I’m building out an idea that’s been sitting on the shelf – a platform that allows users to gather around a shared prompt and share incremental updates, centering focus on process rather than product. During MVP development there’s been a lot of trial and error: discarding agent-generated code, re-prompting the agent, moving back to older commits, starting the project over during peak frustration*… A PRD, no matter how well defined, isn’t enough at this point in time for an agent to build a complex system that aligns exactly with the intended. I’m learning that guiding agents successfully, in addition to writing clear requirements, involves reducing their assumptions by setting the development context, building in validation, and specifying end goals for tasks.
Job hunting is exhausting. My days are a constant juggle between LinkedIn, Claude, Google Docs, and company websites—filtering jobs, evaluating fit, writing custom resumes. Add the pressure to apply quickly (applications submitted closest to posting date get priority) and networking for internal referrals (also get priority), and suddenly my entire day is gone.
My first attempts using vibe coding to build apps gave results that were flashy but lacked substance. A few weeks ago, I got frustrated using iMessage to coordinate a spontaneous hangout with friends. What was supposed to be an easy trip to our local bar turned into hours of back-and-forth texts. By the time we settled on a time, my internal battery was depleted.
Curious, I did what most people are probably doing—I opened a few vibe coding tools and entered the same prompt:
I want to build an app that helps me schedule casual and spontaneous hangouts with friends.
I haven't coded professionally since 2018, but those late nights debugging in my college's CS lab “dungeon” are burned into my memory. A few weeks ago, I decided to experiment with AI-powered coding assistants to understand how they might fit into my Product Management workflow.
My first attempt was ambitious—I fed the same prompt to multiple assistants to compare their solutions. While there was magic in watching software appear from a few words, the results felt overly complicated. The UI choices were distracting, and I couldn't easily visualize the project structure. These tools excelled at helping me “work backwards” and refine my problem, but I wanted something more tangible.
So I simplified. I decided to build a basic personal website with Cursor, and managed to go from nothing to hosted in just a few hours. It blew my mind.
It's impossible to escape the doom and gloom think pieces about how AI will impact tech—most of them fixated on the “death of the software developer.” As AI enhances existing development tooling, it's natural that developers will work at higher levels of abstraction and existing processes will become obsolete.
But this feels more like evolution than death. The way we communicate with computers is becoming closer to how we communicate with humans. There's a shift from identifying developers by the languages they write to the problems they solve.
Yes, this transformation is important. Yes, we need to understand AI's moral and ethical impact on humanity. But what interests me is how AI might help software move beyond Tech and into Craft.