Full-text search across 58791 articles. Combine with topic and date filters; results sorted by relevance.
A computer repair shop owner who builds apps on the side created Zeltro, a desktop app that gives vibe-coding projects a single home with per-project addresses, a shared database, and an embedded terminal where AI agents…
ProgramDistill, a benchmark for real-world app reconstruction tasks, found that AI coding agents fail 51% of the time, including GPT-6 Astra, according to byteiota. The result highlights a persistent gap between AI codin…
Jindo AI is building Pawsly, a coordination layer for multi-agent coding workflows that shares plans, roles, decisions and commit state across parallel agent sessions. In internal benchmarks on ten two-agent coding tasks…
Researchers developed EMBODIEDSWE-BENCH, a simulation benchmark for coding agents that spans contact-rich manipulation and deformable object tasks in long-horizon, dexterous robotics, to test whether agent-generated solu…
A Hacker News user asked engineers at highly polished product companies such as Linear, Instagram, Google Docs, and Figma how they use agentic coding, noting that agentic tools get projects 90% of the way to "finished" b…
A developer documented how to recover a lost Claude Code session when the `claude --continue` flag fails to find the previous conversation. Because Claude Code stores sessions per project folder under `~/.claude/projects…
Developer kvit released What's Next, a free MIT-licensed Claude Code plugin that ends an agent turn with a keyboard-selectable card of two to four options instead of a prose report, built on Claude Code's AskUserQuestion…
A developer analyzed 2,116 messages sent to coding agents including Claude Code, Kimi, Cursor, and Copilot over three weeks and found that roughly 40% of typed input was overhead rather than real work. The breakdown show…
A developer built LucentraCode, a multi-model coding CLI designed for long sessions and real projects, aiming to offer strong AI coding usage without requiring $100–$200/month max-tier subscriptions. The tool targets dev…
A coding agent marketed as the fastest available claims to run more than 5x faster than Codex, according to promotional material for the tool. The agent is powered by OpenCode and is pitched at "high velocity individuals…
Justin Johnsen, lead technical architect at KPMG, said AI has shifted his forward-deployed engineering role away from hand-writing code toward gathering context and refining intent, and that on one client project AI help…
A new empirical study examines how coding harness design affects autonomous coding agents' long-horizon software-engineering performance, arguing that prior work evaluates harnesses as monolithic systems and leaves the e…
A newly discovered vulnerability dubbed Plugin4Shell allows attackers to execute remote code with a single click or none at all in AI coding agents, according to the report. The zero-click flaw grants attackers access to…
Tines published "The IT leader's guide to AI code sprawl," a downloadable guide warning that employees across departments are shipping apps, agents and automations built with AI tools, often without recognizing the outpu…
Developer Nurysso released Eulix, an open-source, locally run code navigation tool that parses repositories into symbols and call graphs and combines that structural data with keyword and semantic retrieval. Nurysso test…
A 12-year-old developer built and ran a global coding competition, the KODA Code Jam, entirely from a $150 Android phone, using Supabase for authentication, PostgreSQL storage, and Row Level Security to manage submission…
Software developer @javisrike said AI coding has reached diminishing returns, reporting that their default orchestrator, luna, routes work by complexity and that actual usage lands around 16% luna, 48% sol, 30% astra, an…
Emacs upstream maintainer Sean Whitton announced that the original fix for the arbitrary code execution flaw CVE-2024-53920 was incomplete, after Bas Alberts found that viewing or editing untrusted files in modes other t…
Watson Labs reported that its stealth software factory Gymwasp processed 372 issues over six months using agentic workflows with no human code review, routing humans instead to plan validation and product decisions. The …
A developer built Loomwright, a Claude Code plugin that enforces a plan-first workflow for coding agents, requiring a reviewed pull request instead of letting agents edit files before 'done' is defined. The plugin, publi…