avatar
Christian2022.eth
30,831 Twitter followers
Follow
Partner @NextGen_Venture Contributor @WeirdoGhostGang “All in Crypto to achieve common prosperity”
Posts
avatar
Christian2022.eth
01-08
Thread
A few months ago, I saw a top Silicon Valley product manager say that the ratio of engineers to project managers would gradually approach 1:2. Now it seems less reasonable. AI is actually improving the efficiency of product managers' work more, because engineering implementation requires accuracy, while PRDs (Product Requirements Documents) can have relatively more lenient quality standards. Here's my experience from the past two days of hands-on practice on how to quickly produce a PRD for a medium-complexity feature: 1. After determining the priorities and scope of the features, directly send a few sentences in natural language to assistants like Manus/GPT/Gemini, requesting a comprehensive PRD for comparison (takes about 10 minutes). 2. After receiving the results (which are usually not good enough), reduce some features according to your understanding to achieve a streamlined MVP, add some boundary case considerations, and refer to some excellent competitors. (Takes about 30-60 minutes) 3. If there's no prototype, you can only send it to Lovable. If it's an iterative requirement, you can take screenshots of the UI and send the optimized prompt to Alloy. Remember to note that it must strictly adhere to the existing style, and remember to plan ahead for the number of pages needed and their functions. (Time taken 15-20 minutes) 4. After receiving the interactive prototype, adjust it according to your own aesthetic preferences (Time taken 30 minutes) 5. Submit the final prototype (screenshots or code) back to the agent who generated your PRD and ask them to modify the PRD accordingly (Time taken 10 minutes) 6. Combine the two relatively satisfactory versions into a final document and submit it to the designers and developers for evaluation and further refinement of details. I've recently interviewed many senior PMs at large companies, and hearing their lengthy product processes almost made me roll my eyes. This era belongs to teams with good intuition, who can quickly understand requirements, produce quickly, and validate quickly.
GPT
5.73%
loading indicator
Loading..