Pay more, hire less

Hey everyone, Neil here. You're reading High-Signal Hiring. Hiring systems from 20+ years of global recruitment experience and 500+ technical hires. Zero noise and instantly actionable.

Last issue, we covered why the best AI-era hire is the engineer who's sceptical of AI output, not the one who trusts it blindly. If you've started testing candidates on their ability to review AI-generated code and catch subtle bugs, you're building a team that can ship quality.

This week, we're stepping back from the interview itself. You know what to look for (Issue 13), how to interview for it (Issue 14), and what separates good from great (Issue 15). The question founders keep asking me now is simpler. How many engineers do I actually need?

You'll learn how many engineers you probably need, what you should pay them, and how to reallocate your hiring budget so every head you add compounds instead of adding overhead.

Not a subscriber yet? Sign up here

A 5-person engineering team in 2026, properly set up with AI tools, can ship what a 50-person team shipped a decade ago. James Currier at NFX predicts that teams of three people will build $100M+ revenue companies with AI agents handling the rest. Gartner says 80% of engineering teams will be smaller, AI-augmented units by 2030.

Founders already know smaller teams can move faster with AI. That's not the insight. The hard part is the questions that follow. How many engineers do I actually need? How quickly should I hire them? And what should I pay?

That's what this issue is about.

Here's where founders get tripped up. They hear "smaller team" and think "lower burn." Fewer salaries, less coordination overhead, tighter team. All true.

But the per-head cost goes up. Significantly.

When your entire engineering capability sits on 3-5 people, each person carries disproportionate weight. A bad hire in a 20-person team is annoying. A bad hire in a 5-person team is existential. The risk per seat is 4x higher, and the compensation needs to reflect that.

The engineers who can do this, orchestrate multiple AI agents, verify output quality, think in systems, and ship without hand-holding, are not junior developers. They're not even average senior developers. They're the top 10-15% of the senior pool. And they know what they're worth.

Junior developer hiring is already down 73% year on year. Companies are consolidating around fewer, more experienced engineers. The supply of people who can run an AI-augmented workflow well is thin, and getting thinner.

If you're offering market rate for a "senior engineer" and wondering why your pipeline is empty, this is why. You're fishing for a different species now.

Stop thinking about headcount. Start thinking about capability per head.

Here's what I'm seeing work for early-stage founders right now:

Cut your planned headcount by at least 30%
If you were going to hire 10 engineers, hire 6-7. Maybe fewer. Pressure-test every role against one question. Could an existing team member, properly equipped with AI tools, cover this? If yes, don't hire.

Invest more in how you hire, not how long
When every seat carries this much weight, you can't afford a sloppy process. But you can't afford a six-month search either. Build a hiring process that tests for the AI-era skills that matter, scoping ability, orchestration, code verification (Issues 13-15), and still gets an engineer through the door in 30 days. Precision and speed aren't opposites. A tighter process is a faster process.

Redirect that budget into compensation
Take the salary you saved from those 3-4 roles you didn't fill and spread it across the people you do hire. Pay 15-20% above market. (Yes, even at Series A.) You're not overpaying. You're pricing correctly for the level of impact each person needs to deliver. And carve out 5-10% of the saved budget for AI tooling. The smaller team only works if they're properly equipped.

This isn't a cost-saving exercise. It's a speed play.

Small, senior teams make decisions faster. Less coordination overhead, fewer meetings, fewer people waiting on each other. Amazon just mandated senior sign-off on all AI-assisted code changes after a string of outages. That's what happens when too many people are generating AI code and not enough people can tell whether it's good.

Your goal isn't to hire the most engineers. It's to hire the fewest who can deliver the most. AI just made that gap wider.

There's a flip side founders miss. If a bad hire in a 5-person team is existential, so is losing a good one. When one engineer carries 20% of your entire engineering capability, their departure isn't a vacancy. It's a crisis. Retention stops being an HR function and becomes a survival strategy. The higher comp helps, but it's not enough on its own. Autonomy, interesting problems, low bureaucracy, these are the things that keep senior engineers from answering recruiter messages. Build the environment, not just the package.

The engineers we covered in Issues 13 through 15, scoped for AI, interviewed for orchestration, tested for verification, are individually worth more than any three average hires you could make. Pay them like it.

Your next hire should cost more. Not because the market is inflated, but because the job is bigger than it used to be.

Cheers
Neil