Blog / Archive / February 2025
February 2025
28 published posts filed under February 2025. Each article shows its actual publication date.
-
JetBrains AI Assistant: First inspection the agent actually listened to
A field-tested take on the first inspection the agent actually listened to with JetBrains AI Assistant: what it rewards, where it breaks, and how to keep the workflow honest.
-
Windsurf: Windsurf defaults we kept and the ones we overrode
A field-tested take on Windsurf defaults we kept and the ones we overrode with Windsurf: what it rewards, where it breaks, and how to keep the workflow honest.
-
Cursor: Composer for the first multi-file refactor we trusted
A field-tested take on composer for the first multi-file refactor we trusted with Cursor: what it rewards, where it breaks, and how to keep the workflow honest.
-
Google Gemini: First migration we ran through a single long prompt
A field-tested take on the first migration we ran through a single long prompt with Google Gemini: what it rewards, where it breaks, and how to keep the workflow honest.
-
ChatGPT Codex: First Codex PR that shipped without a human touch
A field-tested take on the first Codex PR that shipped without a human touch with ChatGPT Codex: what it rewards, where it breaks, and how to keep the workflow honest.
-
Claude Code: Cost spike that taught us to checkpoint
A field-tested take on the cost spike that taught us to checkpoint with Claude Code: what it rewards, where it breaks, and how to keep the workflow honest.
-
GitHub Copilot: Copilot Chat in the terminal versus the editor, a first look
A field-tested take on Copilot Chat in the terminal versus the editor, a first look with GitHub Copilot: what it rewards, where it breaks, and how to keep the workflow honest.
-
Local LLMs with Ollama: Thermal limits on the first serious local code session
A field-tested take on thermal limits on the first serious local code session with Local LLMs with Ollama: what it rewards, where it breaks, and how to keep the workflow honest.
-
Agentic coding: Guardrails before autonomy, learned the expensive way
A field-tested take on guardrails before autonomy, learned the expensive way with Agentic coding: what it rewards, where it breaks, and how to keep the workflow honest.
-
MCP servers: First auth model we picked, and later regretted
A field-tested take on the first auth model we picked, and later regretted with MCP servers: what it rewards, where it breaks, and how to keep the workflow honest.
-
AI coding workflows: Scorecard that replaced vanity metrics in month three
A field-tested take on the scorecard that replaced vanity metrics in month three with AI coding workflows: what it rewards, where it breaks, and how to keep the workflow honest.
-
Prompt engineering for code: Role prompts the first time we stopped cosplaying with them
A field-tested take on role prompts the first time we stopped cosplaying with them with Prompt engineering for code: what it rewards, where it breaks, and how to keep the workflow honest.
-
Cline and Roo Code: Roo modes that encoded our first team conventions
A field-tested take on Roo modes that encoded our first team conventions with Cline and Roo Code: what it rewards, where it breaks, and how to keep the workflow honest.
-
JetBrains AI Assistant: Refactoring across types with AI Assistant and not losing sleep
A field-tested take on refactoring across types with AI Assistant and not losing sleep with JetBrains AI Assistant: what it rewards, where it breaks, and how to keep the workflow honest.
-
Windsurf: What the first long Cascade session actually cost
A field-tested take on what the first long Cascade session actually cost with Windsurf: what it rewards, where it breaks, and how to keep the workflow honest.
-
Cursor: Tab predict surprises that made us rewrite comments
A field-tested take on tab predict surprises that made us rewrite comments with Cursor: what it rewards, where it breaks, and how to keep the workflow honest.
-
Google Gemini: Gemini CLI as a weekend scripting partner
A field-tested take on Gemini CLI as a weekend scripting partner with Google Gemini: what it rewards, where it breaks, and how to keep the workflow honest.
-
ChatGPT Codex: Parallel tasks and the temptation to scope too wide
A field-tested take on parallel tasks and the temptation to scope too wide with ChatGPT Codex: what it rewards, where it breaks, and how to keep the workflow honest.
-
Claude Code: Sub-agents the first time they saved an afternoon
A field-tested take on sub-agents the first time they saved an afternoon with Claude Code: what it rewards, where it breaks, and how to keep the workflow honest.
-
GitHub Copilot: Early art of rejecting suggestions without hurting velocity
A field-tested take on the early art of rejecting suggestions without hurting velocity with GitHub Copilot: what it rewards, where it breaks, and how to keep the workflow honest.
-
Local LLMs with Ollama: Private RAG pipelines, assembled out of a weekend of curiosity
A field-tested take on private RAG pipelines, assembled out of a weekend of curiosity with Local LLMs with Ollama: what it rewards, where it breaks, and how to keep the workflow honest.
-
Agentic coding: Planning outputs as the real product of the initial turn
A field-tested take on planning outputs as the real product of the initial turn with Agentic coding: what it rewards, where it breaks, and how to keep the workflow honest.
-
MCP servers: Early protocol tour, done with a plain JSON-RPC client
A field-tested take on an early protocol tour, done with a plain JSON-RPC client with MCP servers: what it rewards, where it breaks, and how to keep the workflow honest.
-
AI coding workflows: Pair-with-AI rituals for a team that had never tried them
A field-tested take on pair-with-AI rituals for a team that had never tried them with AI coding workflows: what it rewards, where it breaks, and how to keep the workflow honest.
-
Prompt engineering for code: Structured outputs, discovered the hard way
A field-tested take on structured outputs, discovered the hard way with Prompt engineering for code: what it rewards, where it breaks, and how to keep the workflow honest.
-
Cline and Roo Code: Plan mode as a cheap insurance policy from day one
A field-tested take on plan mode as a cheap insurance policy from day one with Cline and Roo Code: what it rewards, where it breaks, and how to keep the workflow honest.
-
JetBrains AI Assistant: Structural search as the hidden rail for early prompts
A field-tested take on structural search as the hidden rail for early prompts with JetBrains AI Assistant: what it rewards, where it breaks, and how to keep the workflow honest.
-
Windsurf: Moment Windsurf made a skeptical reviewer pause
A field-tested take on the moment Windsurf made a skeptical reviewer pause with Windsurf: what it rewards, where it breaks, and how to keep the workflow honest.