Just sharing my thoughts here because I’ve used comfyui and other workflow building tools—prompts are actually just a small (but important) part of the whole process. Personally, I don’t think there’s a need to systematically study prompts, and it’s definitely not about writing essays. It’s more about guiding AI, and there are various ways to do this, like using controlnet lora, depth maps, normal maps, Cenny edge detection, and open pose. When these models first came out, there was even a job called a Prompt Engineer, but that role has faded over time. Mastering a model’s prompts doesn’t guarantee that the same prompts will work effectively after the model upgrades or integrates other guiding models.

I’ve noticed that the review process for Chrome extensions is super simple and efficient. Usually, you can get something up and running in just a day, which is a whole different ballgame compared to the App Store and Play Store. So now, whenever I come up with a new idea, I ask myself: Do I really need to develop an app for this? Can it be done just as well via a web solution? If so, I might just create a plugin to test the waters. Plus, plugins have a big advantage: they can use the user’s Google account that’s already logged in, which is super convenient. Chrome used to handle payments, but now that’s up to the developers, making it possible to create subscription-based products.