@artheaiguy@surim0n I was copy/pasting code in and out of ChatGPT from my browser back in 2023, before Codex, Cursor, Claude Code, etc.
I built a stupid simple game back then, which now has an install base of over 3M, and has about 180 DAU.
The Product Engineer was always the better model.
It just took AI making implementation cheap for companies to admit it.
For years, we split the work like this:
- PM decides what to build
- Engineer builds it
Two roles. Two perspectives. Constant handoffs.
It looked efficient. But it wasn’t.
Every handoff lost context.
Every spec introduced ambiguity.
Every iteration took longer than it should.
The split only existed because implementation was lengthy and expensive.
You needed specialists focused on execution. Now that constraint is gone.
AI made building fast. And that exposed the inefficiency of separating thinking from doing.
The people who decide what to build should be close to building it.
The people who build it should understand why it matters.
That’s the Product Engineer role.
Not a new role. Just the natural shape of how software should have been built all along.
AI didn’t create this shift. It revealed it.
@catchintherift@ztisdale The only reason their elbows are up are to stick their fingers in their ears and pretend to ignore a proposal for a possibly good idea. Just like anything in life, there’s pros/cons - but it’s easy to get caught up in one and ignore the other.
PMing with o1 pro, v0, and DeepSeek-R1
The days of just writing specs are over. AI can now generate detailed PRDs in minutes. The future belongs to technical PMs who can design and build, not just document.