Where Talent Hides

If you want to catch trout, don’t fish in the bathtub.

—Popular saying

1. Rough awakening

I hit the submit button on AngelList, finally uploading the job description (JD) that I had spent an hour putting together, format entirely cargo-culted from some trendy tech company’s careers page. For the first time in my life, full of hubris and hope, I was advertising a job for a company I had started.

I figured getting someone to apply couldn’t be that tricky. We were a VC-backed startup in San Francisco doing interesting mission-driven work. I had read and re-read dozens of Paul Graham essays, I was pumped from the pep talk at the latest startup founder meetup, we had money in the bank. Who would be crazy enough to turn down this opportunity?

A trickle of applicants came through. Half hadn’t read the listing. The other half was uninspiring. Not the caliber you need for your first few eng hires, the ones who lock in the DNA of the company for the rest of its life.

Ok, must be the JD. I read every blog post and whitepaper on writing better job descriptions. Stack Overflow had a great one, unfortunately long-ago lost to the sands of Internet time. Made it clearer, sharper, friendlier, more human. Reposted it on Hacker News Who’s Hiring.

Again, crickets.

I tried hiring a recruiter, a personal connection. I had never worked with one; I had no idea if they were good, but I already knew this one. I was given a few dozen candidates. Meh, I didn’t feel like these were quite the profile. I wanted the functional programming whiz rotating monoids in the category of endofunctors, not Mr. 15 Years of Java at Oracle. I gave feedback, but still nothing great. It didn’t feel like they got us, the sort of vibe and mindset we were trying to get more of.

I tried employee referrals next. Got a few warm intros from our tiny team. Exhausted that well in about two weeks.

Maybe the title was wrong. Maybe the benefits weren’t compelling enough, or I needed to be quirkier, or more specific about the tech stack. I was always one more perfectly-worded tweak to the listing away from cracking it.

Then one week, two things happened almost simultaneously. I went to a small local SF meetup for functional programming, not to recruit, just to hang out and shoot the breeze. I mentioned to someone between talks what we were building. They emailed me the following week asking if we were hiring. Their profile was phenomenal.

Around the same time, a teammate’s blog post about a gnarly technical problem we’d solved got some traction online. Someone who read it reached out cold. Also phenomenal.

Neither of these people had come through any channel I had been deliberately working. They came from places I hadn’t really thought of as “hiring channels” at all. And yet there they were, exactly the kind of people I’d been struggling to find.

It wasn’t that the job boards and recruiters were useless; they’d eventually produce good hires, too, once I got better at the whole game. But I’d been pouring all of my energy into two or three channels while ignoring an entire universe of others. The problem wasn’t where I was fishing. It was that I was ignoring many other interesting spots that offered perfectly viable alternatives.

Every startup CTO past the “two guys in a garage” stage faces the need to find talent to realize their company’s bold vision. Great people are just about everywhere, but you have only so much bandwidth to explore the different avenues, and many of the roads you take will net rather little. Not wasting your time and capital on dead ends, while other companies at your stage are gobbling up the remaining talented folks on the market: that is a life-or-death skill for every CTO. Finding great people takes lots of luck, most of which you have very little say in. But increasing the surface area of luck? That is entirely within your control.

So, with limited time and heavy competition, how do you find the right fishing holes without burning through your budget and sanity? There are many options available, most of which I have direct experience with as a startup CTO. They fall into three broad categories: fast and transactional, slow and compounding, and alternative pools.

2. Fast, transactional

With these, you can start getting results almost right away.

Recruiting agencies, AKA recruiters

Recruiters are what founders and CTOs will often instinctively reach for when it’s time to get the hiring funnel filled. The logic is seductive: find a pro, hand them a brief, watch the perfect candidates roll in. You’ll pay 20–25% of first-year salary, sure, but think of all the time savings!

In practice, tons of startups have negative experiences with recruiters, and many swear off ever using them. They try to communicate what sort of person they’re looking for, but often get sent generic applicants who aren’t any different from the hundreds they’ve passed on so far. You give feedback. You get more candidates. They look basically the same. You give more feedback. Nothing shifts. A few rounds of this and you start feeling like you’re doing donuts in a parking lot. Eventually, someone ends the relationship, usually you.

Who’s to blame? Honestly, usually both parties: the CTO and the recruiter. The founder communicates vibes instead of specifics. The recruiter pattern-matches to their existing rolodex. Neither party pushes hard enough on the mismatch, and everyone moves on quietly disappointed.

I’ve heard the claim that 95% of recruiters aren’t worth the time from founders who’ve cycled through many shops before finally landing one that clicked. Those lucky few swear they’re getting tremendous value out of the partnership. For them, the partnership was ultimately worth it, but it took many tries to get there.

How do you find a great recruiting shop? Same as with most consultancies: you ask around. Personal recommendations from people you know. Check with past clients. Does anybody swear by them? How long have they been in the industry? Have they consistently delivered great results for companies like yours? Are they able to tap into sources of talent you might not be reaching with your other channels, increasing the surface area of luck, rather than doing more of the same?

It’s worth at least experimenting with; maybe you get lucky and find a good one. But don’t feel bad if it doesn’t work out for you; it’s a common experience in the industry. Move on to other sources if it doesn’t. Some companies get hooked on the convenience of not having to build their own talent acquisition muscle and keep trying to make recruiting shops work when they should be trying something else.

Employee referrals

Your employees likely know other people just like them, with all of the pros and cons of that. They might be their college roommate, their hackathon partner, a former coworker, or the mullet guy from the Hayes Valley AI hacker house.

If you liked the archetype of your employee, you can likely get more just like them through those networks. So far, so good, but a team built entirely from the same mold will develop blind spots fast. There are other real limitations to this approach as well.

First of all, your employees only know so many truly stellar people, and you can exhaust that network faster than you’d expect. Shouldn’t stop you from reaching out to them, shoot your shot, but the odds are middling. Doesn’t mean skip it, just don’t make it your primary strategy.

Secondly, a referral is not a vetting. Early-stage employees, even great ones, often haven’t yet developed a fully calibrated eye for talent. They’ll recommend people they like, people who remind them of themselves, people they’ve had fun working with. But that’s not the same as recommending someone who will thrive in your specific environment at this stage.

The candidates still need to go through the usual checks to make sure they meet the bar. Folks will often recommend people they personally like, but not always the ones who will also be successful at your company. That’s ok, that’s why we trust but verify. I’ve personally seen absolutely stellar employees recommend friends who were nowhere near the caliber of their endorser.

Specialized job boards

This has worked well for me over the years. Generic job boards and LinkedIn are fine, but they also lead to generic applicants and a worse signal-to-noise ratio. The people whom I want to hire often dread opening the site, and would rather use something more off-the-beaten-path. You have to filter through more people who are flinging resumes at everything, rather than people who are specializing in the thing you are looking to hire for. Niche boards aren’t a panacea, but they let you zone in faster. When hiring in Clojure and Haskell, we had a good time working with boards like Functional Jobs and Functional Works, or using smaller communities like Functional Programming Slack to reach the people who wanted exactly what we were offering. When hiring in early-stage game dev, boards like Work With Indies or Remote Game Jobs are great intersections of multiple candidate segments, such as gamedev + early stage or gamedev + remote work. You can also focus on work style boards like We Work Remotely and Remote OK.

Newsletter job postings

These have become more popular recently. You can post your job on a newsletter like Lenny’s, Ben Lang’s, or something niche like Haskell Weekly if you know your target demographic reads those regularly. I’ve had mixed results with these in the past, but it’s an option; there are lots of ways in which this can work, and in which it’s a total waste of marketing dollars.

VC networks

If you’re part of a great VC network, you likely have access to extra resources that the average person doesn’t. For example, a large VC may have a talent team of people who are spending all day interviewing candidates who want to apply to portfolio companies of the VC and be matched with them. Someone is working full-time just to make sure that the people you end up talking to through this channel are worth considering and seem like sensible people.

Startup incubators

If you’re part of a big startup incubator, you can often hire other former founders who may be looking for an opportunity once their original startup didn’t work out. They may be quite happy to work with you for a couple of years, recuperate, regain some strength, bring to your company all the things they had to learn the hard way while building their own, and be already a little vetted due to being part of a network. They likely won’t stick around forever, but for at least a few years, you will have someone of quite high quality and fairly trusted come on board and contribute to the mission.

All of these are tools you can pick up and put down relatively quickly. The next set is different. You plant them now, you harvest later… if you have the patience.

3. Slow, compounding

There are no quick wins here, but these will pay for themselves over the long term.

Direct cold outreach

This works well as the CTO, but you need to be systematic about it, and it takes a while to reach the volume at which it starts to pay dividends. This is just like any other enterprise sale. You reach out to folks, let them know that you want to connect and build a relationship, learn more about them and their work, and keep them in the loop as far as your company goes. This is a long game for you for candidates you’d be excited to bring on board one day. It’s insurance for them: someone who personally cares to get to know them wants to build a relationship and provide them with an option to exercise if things go south at their current role.

The main downside of this one is that it takes time and organization, so most short-term-minded CTOs don’t do it. You need to make time to meet a few interesting potential candidates every week, you need them in a CRM so that you can track your touch points, you want to get updates about their status on LinkedIn and X so that you can intervene if their employment situation ever changes, and you want to touch base on the regular to remind them that you exist. Not sustainable without a system and some automation in place, but also, this may work better than many other approaches because it’s about that personal connection, going beyond being yet another JD in a crowd of startups telling you how they’re the next great thing.

I’m only aware of a couple of CTOs in the early-stage scene that do this, but they all swear by the method. It’s also why many top founders end up spending over 50% of their time on hiring; this stuff doesn’t scale that well for the people on top unless you want to let go of the wheel, which is generally a bad call when you’re in the rapid growth phase of your company.

Content marketing and thought leadership

This is tried and true. You do interesting work, you talk about it. You have your team talk about the interesting hurdles you had to overcome and what clever solutions you invented to get around them. It doesn’t have to be crazy innovative or fresh, you don’t have to be the first to solve a problem, and there are likely hordes of people who will enjoy reading about your engineering adventures because they missed all the other posts about a similar subject. The Internet’s short term memory is your opportunity. It’s a great chance for someone on your team to get their name out there, get some visibility, and build a reputation, while your company builds its own brand as serious professionals in the game.

You have probably read your fair share of company blogs about their best practices, how they hire, how they ship, and what lessons they have learned from both wins and losses. “We Moved to Microservices! Here’s Why and What We Learned!” “Nevermind, LOL! We Moved Back to the Monolith. Here’s What We Learned!” Either way, you now have content. We all love a good postmortem that we can pass around as a cautionary tale, great stuff.

They become “thought leaders” in the industry. They gain prestige from the visibility, attracting those who want to be part of the prestigious club. But at the base level, they simply let the world know that you exist, increasing the chance of someone discovering you. A top Hacker News post, a viral X tweet, and a great video are all excellent paths to visibility and relevance.

Honesty and humility will go a long way here. Talking openly about what didn’t go according to plan, what ended up surprising you, and what humbled you in the process of experiencing that story arc will win readers over instead of feeling like they’re reading someone bragging about how great they are. That vibe won’t win over many or make them want to connect with you on a personal, emotional level. Be human.

Running open-source projects

Open source is another solid way to get more visibility for your team. If you’ve solved a problem that many others are bumping into, and you’re doing it in a way that’s interesting, unique, and opinionated, you may as well share it with the world. Often, you’re sitting on interesting modules of functionality that you suspect the outside world might benefit from. Still, you don’t bother releasing them because you’re not sure how that’s going to actually benefit you. Now you have an excuse. All of the same rules from content marketing apply here.

Podcasts, talks, and social presence

These all make a big difference. Appearing on a popular technology industry podcast, say Lenny’s Newsletter or The Pragmatic Engineer, is a big deal for both the team and the individual as it’s the validation of your relevance on “the scene” by a trusted tastemaker. Countless people pay attention to what the top industry influencers are talking about and showcasing, and if your team is there, you’re being legitimized as well.

Doing a recorded talk at a conference, think AI Engineer or the Game Developers Conference, makes you legit. You did something that the community wanted to hear about enough to put you on stage. It’s a big deal. Or people simply talk about you on social media as a company that’s doing something interesting and relevant, that’s again both validation and visibility.

Local and online events, meet-ups, game jams, hackathons, workshops

Organize these and become the regular ritual in your area for a particular interest group. That may be game dev. That may be your language of choice. That might be a particular framework or perhaps a get together for industry professionals from your vertical. If you have the scale and the resources, you can take it as far as even hosting your own conference; the sky is the limit.

It should be relevant to what you’re doing, but at the end of the day, just getting people with similar niche interests together to learn from each other, teach each other, demo for each other, and meet other people doing interesting work on the frontier is great. It allows you to build a mailing list of people who may be relevant to your company. They get to learn more about you, what the team is working on, as the sponsors of the event, and you get to learn a bunch about them. You get to display the interesting work you’re doing at your company to everyone there, and they may then decide they want more of it and do it alongside you, thanks to you being congenial, human and personable, and doing work that is up their alley.

But it doesn’t have to be just related to work. Some companies in San Francisco simply throw together great parties at their offices, and they ask their staff to invite their smart friends to these regular get-togethers. The founders will spend time getting to know everyone, learning more about them, and connecting with people who may be a really great fit for what they’re looking for, either immediately or months or years down the line.

The important bit is that smart people are exposed to the vibes, culture, and people working at your company. If they get along, if they’re impressed, if they really enjoy being around this set of people, and they themselves are pretty sharp and hungry, it’s a great way to take advantage of an existing network to keep growing your ranks.

One rather successful early-stage startup in San Francisco became known for hiring a ton of their employees’ roommates who just happened to be living in the same big building they had all moved into. They were all fairly young and fresh out of college, and they wanted to keep those college/dorm-room vibes going by hiring neighbors in the building. Getting more and more people on board this way seems to have worked pretty well for them, albeit at the cost of the staff all being roughly the same archetype.

It’s still working for them, for now. But homogeneous teams tend to build homogeneous products, and at some point, the monoculture becomes the ceiling. They’ll want to diversify eventually… the only question is when.

4. Alternative pools

Often slept on, but it can be a powerful resource if used right.

Hiring outside of the US

Consider this: there are many markets out there that early-stage startups regularly sleep on. Eastern and Southern Europe, Southeast Asia, South America, you name them. If you can offer compelling compensation and reasonable work hours, you can lock in talent that no one else is even looking at. At Double Dusk, and with async-first companies I’ve watched up close, the default work style is fully asynchronous and averse to most meetings. You can still meet ad hoc if you want to, but by default, everything is done over shared documents across the company, over Loom recordings, and in your main communication chat app (like Slack or Discord).

This allows people in time zones all over the world to participate in building the company regardless of where they are, their obligations, or how many contiguous hours of work they’re able to pull off. This unlocks a whole new pool of talent that you wouldn’t have had access to otherwise.

When everyone’s remote and async, it’s easy for the team to slip into clock-in, clock-out mode. No hallway energy, no spontaneous celebrations, no shared sense of urgency. You lose the feeling of chasing something together. That’s fine for a regular business, but a death sentence for a winner-take-all market startup with 18 months of runway in the bank.

Getting people genuinely stoked at the same time, across time zones and Slack channels is one of the hardest leadership challenges in a distributed startup. You’ll get work done, just less of it than you’d expect in person. The quality, variety and density of innovative ideas will drop as well. We don’t have great data on this gap yet. Vibes, for now, are the only metric.

Contract-to-hire

Some CTOs swear by contract-to-hire engineers as a protracted work trial; they claim they have found great talent that way. I have mixed feelings about this; I think it really depends on the market. In hot markets, people who are okay with spending several months contracting for the potential of being hired one day are not going to be the best in class. Everybody is getting great full-time offers with great benefits, and the great people get snatched at a moment’s notice.

If someone is on the market in that type of money-printing environment and is okay with contract-to-hire for months at a time, it’s either a deliberate lifestyle choice—understandable—or because they’re not that great, in which case you should pass. In Hiring Winters, though there is a ton of great talent looking for a role, because nobody is hiring, they’re ok with getting a contract gig with you to see what happens.

In my experience, you can usually tell within a couple of days of trialing if someone is a real killer; you don’t need three months. I don’t find it necessary to wait that long to decide if someone is a great fit for what the team is looking for.

Open-source contributors

People with plenty of solid work on GitHub can be a great source. These are folks who have demonstrated the ability to write code, respond to feedback, and work tirelessly on projects that they’re passionate about.

The one caveat that I’ve personally seen is that being a great open-source contributor doesn’t always translate to being a great commercial software team player. Commercial software demands a different muscle: tight deadlines, real constraints, constant trade-offs, and a bias toward shipping over perfecting. That’s a different mode than most open-source contributors operate in. Their skill set and work style, one that involves working mostly individually at their own cadence, may not translate well to a dynamic team-based environment.

Social media

Looking through social media for indie developers and up-and-coming builders who are creating lots of interesting projects in their spare time and actively talking about them can work really well. X, Hacker News, Reddit, YouTube, all places where talented people showcase their work. Those folks demonstrate not only the ability to build but also the ability to talk about it and explain why it’s interesting. They are part of that very small percentage of people who are passionate enough about their work to go out of their way to do more of it and showcase it to the world for the world to judge. Even better if they’ve attempted to monetize and turn several of these projects into cash flow machines, regardless of whether they succeeded or not. They showed a desire to prove themselves and to do what’s difficult, rather than do the absolute bare minimum to stay employed at a large organization. They’re the kind you want at a scrappy early-stage startup that needs maximum agency.

Hiring your existing customers

Sometimes you can even hire your customers. They might be really excited about your product, and they’ve had nothing but positive interactions with the people on your team. Perhaps they were early adopters, or design partners, or people you got to work with for months or years as you were building out the product. They built a real personal connection with you as the CTO or with folks on your team, and one day they decide they want to be part of that journey themselves.

It’s a great way to get someone who genuinely cares about the product and the mission. That enthusiasm is night and day compared to interviewing someone who’s hearing about your company for the first time. Just make sure they’re the ones initiating the conversation. Actively poaching from your customers’ teams is a fast way to torch a business relationship, lose the customer entirely, and build a reputation you don’t want to acquire.

5. The bigger picture

Ultimately, none of this matters if your company is a rocketship everyone wants to board. If you’re an early Anthropic, OpenAI, ElevenLabs, or Mercor, a company the market has decided is about to take off, the sophisticated tactics in this piece become largely optional. Product-market fit, investor momentum, and talent density create their own gravity. The universe is on your side. Smart, hungry, talented people are banging on your door to get on the spacecraft before it departs. There’s a self-reinforcing loop to this: the more great people join you, the higher the chance of success, the more other great people want to join you, at least until the growth tapers off and there isn’t as much pie to distribute. Just the cycle of life in the world of VC-backed startups on the frontier of major market changes.

For everyone else: finding and attracting talent is sales and marketing at the same time. As I mention in Hiring - The Big Picture, as a startup CTO, you’re always pitching two products: whatever makes your company money, and the company itself. Both are equally important. Both often require outbound and inbound activity to get people into the funnel.

The product you’re selling is complex, so you have plenty of dimensions to describe it. You’re not only offering a paycheck, but also belonging, identity, purpose, a tribe, camaraderie, a way of working, an opportunity to deepen one’s mastery and being part of an adventure. Just like with any other product, if you’re not crisp about who you are and what someone is going to feel like once they acquire said product, you’ll get a lukewarm response and lukewarm applicants.

The brutal truth, as the LLMs like to say, is that there’s no standard, spammable move that will net you great access to talent. Every employer on the market is looking for alpha. They want great talent that they can get the best deal on. If you’re doing what everybody else is doing, you will get what everybody else is getting: perfectly average results. Maybe ok for a stable, big co that can keep exploiting a dominant market position. This works in the short term as you’re coasting on momentum, but you can’t sustain that forever. At least not as a startup that needs to generate a lot of high quality work on barely any cash flow and a disappearing runway.

Where you look for talent and how you stand out have to be unique to the company stage, the market cycle, the company trajectory, who you are as a team, as founders, as a CTO. What works miraculously for one firm might feel wrong for you, and vice-versa. And where you search for greatness also evolves with your hiring taste, and the evidence you collect about what’s working, and what’s no longer bearing fruit.

Talent channels are effectively growth channels. And just like with any growth channel, you have to run experiments on them, decide if you should pivot or persevere depending on the results, and eventually move away, as every channel inevitably hits The Law of Shitty Clickthroughs, the idea that every growth channel degrades over time as it gets discovered, crowded, and gamed. Every channel will eventually be overfished, by you or your competition, but as an early-stage startup you likely don’t have the volume of hiring necessary to tip those scales just by yourself. And over time, channels that used to work may fall out of fashion for myriad reasons outside your control. That’s just the nature of the game.

The key is to identify a few channels that meet your needs for the foreseeable future and stick with them until something changes. In retrospect, I wish I had spent far more time on building personal connections with promising future hires through in-person events, direct outreach to people doing interesting work, and social media thought leadership. It would have made hiring pushes feel like less of a lottery, and more akin to going through a list of warm leads you’re already excited about.

Experiment with these few channels, treat them as a portfolio of sources, and assume you’re going to eventually rotate some of those sources out. Some will bear more and more fruit until saturation, some will never take off, etc. None of the channels, individually, will typically do miracles; the strategy is to pick a few you can reasonably sustain over time, get a feel for how the market responds, and iterate from there.

It’s also worth remembering that being in a Hiring Summer vs Hiring Winter will dramatically impact the volume of candidates you will receive. A rough year for employees will make your life much easier, and a rough year for employers will force you to work much harder to get the attention of those who you really want to bring on board.

In What The Best Looks Like I talk about what that talent looks like and how to identify it. In Telling Your Company’s Story, I talk about how to tell the tale of your company’s journey in a way that will attract the right people to the gates and get them to want to join the crew on an adventure.

6. What to do with all of this?

Now it’s your turn to venture out there and experiment. Nobody can tell you in advance which of these channels will resonate with the world and showcase your company’s unique identity. You have to try it yourself. Every company is unique and what worked for one will not work for another.

Pick two or three channels that feel right for who you are right now, commit to them for a few months, and see what the data tells you. Pay attention not just to volume but to the quality of who’s showing up. If a channel is too costly, or isn’t bringing you the kind of people you’d be excited to work with, rotate it out and try something else. The portfolio evolves.

Looking back at my own story, the two best candidates I found early on didn’t come from any channel I had deliberately set up. They came from a meetup I attended for fun and a blog post a teammate wrote because they thought it was interesting. I wasn’t recruiting in either case. I was just being present, doing interesting work, and talking about it. That turned out to be the channel.

The best sourcing strategy is the one that’s uniquely yours. Go find it.