"We went from multi-day prototyping cycles to a few hours. Everyone uses Cursor now, even our ops team."
Gino Ferrand, writing today from Santa Fe, New Mexico 🌞
Ask any engineering leader where time goes, and you’ll get the same answers. Meetings. Reviews. Scope creep. And prototyping. Always prototyping.
At a recent Y Combinator event, Perplexity’s CEO, Aravind Srinivas, let slip something that shouldn’t be surprising but still feels jolting: his engineering team cut prototype dev time from days to hours by using AI tools like Cursor and GitHub Copilot. That’s not the headline. The headline is that their operations team is building prototypes too.
We’re not just accelerating engineering anymore. We’re reassigning it.
AI Doesn’t Just Speed Up Code, It Expands Who Writes It
This isn’t the first time we’ve seen non-developers experimenting with AI-powered dev tools. The accessibility of natural language coding means anyone who can describe a feature can now sketch its skeleton. But what’s striking in Perplexity’s case is the organizational normalization of that behavior. When everyone prototypes, the prototyping bottleneck vanishes. So does the gatekeeping.
Engineering’s monopoly on “build” is ending. AI is making prototyping a cross-functional activity. Ops teams, product managers, even finance analysts, anyone can wire together a rough feature idea. The cost of iteration has collapsed.
Which sounds great. Until it doesn’t.
Tooling Speed Creates Process Debt
Perplexity’s engineering adoption rate for AI coding tools jumped from 61 percent to 90 percent after the shift. That kind of growth usually signals real productivity gain. But it also comes with new complexity. According to Srinivas, bugs cropped up, not because the AI tools failed, but because the org was shipping faster than it could coordinate.
This is the new AI dilemma. The tools work too well. They let teams bypass traditional engineering friction, which means they also bypass some of the safety rails. In this new world, devs aren’t always the authors of early code. Review processes aren’t always triggered. Shadow features get pushed into production. And the debugging becomes... collaborative.
Which means engineering managers aren’t just overseeing code quality anymore. They’re overseeing code authorship.
Engineering Is No Longer the Gate. It’s the Filter.
When product and ops teams can ship early prototypes without writing a line of code, engineering becomes the editing desk, not the newsroom. This might sound like a demotion. It’s not. It’s an evolution.
But it changes everything.
Teams will need stronger observability, not just in prod but in pre-prod. Version control isn’t just a developer tool now. It’s a system of truth for the entire company. And that means AI-enabled tooling isn’t just about coding assistance. It’s about organizational hygiene.
Because when everyone builds, someone still has to clean up.
Build faster with LATAM engineers who get AI
Hiring great developers used to be the bottleneck. Now it's about finding people who can move fast, collaborate across tools, and co-build with AI.
TECLA helps U.S. tech teams hire senior-level, English-proficient engineers from across Latin America at up to 60% less cost, and in your time zone.
Want to see how fast you could scale?
What Engineering Leaders Should Ask Right Now
Do we know who’s writing code in our org?
Are our review and merge policies scoped for AI-accelerated volume?
Are non-engineers pushing prototypes into environments where security, performance, or cost could be impacted?
These aren’t philosophical questions. They’re operational ones. AI didn’t just speed up prototyping. It broke the boundaries around who participates in building software.
That’s exciting. But only if you’re ready.
More to come...
Recommended Reads
✔️ ‘Vibe Coding’ Has Arrived for Businesses (The Wall Street Journal)
✔️ How to Guide Non-Technical Colleagues Who Want to Write Code (How To Use Linux)
✔️ No code? No problem: How non-developers can start building with AI today (Inicio Digital)
– Gino Ferrand, Founder @ Tecla