Claude Desktop fails to initialize Model Context Protocol (MCP) servers when the claude_desktop_config.json file specifies "npx" as the command, because the app does not inherit shell environment vari…
GPTY, an open-source terminal multiplexer that combines the Godot game engine for its UI with Rust for performance-critical work, is available on GitHub at github.com/godot-pty/gpty. The tool offers t…
A practitioner's test of SimpleDesign 1.0 found the joint protein sequence-and-structure model fails when its sequence and structure latent spaces are not aligned, producing physically impossible "hal…
A new Firefox extension and public API called Pangram flags and hides LLM-generated posts on the Hacker News front page by assigning each thread an AI-probability score, with users able to set a thres…
The OWASP LLM Top 10 framework can be applied to prompt optimization by constraining prompt logic, validating outputs, and adding structural guardrails, according to a technical guide. The guide recom…
AI-generated code most often fails because models hallucinate library versions, invent nonexistent function parameters, or lose context in long conversations, according to a debugging guide that recom…
Claude 3.5 Sonnet can generate clean code, but shipping it to production without a strict validation pipeline risks subtle regressions and security holes, according to an analysis of agentic coding pr…
OpenDiscoveryTrace released a dataset of 558 full AI-agent trajectories across 124 scientific tasks in drug discovery, genomics, and materials science, hosted on arXiv as 2609.09203v1, to expose how o…
A developer comparison of Anthropic's Claude Code CLI agent and Cursor's IDE found Claude Code faster for complex refactoring (6 minutes versus 15 minutes) and runtime debugging (2 minutes versus 5 mi…
A proposed evaluation method for video captioning replaces BLEU and METEOR with Multiple-Choice Question Answering (MCQA), scoring captions by the percentage of video-grounded questions an LLM can ans…
Nexpath, a middleware layer that appends verification steps, constraints, and acceptance criteria to AI coding prompts, improved Claude Code's performance on SWE-bench Verified from 27 of 40 tasks sol…
A developer-focused guide details specific Cursor configurations — including a `.cursorrules` file, targeted `@Files`/`@Folders`/`@Docs` references instead of `@Codebase`, and the Composer (Cmd+I) for…
A roundup of AI programming communities recommends Hugging Face for model weights and datasets, Stack Overflow for specific bug fixes, and PromptCube for structured prompt engineering, positioning the…
A developer detailed a "Constraint-First" prompting workflow for Cursor after a race condition in a Next.js 15 project led Claude 3.5 Sonnet to introduce three new bugs while attempting fixes. The dev…
A guide published on an AI-focused site recommends seven alternatives to Reddit for AI coding discussion: Hugging Face, the Cursor Discord, Stack Overflow's AI-tagged section, Dev.to, PromptCube, X, a…
GPT-Live-1, a new full-duplex voice API, is now available to developers at $0.05 per minute, allowing the model to listen and speak simultaneously and detect user interruptions in real time. The model…
Anthropic's Claude Code is a command-line interface agent that installs via `npm install -g @anthropic-ai/claude-code` on Node.js v18 or later, authenticates through an Anthropic Console account, and …
A head-to-head test of Cursor, Windsurf, and Claude Code on a 40k-line Next.js project found Cursor's Composer needed 4 tries to fix a React useEffect race condition, Windsurf's Flow took 2 attempts, …
Shopify is abandoning React Native and returning to separate Swift and Kotlin native codebases, citing AI coding agents that make cross-platform code sharing less necessary. The company plans to archi…
Developers can stop large language models from hallucinating nonexistent API calls by injecting local documentation or current SDK source into the context window rather than relying on the model's tra…