Journal
The Blog
A fresh, in-depth review of an AI coding tool every week — plus benchmarks and unfiltered takes.
- 8 min read
Claude Code, three months in: what changed
A candid retrospective after 90 days of daily Claude Code use on a production TypeScript codebase.
Claude CodeReviewsRetrospectiveRead post → - 7 min read
Bolt.new: AI-Driven Development for the Enterprise
Bolt.new promises to revolutionize app and website creation through AI, integrating design systems and robust backend infrastructure. We critically examine its strengths, weaknesses, and ideal user base.
AI DevelopmentLow-CodeEnterprise ToolsWeb DevelopmentDesign SystemsRead post → - 9 min read
Cursor 2 review: is the composer mode enough?
Cursor 2 leans into the agent experience. We spent two weeks with it on real work.
CursorReviewsIDERead post → - 7 min read
GitHub Copilot Workspace: worth another look in 2026?
Copilot Workspace was underwhelming at launch. The 2026 rebuild changes the calculation.
GitHub CopilotReviewsRead post → - 10 min read
Windsurf Editor review: the quiet contender
Codeium's Windsurf has matured into a serious daily driver. Here's what it does better than Cursor, and where it still trails.
ReviewsWindsurfIDERead post → - 7 min read
Pieces for Developers: A Privacy-First AI Copilot for Your Code Snippets
Pieces for Developers offers a compelling vision for AI-assisted coding, prioritizing privacy and local processing. We delve into its strengths, weaknesses, and ideal user base.
AI DevelopmentCode SnippetsDeveloper ToolsPrivacyLocal LLMRead post → - 10 min read
Windsurf vs Cursor in 2026: which editor wins?
Two agentic IDEs, one production repo, thirty tasks. Here's what actually happened.
WindsurfCursorBenchmarksRead post → - 8 min read
Gemini CLI vs Codex CLI: terminal agent showdown
Two Google-and-OpenAI terminal agents, one week of unattended work. Which one earned its keep?
Gemini CLICodex CLITerminalRead post → - 7 min read
AI pair programming etiquette for teams
When everyone on the team uses an AI assistant, new social rules emerge. Here are the ones worth writing down.
TeamsProcessCultureRead post → - 7 min read
Should juniors use AI agents? An honest guide
AI agents make juniors faster and, sometimes, worse. Here's how to get the good half without the bad.
CareerBest PracticesRead post → - 7 min read
v0 by Vercel: A Sober Look at AI-Powered Application Generation
Vercel's v0 promises to revolutionize app creation with AI. This review cuts through the marketing to assess its actual utility for developers, highlighting its strengths and critical limitations.
AI DevelopmentApplication GenerationVercelDeveloper ToolsRead post → - 8 min read
Replit Agent for MVPs: honest review
We built three small products in Replit Agent to see where it shines and where it forces a rewrite.
ReplitApp BuildersReviewsRead post → - 9 min read
Gemini CLI deep dive: a terminal agent worth trying
Google's Gemini CLI quietly became one of the most capable terminal agents. Here's what it does well.
ReviewsGeminiCLIRead post → - 6 min read
When not to use AI coding tools
Some tasks are still faster, safer, and cheaper without AI. Here's the list we keep on the wall.
Best PracticesOpinionRead post → - 7 min read
How to evaluate AI coding tool privacy claims
Every vendor says they don't train on your code. Some are telling the truth. Here's how to tell.
SecurityPrivacyRead post → - 9 min read
AI code review tools, compared
Six AI PR reviewers ran across the same 40 pull requests. Only two we'd keep.
Code ReviewReviewsBenchmarksRead post → - 11 min read
Claude Code vs Cursor in 2026: which one ships faster?
Two weeks, two tools, one mid-size SaaS codebase. A side-by-side look at how Claude Code and Cursor actually feel under deadline pressure.
Claude CodeCursorComparisonsReviewsRead post → - 8 min read
Are AI-generated tests safe to ship?
Test suites written by AI can either lock in correct behavior or paper over bugs. Here's how to tell which you're getting.
TestingQualityBest PracticesRead post → - 8 min read
Prompting patterns for large refactors
The refactor prompts that actually work — plus the ones that produce plausible garbage.
PromptingBest PracticesRead post → - 7 min read
Feeding a monorepo to an AI agent without going broke
Naive context loading in a monorepo will blow your token bill. Here's the strategy we use.
MonoreposBest PracticesRead post → - 9 min read
Context windows in 2026: what 1M tokens actually buys you
Big context windows sound great in marketing. Here is what they actually change in your day-to-day coding work, and where they still let you down.
ModelsContextExplainersRead post → - 7 min read
Choosing AI coding tools for a five-person startup
Small teams can't afford a procurement process. Here's a practical playbook for picking tools in a week.
StartupsBuyer's GuideTeamsRead post → - 8 min read
The best AI tools for mobile developers in 2026
iOS and Android have their own quirks. Not every AI tool is fluent in either.
MobileiOSAndroidReviewsRead post → - 9 min read
Self-hosting AI coding models in 2026: is it worth it?
Open models finally rival hosted ones. The economics still surprise most teams.
Self-HostingOpen SourceRead post → - 12 min read
The best AI coding assistant in 2026
We tested every major tool on a real production codebase. Here's the ranking, the scores, and the surprises.
ReviewsBenchmarksCoding AssistantsRead post → - 7 min read
How to migrate your team from one AI tool to another
Switching AI tools is disruptive. A staged rollout beats a big-bang cutover every time.
Team OpsBest PracticesRead post → - 8 min read
When to trust an agent loop, and when to pull the plug
Autonomous agents can run for hours and produce great work — or quietly burn $40 of API credit going in circles. Here is how to tell them apart.
AgentsBest PracticesWorkflowRead post → - 9 min read
Using AI assistants on legacy codebases
AI tools shine on greenfield work. On a fifteen-year-old codebase, they need a different playbook.
LegacyRefactoringBest PracticesRead post → - 9 min read
Cursor vs Copilot: a real-world benchmark
Same task, two tools, head-to-head timings and quality scores across 30 real engineering tickets.
BenchmarksCursorGitHub CopilotRead post → - 10 min read
GitHub Copilot for Enterprise: an honest review
What you actually get for the enterprise tier, what stays the same as the team plan, and whether the price tag makes sense at 200 seats.
GitHub CopilotEnterpriseReviewsRead post → - 10 min read
GitHub Copilot Workspace: an honest six-month review
Workspace promised an end-to-end agent inside GitHub. Half a year in, here's what's living up to the pitch.
ReviewsGitHub CopilotAgentsRead post → - 7 min read
Should solo devs pay for AI tools?
When the free tiers are enough — and when paying $20/mo earns itself back in a week.
Solo DevsPricingBuyer's GuideRead post → - 8 min read
The best AI tools for junior developers in 2026
Tools that teach as they help — and a clear list of ones to avoid until you have more reps.
Junior DevsBuyer's GuideLearningRead post → - 8 min read
AI assistants and the open-source maintainer's inbox
Maintainers are drowning in AI-generated issues and PRs. Here's how the healthier projects are coping.
Open SourceCommunityProcessRead post → - 10 min read
AI code review tools: which one actually catches bugs?
We seeded 50 real bugs into PRs and ran them past CodeRabbit, Greptile, Korbit, and Qodo Merge. Catch rates inside.
Code ReviewBenchmarksReviewsRead post → - 9 min read
Your code, their training data: an AI privacy primer
What every major AI coding tool does with the code you paste in — and the settings that actually protect proprietary work.
PrivacySecurityComplianceRead post → - 10 min read
Which AI tools actually handle monorepos well?
Monorepos break a lot of AI tooling. We tested the major options against a real 200k-line workspace.
MonoreposBenchmarksToolingRead post → - 8 min read
Vibe coding with AI app builders: a 30-day report
I built and shipped seven small apps using only AI app builders. Here's what worked, what didn't, and which builder I'd use again.
App BuildersReplitBoltv0Read post → - 11 min read
Self-hosted AI coding stacks: are we there yet?
Open weights, open agents, your own GPUs. We benchmarked a fully self-hosted setup against the frontier hosted tools.
Self-HostedOpen SourceBenchmarksRead post → - 8 min read
Reviewing AI-generated code for security: a checklist
AI assistants are confident, fast, and routinely produce subtly insecure code. Here's a pragmatic review checklist.
SecurityCode ReviewBest PracticesRead post → - 7 min read
Better prompts for coding: a practical field guide
Stop writing 'fix this'. A short, opinionated guide to prompts that consistently get usable code on the first try.
PromptingBest PracticesWorkflowRead post → - 8 min read
AI for debugging: what works, what wastes your time
Debugging is where AI tools either save your week or make it worse. A pragmatic guide to the situations that suit each pattern.
DebuggingBest PracticesWorkflowRead post → - 9 min read
Calculating AI tool ROI for small teams (without lying to yourself)
A spreadsheet-grade method for figuring out whether your AI subscriptions are actually paying off — or just feel like they are.
ROIBuyer's GuideManagementRead post →