A hiring funnel is guaranteed to land you mediocre people; 10X only exists in teams
Treat hiring like a sales funnel and you don't get the strongest 20% — you get the 20 people who happened to reply that day. Adam Ward says talent density is a collective property: there is no 10X individual, only a 10X team.
The video won't play here. Listen to the audio instead:
The argument · tap a timestamp to hear it
A hiring funnel is guaranteed to land you mediocre people
Treat hiring like a sales funnel and you start from "100 people, 20 reply" — and those 20 aren't the strongest 20, they're the 20 who happened to be in a bad mood that day. Then each stage filters a batch out, and you hire whoever's left. Adam calls this the funnel of doom; it's essentially remainder hiring: you're not picking the best person, you're picking the person who didn't get filtered out. At scale it regresses to the mean, and all you can hire is "slightly above average." The right approach is to first confidently identify who the top 20% are, then use an assessment process to validate that judgment — not use a process to discover people.
— Adam WardDon't ask who's best, ask who's best at one thing
When sourcing top talent, the worst question is "who's the best product engineer you know." What you should ask is "of the product engineers you've worked with, who collaborates best with designers" or "who can translate a framework into product better than anyone you've seen." Put the specific trait the team actually needs into the question, and it's much easier for the person to recall a concrete name. Once several credible people are asked the same question, names start to overlap — Lenny mentions him, Sally mentions him, and that person goes on the target list. Adam calls this triangulation, and says it's far more effective than vaguely asking "who's the best."
— Adam WardRecruiting shouldn't carry the hiring goal
Adam's pet peeve: many companies put the hiring goal on recruiting, but recruiters don't make the hire decision — the hiring manager or a committee does. The goal and the power are misaligned. He says a recruiter is a confidence engine, not a decision-making engine; the job is to give the hiring manager the confidence to make the right call. Nobody benefits more from a good hire or suffers more from a bad one than the hiring manager and the team, so accountability should sit with them. He also quotes "what gets measured gets done": if you don't measure it, the responsibility gets pushed away.
— Adam WardDon't ask for an interview, ask for the next conversation
The most common reply from top candidates is "the timing isn't right." Adam's move is to downgrade the ask: I'm not asking you to interview, I'm not asking you to consider the role, I just want to set up the next conversation. If they say talk to me in the fall, then book a coffee at the office first and introduce them to people on the team who are interested in their work. Even if they genuinely can't start until the fall, that doesn't mean you can't stay in touch every week for those months — do nothing and the best outcome by fall is that they're still interested, or someone else has already snapped them up. Cursor has cases where they talked to a candidate for a full year.
— Adam WardWork samples are the top predictor, yet most companies still do one-on-ones
Adam cites the research finding: work samples are the strongest predictor of how well a candidate will succeed once hired, yet most companies still spend their time on one-on-one desk assessment. Cursor's approach is to have candidates do real projects and work side by side with the team, which gives both the company and the candidate a strong signal. He criticizes most companies' onsites as one-directional — "we're assessing, we're extracting information, and we give almost nothing back to the candidate." He believes in the power of self-selection: good self-selection comes from good inputs and data, so the onsite should be designed to be two-way, including who greets them and who sits next to them at lunch — all of it deliberately chosen.
— Adam WardCut the work trial and the signal collapses
Adam mentions an interview with Michael Terrell: they once tried to remove the work trial because it was too heavy a lift and many candidates didn't want to do it, and "the signal completely collapsed," so they had to add it back. Adam says Cursor ran the same experiment, because a work trial really is expensive and a big ask for the candidate, and after removing it "we had a hard time being confident in our decisions." His conclusion is that it's a confidence-building function, so they brought it back. He adds one qualifier: there aren't many absolutes in this work, but for a lot of companies it's hard to think of a better way to get signal than working side by side on real work.
— Adam WardOne Slack channel per candidate, a stand-up every day
Adam says closing is a team sport. For high-profile, high-caliber candidates, the team often holds a daily stand-up on how to land the person — it might only be 10 minutes, but the point isn't the length, it's the focused energy invested every day around "what are we collectively doing." Participants can be the hiring manager, the recruiter, or the candidate's future colleagues. Every candidate gets their own Slack channel for strategy syncs and updates. He gives a concrete example: a candidate in New York who went to a performing arts high school and is a classically trained violinist — the team found a specialty instrument shop in New York, bought a few-hundred-dollar Turkish guitar-like instrument, and gave it to him at the offer dinner.
— Adam WardThere is no 10X individual, only a 10X team
Adam thinks companies over-focus on the individual when it comes to talent density. Unless the team is one person, talent density is collective — the numerator and denominator aren't 1 to 1, they're 1 to however many people. He says great managers are great puzzle-solvers: they assemble complementary teams and then help them perform. He states it plainly: "I don't think a 10X person can be on a bad team — that just doesn't exist." A great person who isn't on a strong, mutually supportive team gets throttled. He'd rather think about the team's 10X than the individual's 10X.
— Adam WardThe biggest early-stage mistake: cramming two conflicting skills into one head of talent
Adam says the most common early-founder mistake is looking for a silver bullet — the first head of talent — and merging two conflicting skill sets into one role. By the time a founder is ready to hire their first recruiter, they're usually already behind on hiring, with a backlog of 5, 10, 15 roles; but at the same time they want this person to build systems. Those two demands conflict: someone who wants to prove themselves will make hires first and not build systems, while someone who wants to build systems may not make a single hire, and the founder can't sleep. His advice is to decouple: you can hire just one person, but get outside help for the other direction — pair the systems-builder with a contingent recruiter for short-term momentum, or build the foundation first and then hire a high-energy IC recruiter.
— Adam WardThe gap between offer accepted and start date is a no-man's-land
Adam says more and more companies get stuck after the offer is accepted — the more senior the person, the more often the window between accepting and actually starting exceeds two weeks. The industry habit is "they said yes, let's go hire the next one," but there are actually two things to do in that window: make sure the person really shows up, and pre-board them so they ramp faster once they start. Cursor's approach is dinners with other people who've accepted offers, shipping a Cursor laptop to them, and using the posture of "you're already part of this" to create a sense of community, so people start imagining themselves at the company ahead of time.
— Adam WardEvery recruiter should be a talent engineer
Adam has observed a new function emerging on recruiting teams — the talent engineer, often a former engineer, who builds tools for the hiring team. His warm take (not hot take) is that every recruiter should be a talent engineer. The reasoning: it used to be that to build a tool or a feature you had to beg the engineering team for roadmap time; now recruiters can build things themselves in tools like Cursor. He wants his recruiting team to no longer depend on one person to help build tools, but for everyone to have that ability.
— Adam WardModels iterate daily, so switching cost became a product selling point
Adam says one of Cursor's most popular features is how easily you can switch between models: you're not locked into Codex or Claude Code, and when a new Opus or GPT comes out you can switch over and try it immediately. He notes that models used to ship a version every three to four months, and the interval has now shrunk to days, and people who love building product want to use the best tool at any given moment. Cost is also a legitimate company consideration now, so optimizing to "use the right model for the right task" lands right in Cursor's sweet spot. He also mentions that Cursor now trains its own models, running in SpaceX data centers.
— Adam WardIn their own words · checked verbatim
I am going to reach out to 100 people. 20 people will reply just for round numbers. So 20% of those will reply. By definition that is not the top 20% that is just the 20 people who you caught on a bad day for them.
Adam Ward13:13
the mistake that most companies make is they'll say, "Hey Lenny, who's the best uh product engineer you know?" That is the worst question you can ask.
Adam Ward20:18
research has shown that work samples are the highest predictor of success at a company. But yet most companies are still spending a lot of time doing one-to-one across the table assessments.
Adam Ward45:38
I don't think there's a lot of maximums or absolutes in this business.
Adam Ward50:40
I don't think you can have a 10X person on a bad team. It just like doesn't exist.
Adam Ward1:10:53
But what are we doing between offer accept and join?
Adam Ward1:18:56
my warm take, not a hot take, my warm take is like every recruiter should be a talent engineer.
Adam Ward1:19:56
I will answer their emails. I will jump on a 10-minute call with them.
Adam Ward1:26:01
Figures
| How tight the hiring market is (Adam's score) | 11 out of 10 | 3:07 |
| Duration of the mobile talent shortage, for comparison | about two years | 4:08 |
| Company size change during Adam's time at Pinterest | 200 people to over 2000 | 2:02 |
| Engineering org size change during Adam's time at Facebook | 1000 people to over 10000 | 2:02 |
| First hires Cursor's team made themselves | about 80 people | 1:07:50 |
| Cost of the gift for the New York candidate | a few hundred dollars | 1:03:47 |
| Backlog of open roles at an early-stage company | 5 to 15 | 1:11:55 |
| Window between offer acceptance and start date (senior candidates) | usually more than two weeks | 1:18:56 |
| Change in model release cadence | from every three to four months to days | 1:29:04 |
Glossary
- funnel of doom
- Treating hiring like a sales funnel, so the person you hire is whoever didn't get filtered out rather than the strongest candidate.
- remainder hiring
- Picking from the candidates left over after the process filters people out, rather than locking onto the strongest people first.
- work sample / work trial
- Having candidates do real projects and work side by side with the team, as the highest-predictor signal.
- talent engineer
- A new function on recruiting teams that builds tools, often filled by a former engineer.
- talent density
- The proportion of high performers on a team; Adam argues it's a collective property, not an individual one.
- pre-boarding
- The actions between offer acceptance and the start date that ensure the person shows up and ramps faster.
How to listen
Founders and hiring managers building an early team or owning hiring, plus HR people and recruiters who want to understand how a high-talent-density team actually operates.
From 1:22:57 they talk about ER dramas and personal creeds — unrelated to hiring methodology, skippable.