eric zakariasson

i asked some people at @cursor_ai how they use cursor

some nuggets in here ↓

@shaoruu
- create dedicated "playgrounds" to iterate on the ui/ux or logic with the ai, kinda like a free-roam space until you get something you like, and then integrate it back to your main app

- use @ pr and @ commit to review changes &amp; undo things you don't like

@ryolu_
i use cursor to… make prototypes + try crazy ideas
and they usually are way easier than i imagined

@baltaaazr
i think the most underrated / hidden feature that i really like is resume from here, like the little + buttons after a codeblock

i use it when the agent gets stuck on lints, so i can resume from its first attempt which is usually 80% of the way there and continue from there

@ajhofmann18
- don't ask agent to do something until you understand the problem and solution that you want to see. If you need help understanding the problem and solution, iterate in ask/codebase chat first

- if agent is consistently one shotting your requests, try aiming higher (do more work in one request). If it's consistently getting stuff wrong, break the problem down into smaller requests

- reference the correct patterns that exist in your codebase as the basis for implementing new work

@danperks_
- new chats very often, almost every 5 minutes probably, cmd+n is my most hit shortcut

- lean into the linter. setting up a linter to flag what you want the AI to see, but ignore what you don’t can be v helpful for the agent fixing its own code

- be scrappy with it. i frequently clone a repo from GitHub, use ask to find out how one area works, then throw it into my own repo to implement

@ericzakariasson
- let cursor generate tool it can execute instead when it's faster

- setup common workflows with rules that the agent can execute. e.g resolve merge conflicts

- be very explicit and provide as much context as possible with @ symbols

i've built 10+ internal tools at @cursor_ai using cursor. when you have a platform in place, it makes it so easy to spin up new tools

here's how ↓

the first piece is to have the primitives in place like database, servers, networking, workers etc

i've setup this on aws with terraform 

second step is to make sure agent is aware of all primitives. here's a cursor rules i've defined to make it aware of what it can use. this outlines building blocks, references examples and points to MCP tools for introspection 

with this, i only need to setup the following once
- security groups in network stack
- auth middleware t1hat intercepts all routes
- reusable frontend components (s/o @shadcn )
- boilerplate workflow to be used as reference for best practices

once configured, i can start directing agent to build what i want for me 

this is one of many ways you can use cursor, hope it was helpful!

i wrote about this in longer form almost a year ago if you're interested in reading more
https://anyblockers.com/posts/the-idea-of-internal-no-code-platforms