Cursor人才主管谈招聘：工作试岗是最强预测指标

My biggest takeaways from @wardadamp (Head of Talent at Cursor): 1. The forward deployed engineer i...

My biggest takeaways from @wardadamp (Head of Talent at Cursor): 1. The forward deployed engineer is the hottest role in tech right now. These are deeply technical people who can partner with sales, be confident in a room with executives, translate complex products into business outcomes, and help customers shift from “token maxing” to optimizing AI bills. Adam likens this to the mobile-engineer hiring boom—except that one played out over two years, and “now we’re in days or weeks.” 2. The traditional recruiting funnel, what Adam calls the “funnel of doom,” is designed to build mediocre teams. The standard hiring funnel (reach out to 100, 20 reply, weed out at each stage, hire who remains) rests on a broken premise: the 100 people who reply in the first place aren’t the top 20%. They’re whoever you caught on a bad day. By the time a candidate gets to the bottom of the funnel, you’re selecting from an already-suboptimal pool. Even with an rigorous interview process, this leads to average teams. 3. The alternative is treating every hire like an executive search: rigorous scoping (what does “great” mean for this specific role?), deliberate candidate mapping (who are the 50 best people in the world for this role?), and relentless pursuit (go after these 50 people and don’t stop). The shift is starting with a “pillar of excellence” rather than a wide funnel. 4. The new danger zone is after the offer is signed. Candidates reneging on accepted offers is a rising trend, and senior hires often sit for weeks between signing and starting. So Cursor treats the post-accept period as its own campaign, with dinners alongside other incoming hires, a laptop shipped before day one, and a community forming before anyone officially starts. It makes sure people actually show up, and it gives them a head start on ramping up. 5. Work trials are the highest predictor of success. Research has consistently shown this to be true, yet the default on-site is still a series of one-on-one conversations across a table. Cursor runs long on-site experiences where candidates work side by side with the team on a real (or realistic) project. That format generates technical signal, core-values signal, and collaboration signal simultaneously—and it gives the candidate data to self-select. When Cursor ran an experiment removing the work trial, hiring confidence dropped significantly. They brought it back. 6. Asking your network “Who’s the best engineer you know?” is a trap. Instead, ask hyper-specific questions tied to your scoping details instead—“Of all the product engineers you’ve worked with, who’s most collaborative with designers?” or “Who can translate a framework into a product better than anyone you’ve seen?”—because specific prompts make names easy to conjure. When multiple trusted people independently mention the same person (“Lenny mentioned them, but so did Sally”), you’ve triangulated a target for your top-50 list. 7. Scoping the role is the most under-invested step in most hiring processes. “I’ll know great when I see it” is a trap. Without spending time up front to define and stack-rank the skills, experiences, and attributes that you need in this specific role at this specific company, you have no sourcing strategy, no pitch, and no way to design a real assessment. Everything downstream flows from this. And resist the logo shortcut: a strong team at a company at the right time is only “possibly” shorthand for great, and copying someone else’s definition means piggybacking on a hiring process you already believe is broken. 8. When a top candidate says “bad timing,” shrink the ask: “I’m not asking you to interview—I’m asking for our next conversation.” Adam plants seeds—a coffee, an office visit, an intro to a teammate fascinated by their work—that bear fruit over weeks, months, and sometimes years, because going silent until “late fall” means, at best, you get them if no one else has grabbed them first. Almost as much effort goes into finding a warm intro as into the message itself, since once someone ignores your first note, ignoring the second and third gets much easier. Cursor also plays “home games,” centering touchpoints on shared meals—“a disarming act and a joining act”—and office visits designed to trigger the campus-tour gut feeling of “this feels like home.” 9. Closing is a team sport, and it starts at the first conversation. When Cursor has a high-caliber candidate in process, there are daily standups focused on that specific person: What did we learn about their motivations? What objections came up? Who on the team can speak to those concerns? Every candidate gets a dedicated Slack channel. The “close” shouldn’t feel like a separate sales push but, rather, the natural conclusion of a process that has been consistently demonstrating the same answer. 10. Caring is free, and it’s the biggest unfair advantage in recruiting. In candidate-satisfaction scores across the history of the field, Adam says the number one driver is feeling that the company genuinely wanted you, so Cursor curates every touchpoint: who greets a candidate, who sits beside them at lunch, which interviewer shares their niche interest—handpicked rather than left to calendar availability. The reframe that makes it sustainable: you’re hiring one person for this role, not 10. “Don’t get caught up in the 10; focus on the one.” Lenny Rachitsky @lennysan Cursor has spent its entire existence competing with every major AI lab and incumbent in the toughest market in tech—and has continued to win. So much so that @elonmusk bought them for $60 billion to help SpaceX win the AI race. They've done this by building one of the most talent-dense teams in history. Adam Ward ( @wardadamp ) is Head of Talent at @Cursor_AI —who acquired his legendary recruiting firm after he spent 20+ years building the most elite teams in tech—and in our conversation he shares the entire playbook for building high talent-density teams. We discuss: 🔸 Why the traditional recruiting funnel—the "funnel of doom"—guarantees mediocre hiring 🔸 Adam's three-step playbook for hiring the top 1% 🔸 The rise of the forward-deployed engineer 🔸 Today's "tale of two cities" talent market 🔸 The biggest mistake founders make with their first recruiting hire 🔸 The worst question you can ask when sourcing talent (you've definitely asked it) Watch the full episode right here 👇 Your browser does not support the video tag. 🔗 View on Twitter 🔗 View Quoted Tweet 💬 5 🔄 1 ❤️ 12 👀 3984 📊 7 ⚡