Full-text search across 59799 articles. Combine with topic and date filters; results sorted by relevance.
In a benchmark of AI coding agents building a Flappy Bird game in Unity, Claude Opus 5 scored 10/10 and shipped a working game with a 306-line PlayMode test suite, while Claude Sonnet 5 scored 5/10 because it placed the …
In 60 coding-agent runs across three inference routes, patch footprint—the number of files a change touches—was the strongest predictor of success, beating every model-side variable measured. Local DeepSeek V4 Flash quan…
Supabase open-sourced Evals, a free benchmark that scores AI coding agents on real backend tasks, on August 1, 2026 under an Apache-2.0 license. Early results show that with access to Supabase's 'skills' context files, S…
A developer built a control plane that supervises multiple Claude Code CLI processes from a single dashboard, enabling five agents to run concurrently without becoming the message bus. The system uses a Flask app to spaw…
MonkeyCode has published a benchmark method for testing whether AI coding tasks survive mobile backgrounding, network switches, and battery pressure. The method classifies outcomes as recovered, restarted, or silent loss…
Researchers introduced VideoCoCo, an agentic dual-engine framework that uses executable Blender code as a chain of thought to generate physically consistent videos. The system improved the OmniWeaving baseline from 0.475…
A developer built Vinkius, an MCP server for Bland AI that transforms telephony integration from manual glue code into an LLM-driven capability. The server provides tools like send_phone_call and create_voice_agent, enab…
A developer introduced twd-cli, a tool that records test-driven UI interactions as smooth, paced videos directly in pull requests, aiming to eliminate the manual review bottleneck created by AI-generated code. The tool c…
Developer atasoglu released claude-nothing, a satirical CLI tool that simulates a productive Claude Code terminal session without modifying any files, displaying "0 files changed, 0 insertions(+), 0 deletions(-)" at the …
A developer who published six articles on DEV.to over two months received only about 200 total views, prompting an investigation into the platform's feed algorithm. By reading the open-source Forem code, the developer fo…
A developer has built a script that uses the GitHub Copilot SDK to automatically review the actual git branch by sending the diff between main and HEAD to a large language model for analysis. The script generates a JSON …
Claude Code, which now accounts for roughly 4% of all public GitHub commits as of April 2026, suffers from a compliance gap where it can articulate best practices like test-driven development but skips straight to implem…
Anthropic's Claude Code agent and Oxide's Project Glasswing hardware are being combined to create AI workflows where the LLM agent is integrated into the system's operational fabric rather than running as a guest, reduci…
A new study from arXiv preprint 2607.24797v1 finds that decoder-only large language models (LLMs) such as GPT-2, OPT, and Pythia entangle reading and writing into a single code, unlike the human brain's double dissociati…
A developer built Doc Bridge, a deterministic tool that provides coding agents with precise package-level context—such as which file to read first, which directories to edit, and which tests to run—rather than relying on…
A developer who passed the Microsoft AB-100 exam as part of the Microsoft Frontier Transformation Engineer journey reports that the exam is not a coding exam, focusing instead on solution architecture and business transf…
Anthropic's Claude Code uses a live classifier that calls claude-sonnet-5 to score each subagent tool spawn from 0 to 100, blocking any call above 50, but the classifier sits on the critical path with latency up to 3.2 s…
A developer created a custom statusline script for Claude Code that displays model, effort, cost, context bar, duration, rate limits, and git branch info in a width-aware two-line format. The script uses a single jq call…
A developer argues that running multiple coding agents in parallel creates a review bottleneck, not a compute bottleneck. The solution is to treat each agent session as a queue of reviewable artifacts with clear ownershi…
A developer has created a PreToolUse hook for Claude Code that blocks the AI agent from modifying .env files, preventing accidental overwrites of database passwords and API keys. The 60-line hook intercepts write_file an…