TODO for AI
vs Cursor
A VS Code fork with built-in AI chat, tab completion and background agents. Here is where each one wins, with numbers, and who should pick which.
Multi-vendor, metered
Every providerWho should pick which
- You want autocomplete and inline edits while you type
- You never leave VS Code
- You want to delegate whole tasks and walk away
- You want work running while your laptop is closed
- You need the agent in Chrome, email and ad accounts, not only in the repo
What Cursor is, and where it stops
Cursor is the editor most AI-native teams type in. The tab completion and inline edits are the best in the category, it supports models from every vendor, and it feels like VS Code because it is VS Code. Background agents exist, but the product is built around a human in the editor, accepting suggestions as they appear.
That is a different job from delegating. Cursor is at its best when you are present. It is not built to run for two hours while your laptop is closed, to drive your browser, or to handle anything that is not source code. And with frontier models metered on top of the subscription, heavy agent usage makes the monthly bill hard to predict.
Engineers who want AI inside the editor while they type.
Three reasons Cursor users look elsewhere
The agent works while the editor is open and you are watching. Close the lid and it stops.
The $20 covers a quota; real agentic work on Opus or GPT-5.6 is billed on top.
No browser, no email, no marketing or ops tooling. Everything outside the repo is someone else's problem.
Six dimensions that decide it
What each one ships
Same ladder, different meter
Credits worth roughly the price. Cursor models are cheap, third-party frontier models bill at API cost, overage at API rates.
One pool priced in dollars of model cost, spendable on any vendor. Session and weekly windows pace it, and a 1.3-1.5x plan multiplier applies, so a $20 plan is about $15 of list-price API. Team seat prices go into one shared pool.