@nt
If you think you need to spend X amount of weeks/days/hours working to launch the first version of something, really start to believe that X/2 and X/3 is likely more optimal.
I've progressively learned this over the years from values of X = 4 months, then 1.5, then 3 weeks, and so on... I launched the initial version of /amp after a few days, the semantic search after just 2. I launched the AI-generated homepage (ampcast.xyz/news) with about 3 hours of work on the first prototype - it turned into one of the top launches on the network that day, but I was genuinely surprised and was expecting just a few people to try it out and give me feedback, and instead I've had over a thousand land on the page that day.
Shaving scope is its own challenge - but unnecessary delay in launching hurts your learning. Launch faster.