AI Coding Tools · Autonomous Coding Agents · AI in Software Engineering

A blog and newsletter on Vibe coding workflows and AI coding agents (Claude Code, Codex, etc.).

Vibe Coding Workflows

Prompt-Driven Feature Development in Production Codebases

Context and specification discipline matter more than clever prompts for sustainable AI coding.

November 6, 20269 min read

Autonomous Coding Agents

Autonomous Agent Evaluation on Real GitHub Repositories

October 26, 20268 min read

Vibe Coding Workflows

AI-Assisted Refactoring of Legacy Code

October 22, 20269 min read

Latest

Autonomous Coding Agents

Autonomous Agent Task Scoping to Prevent Runaway Edits

Limit agent tools and resources per task to prevent unintended damage.

Priya SubramaniamOctober 17, 20269 min read

Autonomous Coding Agents

Cost Control for High-Volume Autonomous Agent API Usage

Agents multiply token costs faster than per-unit prices fall.

Simone ValléeOctober 10, 202610 min read

Autonomous Coding Agents

Human-in-the-Loop Checkpoints for Autonomous Code Agents

Checkpoint decisions should follow consequence, not workflow position.

Priya SubramaniamOctober 9, 202610 min read

Autonomous Coding Agents

Agent Memory and State Persistence Across Coding Sessions

LLMs lose all context between sessions, forcing agents to rebuild understanding from scratch.

Priya SubramaniamOctober 8, 202610 min read

AI in Software Engineering

Interviewing and Hiring Software Engineers in an AI-Assisted Era

Senior engineers now lead AI output; junior work has moved to the tools.

Simone ValléeOctober 5, 202610 min read

AI in Software Engineering

Skill Atrophy Risks for Junior Developers Using AI Assistance

Junior developers lose critical debugging skills they don't yet realize they need.

Tomáš BlažekOctober 3, 202610 min read

AI Coding Tools

OpenAI Codex API Integration in a CI Pipeline

The `codex exec` subcommand turns Codex into a scriptable automation tool for CI pipelines.

Priya SubramaniamOctober 2, 202610 min read

AI Coding Tools

Cursor Agent Mode for Multi-File Edits

The autonomy slider moves from tab completion to independent codebase edits.

Priya SubramaniamOctober 1, 202611 min read

AI Coding Tools

GitHub Copilot vs Cursor for TypeScript Development

Cursor's full editor control beats Copilot's reach for TypeScript refactors.

Simone ValléeSeptember 30, 202610 min read

A blog and newsletter on Vibe coding workflows and AI coding agents (Claude Code, Codex, etc.).

Prompt-Driven Feature Development in Production Codebases

Autonomous Agent Evaluation on Real GitHub Repositories

AI-Assisted Refactoring of Legacy Code

Latest

Autonomous Agent Task Scoping to Prevent Runaway Edits

Cost Control for High-Volume Autonomous Agent API Usage

Human-in-the-Loop Checkpoints for Autonomous Code Agents

Agent Memory and State Persistence Across Coding Sessions

Interviewing and Hiring Software Engineers in an AI-Assisted Era

Skill Atrophy Risks for Junior Developers Using AI Assistance

OpenAI Codex API Integration in a CI Pipeline

Cursor Agent Mode for Multi-File Edits

GitHub Copilot vs Cursor for TypeScript Development

Archive