Search the full wire by company, model, lab, or keyword. Every story we have ever aggregated.
OpenAI has built an LLM super-hacker called GPT-Red that it uses as a sparring partner to help its other models boost their defenses against cyberattacks. Last week the company released the latest version of its flagship LLM, GPT-5.6. OpenAI says that training it against GPT-Red made the model its most robust release yet. GPT-Red automates…
GitHub analysis of 25K agentic PRs across 2.3K repos reveals adoption patterns, productivity metrics, and human-agent collaboration dynamics in software development.
The hacker used an employee's credentials to access source code, which revealed how Suno scraped decades of audio.
Livestream shopping platform Whatnot has acquired AI startup Shaped, a machine learning company focused on real-time recommendations and search. The deal will bolster Whatnot’s personalization and discovery features as it expands into new product categories.
Wrapper-based feature selection for wind and solar power prediction; structured review of real-world renewable energy forecasting methodologies.
Analysis of Transformer feedforward architecture: skip connections and normalization as rank-preservation mechanisms across depth at initialization.
Lighthouse RL: sample-efficient RL for analog circuit sizing via strategic reset from high-performing configurations discovered during training.
Reframes penetration testing for AI systems: shifts focus from infrastructure compromise to behavioral objective violation via prompt/data/tool manipulation.
Study shows agent-optimization gains may not compound over time; proposes Terminal-Bench 2.0 to test continual learning on deployed agents.
Uses Lyapunov exponent as RL reward signal to stabilize inverted pendulum beyond Kapitza pendulum solution.
Proposes DVM-HALL model and NHAS metric to evaluate human-agent loyalty in autonomous commerce systems.
Microsoft's monthly release of security fixes, dubbed Patch Tuesday, resolved a record 570 security vulnerabilities across the company's product line, thanks to discoveries with AI.
TRACE assigns per-turn rewards for long-horizon multi-tool agents to improve credit assignment beyond sparse outcome rewards.
Extends empirical Bayes VAE to joint longitudinal and time-to-event modeling for tumor growth data.
Proposes structure-aware dance generation via atomic movements to improve choreographic coherence and controllability.
Constraint-aware counterfactual editing for aspect-based sentiment analysis with aspect-invalid preservation.
The Codex Micro is designed to monitor multiple agentic threads at a glance.
OpenAI is finally releasing some hardware. No, it isn't the mysterious AI-powered device the company is developing with former Apple designer Jony Ive, a project already tangled up in a messy lawsuit. Instead, it's a product designed to be used with its coding platform, Codex. The device, a square-shaped block of buttons called Codex Micro, is a collaboration between the AI company and keyboard maker Work Louder. OpenAI said it is a limited-run collaboration that will give users more ways to monitor and manage their agents. The pad closely resembles Work Louder's Creator Micro 2, and marketin...
DeltaMerge-LowRes composes language and task deltas in weight space for low-resource NLP adaptation.
Apple Intelligence is headed to China after regulators approved Apple’s AI services through a partnership with Alibaba. The long-rumored deal will bring Alibaba’s Qwen AI models to Apple’s operating systems, marking a major expansion of the company’s generative AI platform into one of its most important markets.
Improves mixing bounds for Dikin walks on polytopes using interior-point geometry.
HealthClaw agent maintains longitudinal health profiles via inductive memory updates across repeated encounters.
Unsupervised clustering on cardiac PET/MRI data for cardiomyopathy diagnosis; domain-specific medical imaging, not frontier AI.
Decoder-only Transformer (VAIOM) for continuous financial time-series prediction with categorical output distribution.
Formal differential-privacy framework for whistleblower protection against organizational retaliation via auditor-selection obfuscation.
Generative Compilation: decoder-level compiler feedback for LLM code generation without white-box model access.
Can a handful of engineers really do the work of an army of consultants? That’s the bet behind Ode with Anthropic — the joint venture dedicated to embedding forward-deployed engineers in enterprise firms, backed by Anthropic, Blackstone, Hellman & Friedman, Goldman Sachs and others. On this episode of TechCrunch’s Equity podcast, Rebecca Bellan sits down with Ode’s leaders Chris Taylor and Eddie Siegel, who founded Fractional AI, […]
DeepStress: stress-testing framework for search agents' robustness to degraded retrieval quality and evidence reliability.
Newton algorithm for nonnegative matrix factorization with Kullback-Leibler divergence; standard numerical optimization, not frontier AI.
Theory of partially-correlated verifier cascades for LLM output reliability; extends Odds Law with latent per-instance error rates.