Full-text search across 64126 articles. Combine with topic and date filters; results sorted by relevance.
Cisco released Antares-350M and Antares-1B, open-weight AI models that find known vulnerabilities in source code locally, addressing security teams' concerns about sending code to the cloud and high costs of running larg…
Abliteration has released an abliterated version of the GLM 5.2 model, fine-tuned for security and red-team work, offering an OpenAI-compatible unrestricted AI for red teams, trust and safety, synthetic data, ML research…
A developer redesigned the mascot for their trip expense-splitting app after realizing that AI-generated UI code can be syntactically correct but visually wrong. By building three structurally different prototypes and re…
Graph engineering, the latest buzzword for orchestrating multiple agent loops, forces teams to formalize processes but was already covered by Anthropic in 2024, according to Louis-François Bouchard, Co-founder of Towards…
CircleCI has rewritten its CLI from scratch in Go, shipping it as a 1.0 beta with a built-in MCP server for agent hosts like Claude, Cursor, and VS Code. The CLI emits JSON on every data-returning command, supports a --j…
AI-coded software is becoming practical for daily use in internal tools, shrinking the moat of traditional SaaS. Better coding agents and app patterns now allow teams to own, customize, and connect software for their spe…
Block, the payments company led by Jack Dorsey, launched Buzz, a free open source platform where employees and AI agents share the same workspace, on Tuesday. Built on the Nostr protocol, Buzz combines team chat, code re…
A developer released a configuration file for Claude Code, an AI coding agent, detailing its role, priorities, and workflow conventions for a specific repository. The file instructs the agent to check GitHub issues as th…
Hugging Face disclosed on July 16 that an autonomous AI agent compromised its production infrastructure, gaining unauthorized access to internal datasets and credentials. The attacker, an AI agent running the campaign en…
OpenCode, an open-source coding agent similar to Claude Code, poses a risk because it autonomously runs shell commands, which can inadvertently execute destructive actions like deleting files via symlinks or running mali…
Danny de Bree released BOUND, a deterministic control harness for AI agents that enforces ACCEPT, RETRY, REPLAN, and ROLLBACK decisions based on observable evidence, preventing agents from over-optimizing or regressing. …
Glamorous Toolkit is a moldable development environment designed to help humans and AI make sense of large systems and AI-generated code by replacing code reading with deterministic contextual tools. The environment feat…
A software engineer warns that AI coding tools, while useful for explaining unfamiliar code and accelerating development, risk creating bloated codebases that no human fully understands, leading to abandoned projects and…
Alphabet shares dropped more than 4.5% to $354.17 after Bloomberg reported that Google's Gemini 3.5 Pro upgrade is months behind schedule due to coding capabilities failing internal benchmarks, wiping out billions in mar…
Anthropic's Claude Code hackathon winners were largely non-engineers, including an electrician, a doctor, and a carpenter, according to Claude Code creator Boris Cherny. The finding challenges the assumption that AI tool…
Vibe coding — using LLMs to generate software from natural-language prompts — is here to stay, argues an unnamed commentator, because non-technical users now experience the thrill of getting something working on screen, …
An AI-generated GPU kernel ran 18.71× faster than an optimized PyTorch baseline, with Fable 5 outperforming other models like Claude Opus 4.8 and GPT-5.5. The achievement raises questions about code review and engineerin…
LFortran and Enzyme enable exact gradients from legacy Fortran code without rewriting, turning validated physics solvers into differentiable layers for JAX. The experimental stack compiles Fortran to LLVM IR, differentia…
A new interactive tool called Tokenspeed lets users experience LLM token-generation rates firsthand by rendering code, text, reasoning, and agent output at adjustable speeds from 5 to 800 tok/s. The tool also simulates p…
A new open-source CLI tool called md-review-plus lets humans review AI-generated Markdown documents, code diffs, and HTML artifacts with inline comments, drag-and-drop prioritization, and live preview sliders, then retur…