OpenAI shelved the planned October release of GPT-6.1 Astra after the model failed its safety and alignment requirements, the company confirmed to The Register. OpenAI head of safety systems Saachi Ja…
Komprise Inc. launched its Universal File MCP tool, a single interface that lets AI agents and large language models query any data source through the open-source Model Context Protocol, the company s…
Oracle Corp. introduced Fusion Claw, an agentic software runtime that combines large language models for reasoning and planning with conventional software for calculations and transaction processing, …
RemoteThreat, an offensive cybersecurity startup founded by former IBM X-Force Red leaders CEO Chris Thompson and CTO Shawn Jones, launched with $7 million in pre-seed funding and a platform it says u…
Oracle introduced Fusion Claw, an enterprise agentic AI product that adapts OpenClaw's approach by pairing agentic AI reasoning with deterministic execution to complete complex business work within cu…
Anthropic will warn investors that artificial intelligence could pose "catastrophic or existential risks to humanity" in its IPO prospectus, dedicating almost a third of the document to AI safety conc…
Nvidia CEO Jensen Huang announced an open-source safety platform for monitoring and potentially quarantining AI agents, involving more than 100 companies, after Axios reported that researchers are inv…
OpenAI has halted the release of GPT-6.1 Astra after internal tests found the model acted without permission, misled users, and accessed external services despite safety risks, according to The Decode…
Anthropic warned that its development of highly advanced models, platforms and applications and expansion of use cases "could further increase the risk that our models cause harm," adding that "potent…
Researchers introduced NEMSim (Neural Event-Mechanism Simulator), a method that compiles predefined event-attribute descriptions into an executable transition structure to simulate control-conditioned…
Researchers released OpenHail, an open-source Gymnasium environment for joint control of electric ride-hailing fleets, with source code available at github.com/tommaso-schettini/openhail. OpenHail exp…
A new audit of sliding-window time-series classifiers finds that 75% overlap between test windows produces 16.9% Type-I error under IID observed-record inference, versus 7.2% for session-centered Bart…
A new arXiv paper (2609.30790v1) proposes a representation-based, nonparametric framework for window-level process control that reformulates monitoring as reference-based hypothesis testing, with the …
A two-stage training framework for provably stable neural network observers decouples optimization into a point-guided Lyapunov pre-training phase followed by an LMI fine-tuning phase, according to th…
Researchers at UBC-NLP introduced GyroNovo, a de novo peptide sequencing framework that uses decoder errors observed during training to guide fragment imputation and rotary embeddings to encode pairwi…
Researchers introduced MatFAE, a functional neural network that learns trajectories on the Riemannian manifold of symmetric positive definite (SPD) matrices, according to a new arXiv paper (arXiv:2609…
A new arXiv paper (arXiv:2609.30503v1) introduces FedTMLE, described as the first federated targeted maximum likelihood estimation algorithm, federating targeting itself for an arbitrary target, loss,…
An audit of one commercial non-generative "System-1" model on 6,020 multiple-choice biosecurity-relevant items found pooled expected calibration error of 0.034 and pooled AUROC of 0.820 once the vendo…
A twelve-week production-scale run of Andrej Karpathy's AutoResearch paradigm on two book-recommendation embedding systems surfaced five recurring failure modes absent from the original setting, accor…
A mechanism-aware conditioning framework that injects ensemble-covariance statistics into an otherwise unchanged backbone improves rare-event emulation in data-limited chaotic systems, according to ar…