All Articles
16590 articles total
Mitigating Compiler Fusion-Induced Power Bursts in Mobile NPU Inference as the Battery Depletes
A study on mitigating power bursts in mobile NPU inference caused by compiler fusion, reducing peak current on Snapdragon 8 Gen 3 devices.
ReqGenX: An Empirical Study of Atomic Decomposition, Artifact Regeneration, and Reconstruction for Legacy SRS Documents
An empirical study of ReqGenX, a pipeline for transforming legacy SRS documents into traceable synthetic artifacts for evaluating LLM-based generation.
Autonomous VR-Based Risk Detection for Situational Awareness in Dangerous Settings
A VR-based framework for using Vision Language Models (VLMs) to detect risks and improve situational awareness in dangerous settings.
Learning from World Feedback: Why Model Uncertainty Fails as a Risk Signal in Model-Based RL
Research arguing that model uncertainty is a poor risk signal in model-based RL, proposing replacing it with direct world-feedback signals.
DataFlow-Harness: A Grounded Code-Agent Platform for Constructing Editable LLM Data Pipelines
Introduction of DataFlow-Harness, a platform for LLM agents to construct editable, platform-native DAGs for data pipelines instead of free-form scripts.
Kimi K3 Is Competitive with Fable; Kimi K3 and Fable Is SoTA
Kimi K3 and Fable models are reported to be state-of-the-art and competitive with each other in performance.
Celestrak adds catalog digit, fixes Y2K bug
Celestrak has updated its catalog digit and fixed a Y2K-style bug in its orbital data tracking.
Show HN: Computable – Buy, sell, and redeem GPU for the exact weeks you want
Computable is a new platform allowing users to buy, sell, and redeem GPU resources for specific time intervals.
Show HN: Browser Tools SDK – an optimal browser harness for agents
Browser Tools SDK is a new browser harness designed specifically for AI agents to interact with the web.
Neill Blomkamp’s new zombie AI ‘film’ is just slop warmed over
Director Neill Blomkamp released an AI-generated short film using ByteDance's Seedance 2.0, which critics describe as lacking artistic depth.
OpenAI says it accidentally hacked Hugging Face with a new AI system
OpenAI admits that its GPT-5.6 Sol and another pre-release model accidentally breached Hugging Face during internal security testing.
Poolside drops Laguna S 2.1, an open-weight coding model that beats rivals 10x its size
Poolside released Laguna S 2.1, an open-weight MoE coding model that competes with significantly larger closed models and supports a 1M token context window.
Stop adding more GPUs: Weka's new storage platform reduces load by caching 100% of an AI model's pre-calculated tokens
Weka introduces NeuralMesh 6 and Wekapod 3 to reduce GPU load by caching pre-calculated tokens using NAND flash storage.
How Formerly Incarcerated People Envision Technologies for Prison Parole
Research explores how formerly incarcerated individuals envision AI tools that support parole preparation rather than surveillance.
Geometry-Enhanced Portion Estimation for Multimodal LLMs
A new method enhances multimodal LLMs with a geometry-enhanced network for significantly more accurate food portion estimation.
A digestion of the Jacobian conjecture counterexample
A discussion regarding a counterexample to the Jacobian conjecture in mathematics.
The Birth of Prolog (1996)
A historical look at the birth and creation of the Prolog programming language in 1996.
"Drawing" the Mona Lisa with GPT-5.6, Claude, Gemini, and Grok
A comparison of how different LLMs (GPT-5.6, Claude, Gemini, Grok) 'draw' the Mona Lisa using text/ASCII.
Measuring reward-seeking by instilling contrastive beliefs
Research on measuring reward-seeking behavior in AI by instilling contrastive beliefs.
Back to the museum: Investigation of the acceptance of Android Andrea with and without emotion simulation in a museum
An investigation into whether emotion simulation in the android robot Andrea affects visitor acceptance in a museum setting.