The 14% pass rate: what we reject and why

14% success rate. That figure is stated on the home page. It's not marketing. It's the actual number of applicants who pass our screening. Here's the inside

Reject reason 1: no live URL

If you can't send a working URL of something you've built and sent, you've failed. A GitHub repo full of half-finished side projects is not enough. The product must exist on the open Internet, with users, or at least with the ability to handle one without falling over

Reject reason 2: cannot defend the diff

We pull up a recent commit and ask the candidate to walk through it line by line. If the answer is "Cursor wrote that part," then the call is over. Not because Cursor wrote it. Because they didn't read it. We want the person who treats every diff like a code review from a stranger

Reject reason 3: no taste

We open the product. If the spacing is wrong, the type contradicts itself, the empty states are missing, the loading states are missing, the candidate gets a polite no. Taste is not optional. It is the difference between a demo and a product, and our customers can tell the difference within 30 seconds

Reject reason 4: cannot ship without a team

We ask what the candidate built solo. If every story starts with "we" and every artifact has five contributors, that's a manager, not a vibe coder. We hire the person who can take a briefing at 9 a.m. and deliver it by lunch with no one else in the room

What the 14% have in common

They ship every week. They read their own diffs. They have opinions about the tools and can defend them. They've killed at least one of their own features because it didn't carry enough weight. They would rather work alone with AI than with mediocre people

Nor are they cheap. The 14% are at the top of the market because they have earned it. You can't skip the price by skipping the vetting process

Read the full vetting →