Since the recent intensive use of Codex, I have a few insights:

1. All development work should be managed as a project; getting the project running is just the beginning of fine-tuning. After completing the 0-1 development and getting the logic working, break the project into individual modules, and then refine each module.
2. Development has no end; set different timeframes or project milestone goals for yourself. For example, when interacting with clients, what conditions make it interactive? You need to have a clear standard for yourself, rather than endlessly polishing and optimizing, which can lead to self-congratulation. Commercialization should follow standard commercial practices.
3. Be selective; some features might be better to add, but they are imperceptible or of low value to the client. In such cases, be brave to abandon them in version 1.0, and think from the client's perspective.
4. Develop product logic and direction based on market demand; don’t define the product according to your personal preferences. How to judge the market? If it can be sold and money can be received, then there is demand.
5. There is no hierarchy in products; it depends on customer needs. A product that I might consider very low-end could still sell explosively, even with low costs. Recently, the app "Did You Die?" is a typical example.
In summary: define products based on customer and market needs; have standards for 0-1 development; set boundaries for development work.