Full-text search across 57801 articles. Combine with topic and date filters; results sorted by relevance.
In a head-to-head benchmark of 12 fresh text tasks, Kimi-K2.7-Code scored 103.5 against gpt-5.4's 101.5, with only 50% confidence that either model is better, resulting in a tie. Kimi edged ahead on code-centric reasonin…
A developer testing Grok's adversarial critique of their own AI agent work found that verification failures are not just technical but human. The agent skipped requested telemetry, and the developer rubber-stamped the om…
A developer at a tech company picked a coding agent based on public benchmark scores, only to find it failed on their real codebase due to gaps like measuring others' code, not their own. After analyzing the failures, th…
A developer created pi-tiered-router, an extension that routes each phase of a coding agent to a different model, optimizing cost and performance. The system breaks a coding turn into classify, plan, validate, and execut…
A developer shares 10 essential VS Code snippets for C++ engineers, arguing that snippets maintain coding flow better than AI for repetitive tasks by avoiding context switching. The snippets, available on GitHub, cover i…
Sophos X-Ops collected seven days of endpoint telemetry from Claude Code, Cursor, and OpenAI Codex running on Windows in June 2026. Behavioral detection flagged all three, with 56.2% of blocking rules triggered by creden…
OpenAI's GPT-5.6 Sol matched Anthropic's Claude Fable 5 on the Code Arena benchmark for coding agents while costing 40% less. The performance-per-dollar advantage makes GPT-5.6 Sol the recommended starting point for agen…
AgentKindergarten is a new tool that streams a developer's entire coding environment to their phone, allowing AI coding agents to continue working on a home PC while the user is away. The self-hosted system uses a daemon…
A developer built DevDiff, a real-time pull request risk intelligence platform that combines rule-based security analysis, machine learning, and optional LLM reasoning to identify risky code changes. The platform uses a …
A developer released Agent OS, an open-source local-first AI project operating system that provides a harness around coding models. The system separates responsibilities between a Main Agent and a Coding Agent, with sand…
A team built a prototype that verifies code migration from F# to Python using a proof in the Lean theorem prover, rather than relying on tests. The system translates both the original and rewritten code into Lean, then p…
A developer warns that relying on LLMs to write code without following best practices leads to accumulating code smells, as the model learns from and replicates bad patterns in the codebase, making future maintenance har…
Hamza built Satori, an open-source, read-only codebase map for MCP-compatible AI coding agents. It provides a structured investigation path to help agents efficiently navigate symbols, ownership boundaries, and dependenc…
LangChain released OpenWiki, an open-source tool that uses an AI agent to automatically generate and maintain a structured wiki for codebases, keeping documentation synchronized with code changes via CI integration. The …
AI can now write code dramatically faster than human developers, raising questions about the future value of programmers. The author argues that while AI makes coding cheaper, it does not make building real software easi…
A developer tested ChatGPT (GPT-5.2), Claude (Opus 4.8), Gemini (3.1 Pro), and Grok 4 on the same real coding tasks and found that the best model depends on the specific job. For debugging, reasoning-heavy models like Cl…
SpaceXAI released Grok 4.5, a general-purpose AI model trained alongside Cursor for coding, agentic tasks, and knowledge work, priced at $2 per million input tokens. The model claims superior token efficiency and ranks f…
A developer reports that DeepSeek V3, while fast at generating boilerplate code, over-engineers legacy code reviews by suggesting architectural rewrites that introduce race conditions, whereas Claude 3.5 Sonnet is more c…
Tako Research released TakoVM, an open-source sandboxing tool for AI agent code execution that provides Docker isolation, job queues, execution history, and retry logic out of the box. The tool aims to simplify secure co…
Assembled open-sourced 143.dev, its internal infrastructure for running coding agents as a team, under an MIT license. The system runs multiple agent harnesses in sandboxes and integrates with tools like GitHub and Slack…