The startups building AI tools that turn natural language prompts into software — the so called vibe coding sector — are spending CEO level personal time courting engineers, because the small pool who can actually ship AI coding tools is
On a recent afternoon, Emily Cohen, head of people at the AI coding startup Cognition, drove a candidate to San Francisco International Airport. The drive was not a recruiting perk. It was part of the pitch. Cohen wanted to be the last voice the engineer heard before his flight home, the impression that lingered while he weighed competing offers from rival startups building the same kind of tools.
That level of personal courtship is the defining feature of hiring inside the "vibe-coding" sector, a small cluster of billion-dollar startups building AI tools that generate software from natural-language prompts. The category includes Cursor, the AI code editor; Cognition, the maker of the Devin coding agent; Replit, the browser-based coding platform; and Base44, a vibe-coding platform owned by Wix. These companies share a thesis that software creation is about to be commoditized, and a recruiting practice that treats the humans building the commoditization as irreplaceable.
The contradiction is the story. AI coding tools are pitched as products that let anyone ship software by typing a sentence. The companies making those tools are simultaneously treating their own engineers like scarce draft picks, with chief executives personally scanning X and GitHub profiles, head-of-people staff logging flight times, and former employees describing the outreach as a "love-bomb" campaign.
At Cursor, the practice starts at the top. According to a former employee cited in a recent Business Insider feature, chief executive Michael Truell personally scans X and GitHub profiles for undiscovered engineers, former founders, and staffers from companies like Notion and Figma. Outreach begins with a Truell-authored email and can escalate into what the same source described as love-bombing: a rolling series of coffee invites and office drop-bys that read less like recruiting and more like a personal campaign.
The mechanism behind the courtship is supply-side scarcity at the build layer. The pool of engineers who can ship production-quality AI coding tools is small, and the demand for them is concentrated in a handful of well-funded competitors. Traditional résumé-based hiring is too slow and too noisy for that market. Candidates worth hiring are not applying through portals. They are already employed, often at peer companies, and they are evaluating multiple offers at once. The companies that win them are the ones willing to spend chief-executive time and headcount on individual outreach.
That spending shows up in unusual places. Some candidates go through multi-week work trials rather than whiteboard interviews. Others are run through company-specific "bootcamps" designed to test for the judgment that does not appear on a résumé. Internal Slack channels at these firms carry names like "Talent Spot," where engineers are encouraged to flag potential hires from their own networks. Token burn, the practice of running expensive inference workloads on a candidate's behalf to evaluate real product thinking, is treated as a recruiting cost rather than an engineering cost.
The courtship economy is not without critique. Former employees at Cursor have used the word "love-bomb" to describe the cadence of attention, and the practice of bypassing résumés raises obvious questions about whose work gets seen and whose does not. Personal-campaign hiring is biased toward people with visible online footprints, prior brand-name employers, and existing networks inside the small circle of AI coding investors and founders. The source reporting flags both the intensity and the fairness risks. The companies themselves frame the practices as necessary in a labor market where the alternative is losing the build layer to a competitor.
The watch item is whether the courtship economy survives the next round of funding. Vibe-coding valuations have reached the tens of billions; Cognition alone is reported at roughly $26 billion. If capital tightens, the chief-executive time and head-of-people staff now spent on individual courtship will be among the first line items to compress, and the labor market for AI engineers will look more like a normal tech hiring cycle. If capital does not tighten, expect the airport drop-offs and the founder DMs to scale with the rest of the sector. The defining question for the vibe-coding boom is not whether the tools can replace software engineers. It is whether the companies building the tools can keep hiring the humans who make them work.