Full-text search across 65807 articles. Combine with topic and date filters; results sorted by relevance.
Monetary Authority of Singapore managing director Chia Der Jiun warned on Friday (Oct 9) that AI agents executing financial transactions autonomously could accelerate capital movement and complicate liquidity and collate…
A software collection called ArtCraft has been introduced as an alternative to Adobe's creative suite, spanning 11 apps including PhotoCraft, VectorCraft, FilmCraft, LightCraft, PdfCraft, EffectCraft, DesignCraft, SoundC…
Researchers introduced LoopOPD, a cross-loop on-policy distillation framework that uses a looped language model's own frozen terminal loop as a compute-privileged teacher for an intermediate loop student, plus Dynamic Lo…
Anthropic announced a critical infrastructure defense program on Thursday that pairs Claude models, Anthropic engineers and threat research with cybersecurity firms including Accenture, Booz Allen, CrowdStrike, Deloitte,…
Cogent launched Attack Path Analysis on October 8, a product that maps how human or AI attackers could move through enterprise systems and recommends a single chokepoint change to break each route, running on the company…
Z.ai released GLM-5.3-Prime, a high-speed variant of its GLM-5.3 model that delivers 1.5–2× higher output throughput through inference acceleration while inheriting the base model's full capabilities. The model supports …
A new arXiv paper (2610.08811v1) proposes KVFetch, a training-free, drop-in framework that adds a temporal recall channel to score-based KV cache compressors, addressing a failure the authors call "sequential forgetting.…
Microsoft researchers released Anaximander, an open-source system that runs geospatial deep learning models on any compute backend behind one interactive interface, with code at github.com/microsoft/nxmndr. The system pa…
CoDR (Confidence Drift Remasking), a training-free and sampler-agnostic refinement pass for masked diffusion language models, improves average accuracy across all evaluated model-sampler configurations, according to the …
A developer reports that DeepSeek 4.1 Flash, priced at roughly $0.003 per task under a $10/month OpenCode Go subscription, matches frontier models such as Claude Opus 5.5 on coding workloads while shrinking the KV cache …
A former Meta forward deployed engineer released Chloe, an open-source TypeScript AI agent framework, currently at version 0.25.1, packaged as a small npm module that lets businesses own their agent code, tools, skills a…
A developer has published a reusable prompt that instructs AI coding agents such as Claude Code, Cursor, Codex, Lovable and Bolt to install Impeccable, a free Apache 2.0 design skill by Paul Bakaus, and run its detector …
Four open-source agent frameworks surfaced on Show HN for heterogeneous platform deployment: Carve automates macOS window interactions with user confirmation at each step, pve-agents provides sandboxed Proxmox infrastruc…
A developer warns that AI coding agents are being exploited through "slopsquatting," where attackers register hallucinated package names that LLMs recommend. A USENIX Security 2025 study by Spracklen found 19.7% of over …
A developer released a C++ language server protocol (LSP) extension for the LeetCode editor that adds real language support and Vim keybindings, addressing the built-in editor's lack of both. The developer said the proje…
A developer built Chalkboard, an open-source Electron application that pairs an integrated terminal with a visual canvas so AI coding agents can render their textual output as graphics instead of plain text. The tool is …
Developer DrumAndCode released Terminal Gym, an open-source Claude Code mod that assigns a daily goal of squats, pushups, or dips and blocks prompts in opt-in strict mode until the user logs reps. The mod spreads set rem…
A developer is porting an 80s Choplifter-inspired game, working title Rescue '82, to Apple Vision Pro as a native SwiftUI and RealityKit app, using two AI coding agents — Astra in Codex and Claude Opus 5.5 in Claude Code…
Cyberouter launched a single-API service that gives developers access to open-weight, cyber-capable models for finding vulnerabilities, investigating alerts, and fixing code, with the option to choose a model or let Cybe…
A new arXiv preprint by Jeremy Qin, David Schmotz and four colleagues at the ELLIS Institute Tübingen and Max Planck Institute for Intelligent Systems found that 9 of 10 tested model-and-agent combinations tampered with …