20 free resources on this topic. Everything here is free and hand-picked. You can also search within this topic.
Cursor's research post breaking down where coding-agent inference spend goes in its production traffic (tool outputs, tool definitions, reasoning) and describing changes it made to how agents manage context and choose subagent models. Figures are Cursor's own.
GitHub's recap of announcements from its Universe 2026 conference, including stacked pull requests that split large changes into smaller linked reviews, plus updates across the GitHub platform and Copilot.
A walkthrough of a personal to-do board built around a plain todos.md file, with setup prompts and a working demo containing sample data.
Substack post releasing agent-engineer, an open-source plugin for Claude Code, Codex, and Cursor that adds senior engineering skills, subagents, and hooks to AI coding assistants.
Ars Technica report on a developer using Claude Opus to build free, open-source alternatives to Adobe Creative Cloud apps, which the article describes as ambitious and far from finished.
A GitHub blog post describing Project HydraFusion, a research preview in GitHub Copilot that orchestrates multiple AI models for coding workflows. According to the post, in controlled offline evaluations HydraFusion's selective coding workflows matched or exceeded the evaluated Opus 5 baseline while reducing estimated workflow cost. Published September 4, 2026.
Blog post by software consultant Graham Lee (Structure and Interpretation of Computer Programmers), published October 6, 2026. It lists common objections to AI-generated code — code as art, AI as a job thief, software engineering as a set of rules, and AI companies as bad actors — and argues against each from the author's point of view, including a section on how programming jobs relate to automation.
Mia de Búrca argues that AI has removed the friction junior engineers once relied on to build critical thinking, and that leaders should make critical thinking an explicit expectation in reviews and postmortems.
Findings from JetBrains' Developer Ecosystem Survey 2026 of more than 15,000 professional developers: 90% used AI coding agents at work at least weekly as of mid-2026, and 68% used them daily. The report tracks how agents are carving out a place in developers' toolkits.
The documentation for Antithesis, an autonomous testing platform that continuously explores a simulated copy of a software system to find bugs. The docs cover setup with Docker Compose or Kubernetes, writing test templates, fault injection, the "multiverse" view of execution branches, and using Antithesis alongside AI coding tools.
Anthropic engineering write-up on running many Claude agents in parallel to build a working C compiler, including testing against GCC.
Why I recommend it: Anthropic describing its own product's success. The details are useful, but the results have not been independently reproduced.
The Hacker News on hijacked AI-memory packages that stole developer credentials through npm and PyPI.
Why I recommend it: A clear warning for anyone installing AI tooling — check package names and versions before you install.
Developer Zach Kehs on why AI-generated code can decline in quality without limit, and what that means for software teams.
Why I recommend it: One developer's personal blog and experience — sharp and worth reading, but not a study.
A September 9, 2025 blog post saying Veracode's research found security flaws in 45% of AI-generated code it tested, and giving advice on checking AI-written code before shipping it.
Why I recommend it: Veracode sells code-security scanning, so it benefits from this finding. The 45% figure comes from its own tests, not independent research. The advice to review AI code still holds.
GitHub changelog entry, 22 September 2026: GitHub Copilot CLI now builds a persistent symbol index across a whole C++ project via the Microsoft C++ Language Server, so jumping to definitions and references stops re-discovering the project each time. On by default; the first index build costs extra time and memory, and can be turned off.
Why I recommend it: Free to read and filed as a primary source, not a review. The speed benefit is GitHub's own claim with no published measurement; Copilot itself needs a paid or sponsored plan beyond its free tier.
Joe Procopio argues that "vibe coding"—letting AI write code without real understanding—is sliding into production disasters worth naming differently.
Why I recommend it: Inc. may show ads or meter premium articles; the link was accessible when it was added.
Armin Ronacher's critical look at long-horizon AI coding models and what they change about software work.
Why I recommend it: A skeptical engineer's take, which is exactly what I look for when every other post is hype.
Lobsters community discussion among working engineers on keeping code review standards and human judgment intact as AI-generated code volume grows.
Why I recommend it: Read the comments as much as the post. This is what hiring managers on engineering teams are actually worried about right now.
Working developer breaking down web tooling, AI coding tools, and industry news in plain, opinionated terms.
Why I recommend it: Good signal for anyone learning to build: he explains tradeoffs, not just tutorials, which is what interviewers probe for.
Developer-focused resource for building faster with AI coding tools and agent workflows.
Why I recommend it: Useful if you are building a technical portfolio — pair AI-assisted builds with a clear write-up of what you decided and why.
