FOMO

For the past more than a year now, I’ve been feeling anxious about not having materially incorporated a breakthrough technology in my day-to-day.

AI. Two syllables that continue to shape the world on an unprecedented scale. From applications in drug discovery to sparking conversation about originality and art to being weaponized in disinformation campaigns, the buzzword has become far more pervasive penetrating most of our lives beyond the technology bubble as well.

This was completely caused by the lack of trying. I had actively stayed away, wanting to avoid the hype and give time for the technology to mature. The feeling was exacerbated, more so since I worked at the cutting-edge of deep learning till only a few years back.

I believed that incorporating AI into my PM workday for writing PRDs, user stories and such would eventually result in additional overhead. Primarily due to the lack of context. For existing, nuanced and complex products such as those that exist in the financial services industry, context is everything. I felt that the time spent providing GPT with the necessary context, constraints, and nuances might outweigh the time saved in generating output. Add to that the inevitable rounds of editing, and the net gain seemed questionable.

For those familiar with the field, understand that AI is mostly just an acronym. More abstract than tangible. More spectrum than binary. While generative transformer models represent a quantum leap in prediction and natural language processing, the leap to general intelligence—reasoning like humans—remains a tall order. Yann LeCun has pointed out repeatedly, we’re still far from that milestone.

This left me wrestling with the question: Is it imperative that I learn to work with, interface or even develop with LLMs? While I wasn’t actively working in AI, I’d kept a close eye on industry trends and developments. But keeping up with the relentless pace of innovation eventually became overwhelming. My Type A brain sought an alternative: dive in and cut through the noise.

My thesis for doing this was simple.

Beyond content creation, the strongest use case/largest impact genAI had was on reducing the time to delivery for software products. Widely adopted by the developer community and just disruptive enough to change the way software was getting built. I also knew that one needed technical know-how and working knowledge of how to build software products to leverage these capabilities.

As someone who didn’t have a formal software engineering background but had built software products, from the ground up, genAI seemed to promise something that felt too good to be true. Could this technology truly empower individuals like me to build and deploy digital products without writing a single line of code?

Fueled by this question, I embarked on an experiment. Armed with my experience in product management and software development, I embarked upon a journey to build increasingly complex web/mobile apps-entirely using AI tools. I vowed to not write a single line of code.

This journey is as much about validating the hype as it is about uncovering the limitations of generative AI. Whether this technology lives up to its transformative promise remains to be seen, but diving in has already taught me one thing: the only way to cut through the fear of missing out is to act.

Read about my experience of building software MVPs with generative AI here.