Interesting viewpoints comparison: What is the key to building AI products?

There is a consensus in the industry abroad — technology is the real moat in the AI era. No matter how good the marketing is or how accurate the demand understanding is, if the product itself isn't solid enough, it's all in vain.

In contrast, the domestic startup scene often follows a completely opposite approach: prioritizing marketing, demand exploration, and user growth. Both sides seem to be working on AI, but in reality, they are taking two different paths.

Why? Someone summarized it: creating a truly excellent AI product is far more difficult than imagined. It’s different from traditional SaaS — it requires deeper algorithmic accumulation, stronger engineering capabilities, and a longer polishing cycle. This cannot be compensated for by marketing in the short term.

So the question is: do you choose to bet on technology or on channels?