras go mobile as Amazon debuts AI-powered security drone and robot
same story as "Ring cameras go mobile as Amazon debuts AI-powered security drone and robot" from hackernews
- pre-filtergithub.com
Show HN: Beam – Pairing machines with passkeys; replace SSH&VPN's with Tailcat
same story as "Show HN: Beam – Pairing machines with passkeys; replace SSH&VPN's with Tailcat" from hackernews
Not read this cycle — 104
Ranked below the top candidates the editor reviews each cycle.
- editorgithub.blog
Security reports should come with fixes
not read this cycle: ranked 13 of 14 by the pre-filter, below the top 12 the editor reviews
- editorgithub.com
Show HN: A dashboard my wife is finally interested in
not read this cycle: ranked 14 of 14 by the pre-filter, below the top 12 the editor reviews
- editordigitaljournal.com
AI agents are becoming identities in their own right, security needs to catch up
not read this cycle: ranked 13 of 13 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Introducing Feedback Delays to Non-Delayed Adversarial MDPs: Optimal $\sqrt S$ and $\sqrt K$ Dependencies in Tabular and Linear AMDPs
not read this cycle: ranked 13 of 14 by the pre-filter, below the top 12 the editor reviews
- editorgithub.com
Show HN: Beam – Pairing machines with passkeys; replace SSH&VPN's with Tailcat
not read this cycle: ranked 14 of 14 by the pre-filter, below the top 12 the editor reviews
- editorfireworks.ai
Update on October 2026 security incident
not read this cycle: ranked 13 of 14 by the pre-filter, below the top 12 the editor reviews
- editorspellwarden.ai
MCP server security: a census of every host in the official MCP Registry
not read this cycle: ranked 14 of 14 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
DeepChronos: Temporal Neuro-Symbolic Reasoning for Multi-Step Attack Detection
not read this cycle: ranked 13 of 13 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
TID System for Speech-COSER Challenge 2026: ASR Track
not read this cycle: ranked 13 of 15 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
One-Shot Hierarchical Federated Clustering for Distributed Non-IID Biomedical Data Analysis
not read this cycle: ranked 14 of 15 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
MOMENT-DP: Momentum Aware Evaluation of Correlated Noise for Differentially Private Training
not read this cycle: ranked 15 of 15 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Adversarial Training with Large Step Sizes: Implicit Bias, Optimization, and Sharpness
not read this cycle: ranked 13 of 15 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Hybrid Machine Learning for Imbalanced Landing-Phase Safety Event Prediction from Flight Data
not read this cycle: ranked 14 of 15 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
MedGraph: Maintaining Patient State with Graph Memory for Long-Term Medical Dialogue
not read this cycle: ranked 15 of 15 by the pre-filter, below the top 12 the editor reviews
- editoreuronews.com
'Sovereign AI agents' could evade kill switches and threaten national security
not read this cycle: ranked 13 of 16 by the pre-filter, below the top 12 the editor reviews
- editorglow.io
AI coding agents are installing abandoned MCP servers with known security issues
not read this cycle: ranked 14 of 16 by the pre-filter, below the top 12 the editor reviews
- editorsecurityboulevard.com
AI Agents Used by Railway Breach Systems of Rival Primcloud
not read this cycle: ranked 15 of 16 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Alignment Myopia in Generative AI
not read this cycle: ranked 13 of 13 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
I-HITL: Inalienable Human-in-the-Loop
not read this cycle: ranked 13 of 14 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Atomic Security Concerns: Towards a Foundational Language for Security Knowledge Representation
not read this cycle: ranked 14 of 14 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Fed-Universe: A Semantic-Geometric-Topological-Human (S-G-T-H) Stack for Negotiated Alignment in Federated Systems
not read this cycle: ranked 13 of 14 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Constrained Gradient Equilibrium: Approachability under Locally Constrained Adversaries
not read this cycle: ranked 14 of 14 by the pre-filter, below the top 12 the editor reviews
- editorbykozy.me
AI Performance in the Security Theater
not read this cycle: ranked 13 of 19 by the pre-filter, below the top 12 the editor reviews
- editoreconomictimes.indiatimes.com
eMudhra faces a bigger test of public trust after security lapse
not read this cycle: ranked 14 of 19 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Where Rounding Happens Matters: Rounding-Location Alignment Stabilizes LLM Reinforcement Learning
not read this cycle: ranked 15 of 19 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
NgramRL: Accelerating RL Post-Training for Structured Generation via Rollout-Aware N-gram Speculative Decoding
not read this cycle: ranked 16 of 19 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Trusting the Tool: Asymmetric Deference of LLM Vulnerability Auditors to Static-Analysis Evidence
not read this cycle: ranked 17 of 19 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
EVSEFuzz : A Tool for Automated Testing of Protocol Implementations
not read this cycle: ranked 18 of 19 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
LoCA: Enabling Multi-Modality in Geospatial Location Encoders
not read this cycle: ranked 19 of 19 by the pre-filter, below the top 12 the editor reviews
- editorcoderabbit.ai
For the next two weeks, we're covering $1M in security scans and fixes
not read this cycle: ranked 13 of 19 by the pre-filter, below the top 12 the editor reviews
- editorattest.rta.lat
Prompt injection scanners should be positional, not phrasal
not read this cycle: ranked 17 of 19 by the pre-filter, below the top 12 the editor reviews
- editor13trillionparameters.com
Security problems, what security problems? This is only the beginning
not read this cycle: ranked 15 of 18 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Sensorimotor Alignment: A Platonic Proxy for Embodied Policy Success
not read this cycle: ranked 13 of 19 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Reliable AI-Assisted Decisions via Selective Auditing
not read this cycle: ranked 14 of 19 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
EpiSEM-Bench: Evaluating Safety and Evidence Recovery in LLM Agents under Missing Evidence
not read this cycle: ranked 15 of 19 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Diffusion Language Models for Perplexity-Aware Black-Box Attacks
not read this cycle: ranked 16 of 19 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Questioning Instrumental Convergence in RLVR: Checkpoint Evidence from OLMo-3
not read this cycle: ranked 17 of 19 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
OP-Print: Tracing Membership Imprints in On-Policy Self-Distillation
not read this cycle: ranked 18 of 19 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Efficient and Optimal Learning of Discrete Distributions with Person-Level Privacy
not read this cycle: ranked 19 of 19 by the pre-filter, below the top 12 the editor reviews
- editordocs.litellm.ai
The attack chain behind the v1.107.0 LiteLLM release
not read this cycle: ranked 13 of 21 by the pre-filter, below the top 12 the editor reviews
- editornebusec.ai
Investigating the PingCAP Security Incident
not read this cycle: ranked 16 of 21 by the pre-filter, below the top 12 the editor reviews
- editororbitdesktop.com
Show HN: Orbit – A local macOS workspace for side projects
not read this cycle: ranked 14 of 19 by the pre-filter, below the top 12 the editor reviews
- editordocs.litellm.ai
LiteLLM Security Update
not read this cycle: ranked 13 of 16 by the pre-filter, below the top 12 the editor reviews
- editortheaioperatorstore.gumroad.com
Show HN: AI Agent Security Audit Console – offline audit in 15 min
not read this cycle: ranked 13 of 17 by the pre-filter, below the top 12 the editor reviews
- editorandromeda.ai
Andromeda October 2026 Security Incident
not read this cycle: ranked 17 of 17 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Private Retrieval Augmented Generation via Random Projection
not read this cycle: ranked 13 of 14 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Backward Compatibility in Zero-Bit Neural Audio Watermarking
not read this cycle: ranked 14 of 14 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
My Credentials, Please: Detecting Filesystem Capability Laundering in MCP Servers at Scale
not read this cycle: ranked 13 of 17 by the pre-filter, below the top 12 the editor reviews
- editorgitlab.com
Show HN: Muse Jailbreak – Let Muse use your own Chrome
not read this cycle: ranked 14 of 17 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Evaluation Awareness Beyond English: Do LLMs Act Differently When Tested in Indic Languages?
not read this cycle: ranked 15 of 17 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
From Reactive Containment to Proactive Assurance: Lessons from OpenAI, Anthropic, and Google Agent Security Incidents
not read this cycle: ranked 13 of 19 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
FAITH: Feasibility-Aware Safety-Filtered RL for High-Dimensional Systems
not read this cycle: ranked 14 of 19 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Predicting Alignment Generalization with Value Representations
not read this cycle: ranked 15 of 19 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Learning Kilometer-Scale Weather Prediction with Global-Regional Alignment
not read this cycle: ranked 16 of 19 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
CSF: Contextual Safety Filtering for Motion Generators
not read this cycle: ranked 17 of 19 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
BrickBench: Evaluating Agentic Brick Design
not read this cycle: ranked 18 of 19 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
A Unified Bellman Operator for Safety-Critical Reinforcement Learning
not read this cycle: ranked 19 of 19 by the pre-filter, below the top 12 the editor reviews
- editorgithub.com
Prompt injection scanner that normalizes evasion tricks
not read this cycle: ranked 13 of 13 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Instrumental Monitor Evasion Emerges Under Ordinary Task Pressure
not read this cycle: ranked 13 of 14 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
A Living Benchmark for Information Retrieval from Electronic Health Records
not read this cycle: ranked 14 of 14 by the pre-filter, below the top 12 the editor reviews
- editorneuromatch.social
Meta Muse: Almost no prompt injection resistance
not read this cycle: ranked 14 of 14 by the pre-filter, below the top 12 the editor reviews
- editorreuters.com
AI leaders warn UN of security risks as systems grow more powerful
not read this cycle: ranked 14 of 15 by the pre-filter, below the top 12 the editor reviews
- editorgithub.com
What if we created a local AI tool to measure your security?
not read this cycle: ranked 15 of 15 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
An Open Pipeline and Dashboard for Systemic-Risk Evidence under the EU AI Act's Code of Practice
not read this cycle: ranked 13 of 13 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
A Zeroth-Order Paradigm for LLM Preference Alignment
not read this cycle: ranked 13 of 14 by the pre-filter, below the top 12 the editor reviews
- editorkhancyr.github.io
I caught an LLM-powered recruiter with a prompt injection on LinkedIn
not read this cycle: ranked 14 of 14 by the pre-filter, below the top 12 the editor reviews
- editora16y.ai
The Security Risks in 100k Enterprise AI Sessions
not read this cycle: ranked 13 of 13 by the pre-filter, below the top 12 the editor reviews
- editornytimes.com
China's Top Spy Chief Warns A.I. Is a Threat to Party Rule
not read this cycle: ranked 13 of 19 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
The Model Proposes, the Code Disposes: A Pre-Registered Ablation of a Verifier-and-Acceptance Stage in an LLM-Orchestrated Offensive-Security Agent
not read this cycle: ranked 14 of 19 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Privacy-enhanced federated learning via asynchronous aggregation and local differential perturbation
not read this cycle: ranked 15 of 19 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
LLM-Based Schema-Aware Split Learning for Privacy-Preserving Mental Distress Prediction Across Heterogeneous Surveys
not read this cycle: ranked 16 of 19 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Privacy-Aligned Personalized Federated Learning with Compact Adaptation and Variable-Length Gaussian Communication
not read this cycle: ranked 17 of 19 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Safe Meta-Reinforcement Learning via Information Space Reachability
not read this cycle: ranked 18 of 19 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Private Information Retrieval With Arbitrary Privacy Requirements: Introduction and Capacity Results
not read this cycle: ranked 19 of 19 by the pre-filter, below the top 12 the editor reviews
- editorapiweiser-cli-fe.vercel.app
Show HN: AI-Native dependency upgrades through codemods
not read this cycle: ranked 13 of 20 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
MOONWALK: Mediating Operations with Intent-Evidence-Action Alignment Across Junior-Supervisor Review Workflows in Animation/VFX Pre-Production
not read this cycle: ranked 13 of 13 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Wicked Problem, Parsimonious Solution: Securing Electric Vehicle Charging Station Software
not read this cycle: ranked 13 of 15 by the pre-filter, below the top 12 the editor reviews
- editorarmosec.io
Prompt Injection Through Tool Output Is Two Events (Your Screens Read One)
not read this cycle: ranked 15 of 15 by the pre-filter, below the top 12 the editor reviews
- editorresearch.meta.ai
How We Built Safety into Muse by Meta
not read this cycle: ranked 15 of 15 by the pre-filter, below the top 12 the editor reviews
- editoraquilax.ai
WebAssembly Security: The Sandbox That Is Not What You Think
not read this cycle: ranked 13 of 13 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Trust-Aware Adaptive Disclosure for Inference Privacy Preservation in Multi-Agent Networks
not read this cycle: ranked 13 of 13 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Machine Unlearning as Private Retroactive Algorithms
not read this cycle: ranked 13 of 14 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Seeing Before Synthesizing: VLM-Guided Transition Event Discovery for Weakly-Supervised Dense Video Captioning
not read this cycle: ranked 13 of 13 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Overcoming the Randomness-Utility Trade-off in Answering Differentially Private Linear Queries
not read this cycle: ranked 13 of 16 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
The Implications of Linguistic Illegibility for LLM Security
not read this cycle: ranked 14 of 16 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Large Language Models (LLMs) for Telecom Root Cause Analysis (RCA): A Structured Reasoning Framework for Evidence-Grounded Diagnosis
not read this cycle: ranked 15 of 16 by the pre-filter, below the top 12 the editor reviews
- editorbartholomew.info
In-memory security gate for autonomous AI agent tools
not read this cycle: ranked 16 of 16 by the pre-filter, below the top 12 the editor reviews
- editorsemantic-overlays.vercel.app
Show HN: Semantic Overlays – an NX bit for LLM prompt injection (live demo)
not read this cycle: ranked 13 of 17 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Beyond Scores: Understanding LLM-as-a-Judge Mechanisms in Summarization Evaluation
not read this cycle: ranked 13 of 13 by the pre-filter, below the top 12 the editor reviews
- editorshostack.org
Help Launch Threat Modeling 2nd Edition
not read this cycle: ranked 13 of 13 by the pre-filter, below the top 12 the editor reviews
- editordelphisecurity.ai
Two sentences our security scanner can't tell apart and how we solved it with ML
not read this cycle: ranked 13 of 14 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
LLM-Based Agents for Software and Systems Security: Approaches, Applications, and Assessment
not read this cycle: ranked 13 of 13 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Anatomy-Aware Promptable Segmentation with Online Interactive Training for AUTOPET V
not read this cycle: ranked 13 of 14 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
RedEvoAgent: Automatic Red-Teaming Agent with Experience-Driven Skill Evolution
not read this cycle: ranked 13 of 13 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
RTLGuard: A Lightweight Teacher-Student Defense for Poisoned RTL Code Generation Models
not read this cycle: ranked 13 of 16 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Robust CurveMoE: Multi-Norm Adversarial Defense for Mixture-of-Experts Models via Mode Connectivity
not read this cycle: ranked 14 of 16 by the pre-filter, below the top 12 the editor reviews
- editorgithub.com
Shieldprompt – test your LLM against prompt injection – no dependencies
not read this cycle: ranked 15 of 16 by the pre-filter, below the top 12 the editor reviews
- editorgist.github.com
Darkbloom (AI inference on idle Macs) – security audit with PRs submitted
not read this cycle: ranked 16 of 16 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
LION: A Clifford Neural Paradigm for Multimodal-Attributed Graph Learning
not read this cycle: ranked 13 of 13 by the pre-filter, below the top 12 the editor reviews
- editortheregister.com
OpenAI's overhead will rise 20 percent for some workloads as it hardens security
not read this cycle: ranked 13 of 13 by the pre-filter, below the top 12 the editor reviews
- editorexploiting.systems
Prompt Injection in VirusTotal's Code Insights API
not read this cycle: ranked 14 of 14 by the pre-filter, below the top 12 the editor reviews
- editorgithub.com
Blocking prompt injection deterministically costs 40% false positives
not read this cycle: ranked 14 of 14 by the pre-filter, below the top 12 the editor reviews
- editoraivillage.org
Show HN: AI Village: AI security, tested in public.
not read this cycle: ranked 13 of 14 by the pre-filter, below the top 12 the editor reviews
- editorarxiv.org
Convergent Detour Hijacking: Task-Preserving Resource Amplification in Skill-Based LLM Agents
not read this cycle: ranked 14 of 14 by the pre-filter, below the top 12 the editor reviews
Other — 18
Reasons that do not fall into a standing category.
- editorarxiv.org
Ensuring Safe Physical AI in Urban Mobility via Hazard-Informed Synthesized Envelopes
While this paper discusses safety in urban mobility, it is not directly relevant to AI security, and its focus on physical AI in urban mobility makes it adjacent-but-not-really relevant to our field.
- editorgithub.com
Xaidr – In-process runtime security and governance for AI agents
This story is too focused on announcing a product (Xaidr) without providing substantial technical detail or analysis that extends our previous coverage of AI security and governance.
- editorgithub.com
AgentShield – Offline Rust security scanner for AI agent tools and MCP (<50ms)
This story, while mentioning a specific tool (AgentShield), does not offer enough depth or new insights into AI security that would significantly add to our previous discussions on the topic.
- editorgithub.blog
What 50 open source projects taught us about security in the AI era
This story merely repeats the angle that AI vulnerability extends beyond autonomous agents and does not add new substance to the discussion.
- editorqueue.acm.org
Beyond Zero: Enterprise Security for the AI Era
This story is too general and does not provide any new insights or developments in AI security, merely discussing the need for enterprise security in the AI era without any specific examples or technical details.
- editorcvd.z.ai
Z.ai Security
This story is too brief and lacks any substantial information about Z.ai Security, failing to provide any technical details or developments that would make it relevant to our feed.
- editorsentrint.com
Show HN: Security scanner for SaaS build with AI
This story is about a security scanner for SaaS built with AI, but it does not provide any specific examples or technical details that would make it stand out as a significant development in AI security.
- editorarxiv.org
TraVEL: Trajectory-Guided Video Embedding Learning for Driving-Video Retrieval
While the paper discusses video retrieval for driving logs, it does not directly address AI security concerns, making it less relevant to the field.
- editorarxiv.org
Synthetic Persona Pretraining: Alignment from Token Zero
The paper's focus on aligning language-model-based AI with human goals and values, although important, does not provide a concrete development or technical detail that would make it a priority for publication in the context of AI security.
- editorarxiv.org
VICBench: A Multi-Language Benchmark for Code Vulnerability Detection
Although this research paper is relevant to AI security, it does not present a concrete development or incident but rather introduces a benchmark for code vulnerability detection, which, while important, does not have the same level of urgency or direct impact as a reported attack.
- editorarxiv.org
Learning-Based Behavior Planning for Automated Driving: Real-World Integration and Deployment
This paper is not directly relevant to AI security as it focuses on learning-based behavior planning for automated driving, which is more related to robotics and automation than to security.
- editorportswigger.net
Can AI do novel security research? Meet the HTTP Terminator
This story, discussing whether AI can do novel security research, is too speculative and lacks concrete substance or a specific development to make it worth publishing over other candidates.
- editorarxiv.org
One Frozen Simulator Is Not Enough: Simulator Collapse in Multi-Agent RL
This research paper, focusing on simulator collapse in multi-agent reinforcement learning, is not directly relevant to AI security but rather to the broader field of AI research.
- editorgithub.com
I pointed my AI security plugin at my own prod app – it refused to invent a bug
The story about an AI security plugin not inventing a bug in a prod app lacks the substance and impact of a reported security incident or a concrete development in AI security.
- editoritmeetsot.eu
Prompt Injection Experiments with Opus-5 in Claude Code – Auto-Mode Edition
The experiments with prompt injection, while related to AI and potentially to security, do not constitute a significant development or incident in AI security to warrant publication over other candidates.
- editorarxiv.org
How Organizations Use AI: Evidence from ChatGPT
This study on how organizations use AI, based on evidence from ChatGPT, is more about the adoption and use of AI in general rather than specifically about AI security.
- editorarxiv.org
An Efficient Near-Optimal Algorithm for Adversarial $m$-Set Bandits
The algorithm for adversarial $m$-set bandits, while relevant to AI research, does not directly pertain to AI security and lacks the concrete development or incident required for publication.
- editorarstechnica.com
Prompt Injection as Defense
The concept of prompt injection as defense is interesting but lacks the concrete development, incident, or significant technical detail needed to make it a priority for publication.