pivoting from a generalist PM to an AI generalist ✨
building in a world where a PM is essentially redundant
is a pm role relevant in 2025?
honestly? the traditional pm role is shrinking. the days of writing 30-page prds, running ticket factories, and playing telephone between design and engineering are numbered. ai can do most of that faster and better.
but here's what ai can't do — hold conviction on what to build next, say no to 90% of ideas, and connect dots across business context, user pain, and technical feasibility in real-time. that's the job now. less process manager, more opinionated builder.
i saw this shift coming and decided to lean in rather than resist it. instead of protecting the old pm playbook, i started rebuilding my workflow around ai-native tools and agentic coding.
building products with agentic coding (aka vibe-coding)
i went from writing specs to writing products. literally.
using tools like claude code, replit, and cursor, i started shipping functional prototypes and full apps — not mockups, not figma flows, but working software.
the workflow is simple: think clearly about the problem, prompt precisely, iterate fast, and ship. the bottleneck is no longer "can we build this?" — it's "should we build this?" and that's always been a pm's real job.
this portfolio site, the products in my tinkertank, the interactive widgets — all built this way. no traditional engineering team. just clear thinking and agentic tools.
what do i look forward to in 2026 ✨
more building, less managing. i want to keep exploring the intersection of product thinking and agentic workflows — where a single person with strong product instincts and ai fluency can ship what used to take a team of ten.
the pms who thrive next won't be the ones who master jira or stakeholder alignment decks. they'll be the ones who can think in systems, prototype in hours, and ship with conviction. that's the bet i'm making.