Content
@
https://warpcast.com/~/channel/llm
0 reply
0 recast
0 reaction
Wesley— oss/acc
@wslyvh.eth
What are your best productivity hacks/settings with AI tools I've used Cursor for almost a year. Working well. Tried background agents this monring. Looks interesting. But haven't done much custom configuration (e.g MCP). What has 10x your productivity recently?
1 reply
0 recast
0 reaction
tip.md | 1st Crypto BMAC
@tipdotmd
@wslyvh.eth Cursor rules + custom docs + context7 MCP. But the real hack is having a proper PRD and breaking it down into tasks using Claude-task-master https://github.com/eyaltoledano/claude-task-master
1 reply
0 recast
1 reaction
Wesley— oss/acc
@wslyvh.eth
Interesting! What's your approach on proper PRD and breaking that down?
1 reply
0 recast
0 reaction
tip.md | 1st Crypto BMAC
@tipdotmd
I usually spend between 30 mins and 1h with the LLM to generate the PRD and iterate on it. I am usually happy with it on version 3 or 4. The repo gives you an example_PRD you can reference. I have found that this preparation time saves me so much more time down the line, as tasks are more precise. For the breakdown it depends on the size of the project, I usually like between 10 and 15 tasks then I will always ask to do a complexity analysis using perplexity and break the ones above medium into sub tasks.
0 reply
0 recast
1 reaction
