How do you get to know what will work? I'm referring this. > The difficulty now, as back then, is landing on the technology complement that proves the most tech

How do you get to know what will work? I'm referring this. > The difficulty now, as back then, is landing on the technology complement that proves the most technically no-nonsense and financially viable in the long run. I had my try with Vector Databases as well. But at the same time, I highly feel people have started to add vector databases everywhere. And now generative code makes you feel you don't know anything. Feeling lost in the middle.