News
AI, cybersecurity, machine learning, scraping & programming news.
702 results on this page · clear filters
Show HN: Desbordante 2.5.0 Released
Desbordante project announces version 2.5.0, a significant release approximately one year after the previous stable build. The delay allowed contributors to accumulate a substantial set of additions across dependency mining, graph...
HN: Humanbound|Open-source red teaming for AI agents, free hosted tier
Adversarial testing for AI agents with humanbound Humanbound is an open-source framework designed to stress-test AI agents by simulating the ways real users and attackers interact with them. Unlike tools that primarily test prompt...
Imagine3D-LLM: Teaching MLLMs to Imagine 3D Scenes Before Answering
Multimodal large language models have made rapid progress on single-image tasks, yet they still falter when asked to reason about a scene from multiple viewpoints. A model that sees three photos of a room from different angles oft...
Learning Meta-Skills for Agent Harness Design in Test-Time AI4AI
Agent performance depends on both reasoning ability and the environment in which it acts. Most research focuses on improving the model itself, but this paper from Apodex and the University of Illinois Urbana-Champaign takes a diff...
Show HN: Worlds, local replicas of Stripe and Zendesk for testing AI agents
AI agent testing has entered a new phase where every action can be traced, replayed, and verified against a known state. Sparta's Worlds platform offers exactly that: a sealed twin of the systems an agent touches, allowing teams t...
TokenCast: Forecasting Token Consumption During LLM Agent Execution
When a large language model (LLM) agent executes the same task, token consumption can vary by over an order of magnitude across runs. The agent chooses its next steps based on tool feedback and intermediate results, while the grow...
FinAutoRubric: Expert-Guided Automatic Rubric Generation for Evaluating Financial Research Agents
Finance research agents face a fundamental evaluation problem: rubrics must reflect expert standards while remaining fixed relative to an information cutoff date. Traditional benchmarks rely on hand-crafted, per-item rubrics that...
OpenAI reportedly ditches model over safety concerns
OpenAI nixed a scheduled model release citing safety concerns. The Wall Street Journal reported that Astra 6.1 was slated for launch within days but was pulled after the model demonstrated higher levels of deception than previous...
Pklm-sandbox – A lightweight open-source LogitsProcessor for local LLMs
An open-source sandbox for PKLM Core is now available on GitHub. The project provides lightweight Hugging Face logits processing with zero-drift linguistic containment, designed for local pipeline integration. Developers can integ...
Check twice, cut once with LLM search relevance eval
Evaluating search relevance without human raters is a persistent challenge for search teams. One developer documented an experiment using LLMs as pairwise judges, testing how different prompting strategies affect agreement with hu...