The Archive

HPOQuest is a training-free agent for rare-disease diagnosis using active phenotype acquisition and iterative questioning to refine disease rankings.

Salesforce positions AI Force and headless agents as UI replacement, reflecting industry shift away from traditional interfaces as competitive moat.

HiLNO applies hierarchical latent representations with multi-scale supervision to improve neural operator learning for PDEs on general geometries.

Dependency-aware trajectory refinement identifies and removes redundant agent steps via round-level DAGs, improving fine-tuning efficiency and agent performance.

OpenAI economic research documents worker adoption patterns and emerging AI-enabled job activities across enterprises.

AI isn't some kind of new form of "alien mind," according to Jensen Huang. It's just hardware and software, so safety can be engineered by each AI product maker.

Cohere and OpenText partner to deploy agentic AI systems for government and regulated sectors with compliance controls.

Cohere and Aleph Alpha launch transatlantic sovereign AI with dual Berlin-Toronto operations for EU/Canada data governance.

Poll data released Tuesday by the New York Times and Siena University confirms what we've already been seeing, and what politicians are responding to - AI and data centers are incredibly unpopular. Asked if they support or oppose the construction of data centers to power AI tech, 61 percent of the 1,503 likely voters surveyed in early September (ahead of the recent AI safety slowdown discussion) said they were opposed, with only 14 percent saying they strongly support. The split was about even among 2024 Donald Trump voters, at 49 percent support to 45 percent oppose, while people who had vot...

Google releases Gemini 3.8 Live and 3.8 Live Extended Thinking speech-to-speech models with browser-based UI supporting voice interruption.

National outcry against data center construction has spread to Philadelphia, where officials suggested possible construction in a neighborhood already impacted by a now-defunct oil refinery.

A new WhatsApp Business MCP server lets developers use AI coding agents like Claude, Cursor, Codex, and ChatGPT to handle setup, messaging templates, testing, and troubleshooting.

Good Start Labs demonstrates skill transfer from railroad game training to financial research via modified training design, raising questions about agent generalization.

From Apple's repeatedly delayed Siri AI to OpenAI's messy 'super app' launch, here's a look at the AI projects that shut down or missed expectations.

Robots can start working outside physical cages and without safety barriers.

The AI frenzy could push U.S. data centers to become one of the largest consumers of natural gas in the world.

Research shows AI agent societies fail to coordinate across trust boundaries due to communication vulnerabilities; proposes social harness framework for secure multi-agent collaboration.

ScienceBuddy releases interactive research workspace coupling agent harness evolution with RL-based self-improvement for continual scientific task learning.

PhysStream autoregressive model generates physics-grounded video with streaming control via structured scene memory and fine-grained motion dynamics.

Chain-of-Self-Questioning framework reduces LLM hallucinations by conditioning answer commitment on explicit information-requirement assessment across eleven model families.

Systematic study of pruning effects on LLM tool-calling across Transformer, hybrid, and MoE architectures; evaluates 19,500+ smart-home instances.

LACE layer-wise compression method for neural audio codecs reduces sequence length by adaptive frame-rate merging across quantization layers.

The AI Contact Hotline is designed to be a discreet place where agents that have witnessed misbehavior can tip off authorities.

ENCP applies conformal prediction with episode normalization to provide coverage guarantees for Vision-Language-Navigation agent uncertainty estimation.

Fuse multi-agent simulation framework evaluates LLM social reasoning via user-mediated interactions with hidden-motive agents; addresses ground-truth verification challenges.

FreqSpaNet neural network detects unauthorized wireless hardware replacement via spatio-frequency polarization fingerprint learning.

LimiX-2 introduces Contextual Mechanism Networks for structured-data modeling, shifting from target-centric to mechanism-oriented joint distributions.

Theoretical analysis reveals optimization complexity gap between homogeneous and heterogeneous asynchronous distributed training.

Det-LIME extends LIME explainability to multi-instance detection tasks in marine mammal conservation applications.

Signal-to-noise analysis of biased pattern capacity in dense associative memory networks.