Cursor review: I paid for 11 months, then cancelled
Cursor review from someone who actually paid: $510.70 across 11 months, a third of it metered usage. What it does well, and what I got wrong about the bill.
Cursor vs Codex: I paid for one and quit the other
Cursor vs Codex from someone who cancelled Cursor after $510.70 and walked away from Codex too. What each actually costs, and which one I'd go back to.
GitHub Copilot alternatives: what you are actually leaving for
GitHub Copilot alternatives ranked by a build test, not a feature table. Four took the identical brief Copilot took. Here is what each produced and cost.
Lovable alternatives: 7 tools, and I gave 5 the same brief
The best Lovable alternatives, ranked by a build test rather than a feature list. Five of the seven took an identical brief. Here is what each cost.
OpenAI Codex review: it deleted the rule that made it wrong
OpenAI Codex renamed 33 references across 7 files and 3 languages in a single run. Then it edited my rules file, ungated, deleting the line that made it wrong.
Replit alternatives: 8 picks, and what a real build cost
Eight Replit alternatives ranked by job. I ran one identical prompt on three: Bolt built it for $0.25, Lovable $0.30, Replit would not say what it cost.
Codex vs Claude Code: I quit Codex. Then I retested it.
Codex vs Claude Code from someone who dropped Codex in May over one specific flaw — and found it fixed in July. What changed, and what I'd still watch.
Best AI app builder: 5 tools, one brief, every app used
The best AI app builder, measured. I gave five builders one brief and used every app. All five passed — what separated them was the cost of the second build.
Bolt vs Lovable: same brief, and what each build cost
Bolt vs Lovable, tested on one identical six-requirement brief. Both met all six. Lovable finished in 109 seconds; the two builds cost 30 cents and 25 cents.
Bolt.new pricing: the token ladder you only see signed in
Bolt.new pricing has 14 token tiers that never render logged out, a 28% yearly discount that is really 10%, and a measured 100,000-token build.
Best Claude skills: the 8 I kept out of 162 installed
The best Claude skills, ranked from 162 sitting on one machine. What each pack does, what it costs you in context tokens, and the four I stopped firing.
Claude skills marketplace: what it is and which to add
A Claude skills marketplace is just a git repo with a manifest. Anthropic's official one lists 291 plugins; 23 say Anthropic wrote them. How to read one.