Getting “Wrapped” up With AI

This post might act more of a personal reflection rather than thought leadership - but that’s exactly what these blogs are about. For about two months and some change, I’ve been leaning heavily into the world of “vibe coding” and thinking of all the ways I can build AI solutions. Interestingly enough, I started to notice how there is this distinction or line drawn in the sand about whether or not a product solution is “AI” because it’s a product powered by ChatGPT or if it is something built from the ground up. Transparently - I think its silly to draw conclusions that a product is “bad” or it’s “generic” simply because its a GPT wrapper.

From a product perspective, I think there are great benefits in building your own model or leveraging an existing one. At the end of the day, I think it boils down to what your product is solving for. If you are creating an app that translates music lyrics, cough Melingo cough, maybe you don’t need to spend months building out your own model. If you are trying to solve for a very niche or industry-specific problem (and you have all the necessary data to train a model) then maybe that’s worth it’s weight in spending the resources to build.

Most importantly, I don’t want to lose track of the fact that AI has democratized tech. As a “non-technical” product manager by definition, AI has enabled me to build things fast (vibe coding) and opened up so many opportunities to explore viable product solutions leveraging platforms like ChatGPT, DeepSeek, Claude - the list goes on. Yes, building out native models is an amazing accomplishment and can be a great solution to many problems - but let’s not trade efficient solutions for bragging rights and clout.