Published August 29th, 2025 • Based on Founder Reality Episode 10
Also available on: Apple Podcasts • Spotify • YouTube
I need to tell you about the most counterintuitive business decision I've ever made - and why it saved my company.
Two years ago, I was running a team of 14 people. Today, I have 5.
We're more profitable, we ship faster, and honestly? I actually enjoy going to work again.
Here's the uncomfortable truth Silicon Valley doesn't want you to hear: More people doesn't mean more output. It means more problems.
The Bureaucracy Death Spiral
Let me paint you a picture of what 14 people actually looked like.
Every month, I spent 7 hours just doing one-on-ones. Seven. Freaking. Hours.
That's not including the time to schedule them, prepare for them, or document them afterward.
Add my assistant's time to coordinate everything, and we're talking about a full workday every month just for management theater.
But it gets worse.
We had daily stand-ups at 9 AM where people would robotically recite:
"Yesterday I worked on X, today I'm working on Y, no blockers."
It became toxic. People weren't actually communicating - they were performing.
We had two shadow societies forming within the company: the quant trading group (3-4 researchers and developers) and the product group (everyone else).
It felt like running two different companies, except both were burning my cash.
And here's the moment I knew I'd completely screwed up: I was spending time finding tasks for people instead of having natural work flow.
If you're allocating busy work to justify someone's salary, you're 1000% on the wrong track.
The Sunday Night Test
You know what I call my gut-check for business decisions? The Sunday Night Test.
How do you feel Sunday evening? If you're dreading Monday morning, your spreadsheet might look good but your business is broken. Energy levels don't lie.
Two years ago with 14 people, I failed this test every single week.
I'd lie in bed thinking about payroll, about managing personalities, about performance reviews I didn't want to give for work I should have already known about.
That's not running a startup. That's running a bureaucracy.
When Everything Broke
The breaking point came when one of my developers got married and took time off. Congratulations to him, obviously. But what I didn't realize was how much of his work was interconnected with everyone else's.
Five other developers suddenly couldn't access environment variables.
Nobody knew what was happening in production. We spent weeks untangling the silos he'd unintentionally created.
That's when I realized: complex teams create complex failures.
By late 2023, our cash position was screwed. Fourteen people is a lot of payroll for a bootstrap company. I was losing sleep worrying about making payroll, about all these people depending on me to feed their families.
The stress was eating me alive.
The Great Restructuring
So we made the hard decision. We let people go in waves - first two, then three more, then eventually got down to five total (including me).
I paid every severance I owed. I had difficult conversations with people I genuinely liked working with. It was brutal.
But here's what happened next: AI came along right around that time.
Suddenly, I was back to running something that felt like a startup instead of a corporate job I'd accidentally created for myself.
Life at Five People
Want to know what a five-person team actually looks like?
We don't do performance reviews because I know exactly what everyone's building every day.
We don't have sprint meetings because I can just ask 'how's that feature coming along?' and get a real answer in 30 seconds.
If we need a team-wide call, we hop on Slack and it's done in 10 minutes. Everyone knows exactly what we're working on.
It feels like those college days when you'd pull an all-nighter with your closest friends to hit a deadline, eating pizza and actually enjoying the work.
That's culture. Not some corporate handbook bullshit - actual culture where people know each other and care about what we're building together.
My New Hiring Framework
Now when new tasks come up, here's my process:
- I take it on first. Always. I use AI for brainstorming and breaking down the work.
- If I hate doing it and it drains me, I talk to existing team members to see if anyone's excited about it.
- Only then do I consider contractors. Upwork, Fiverr, whatever. Sometimes outsourcing specific tasks works better than hiring full-time people anyway.
- I never hire full-time unless someone is truly essential. And I mean truly essential.
Here's my thought experiment: If I could only keep one person from my team, who would it be? Start there. Then ask about each person: if they wanted to resign today, would I do everything possible to keep them?
Remove your emotional attachments and be honest. This isn't about firing people - it's about building a team where everyone is genuinely essential.
The Communication Complexity Trap
Here's the math everyone ignores:
Possible pairwise communication paths grow quadratically with team size.
Three people have 3 possible pairs; four people have 6. The formula is n × (n - 1) / 2: 14 people have 91 possible pairs and five have 10. These counts describe possible relationships, not measured communication cost.
At 14 people, we had so many communication layers that information would get distorted by the time it traveled across the team.
Like a game of telephone, except with product decisions and code deployments.
What AI Actually Changed
People ask me about AI replacing workers. Here's the reality:
AI didn't replace my team members. AI replaced the need to hire new team members.
Content creation used to require hiring a content person. Now I can brainstorm with Claude, draft with ChatGPT, and edit myself.
Customer support used to need dedicated staff. Now we built SimpleDirect Desk internally and handle 80% of inquiries automatically.
AI helped us redesign some work on a smaller team. That is my account of the sequence, not proof that AI caused our profitability or makes every planned hire unnecessary.
The Uncomfortable Truth About Scaling
Silicon Valley sells you this narrative: hire fast, scale quick, move fast and break things.
Meta announced more than 11,000 job reductions in November 2022 and another roughly 10,000 in March 2023. Those announcements do not isolate the effect of layoffs on profits, or establish that AI hiring caused the other cuts.
Editorial update, October 11, 2026: Primary records: Meta's November 2022 employee message and March 2023 efficiency update. My own team reduction, severance account and later workload redesign remain a historical postmortem, not a controlled comparison.
Even companies with "endless cash" realized that more people often means more problems, not more progress.
Your competitive advantage isn't team size. It's team efficiency.
For Founders Reading This
If you're feeling drained by management overhead, if you're failing the Sunday Night Test, if you find yourself creating work to justify people's salaries - you might be scaling too fast.
Ask yourself: Do we actually need to hire this person? Or do we just want to be bigger for the sake of being bigger?
Can AI handle this? Can a contractor do this? Can someone on the existing team take it on?
Sometimes the answer is yes, you need to hire. But question the assumption that more people automatically means more output.
I've seen Y Combinator companies jealous of the organic traction our five-person team generates. Think about that.
The Bottom Line
I wouldn't go back to 14 people for anything. We're leaner, more profitable, faster at shipping, and I actually enjoy the work again.
If you're drowning in management overhead, maybe the solution isn't better management systems. Maybe it's fewer people to manage.
Less can be more. Sometimes fewer people doing more is exactly what your startup needs.
Stop optimizing for team size. Start optimizing for team efficiency.
What's your biggest team management challenge? Email me at [email protected] - I read every message and respond to the honest ones.
New episodes of Founder Reality drop Monday, Wednesday, and Friday at 9am EST. Subscribe wherever you get podcasts, or visit founderreality.com
Editorial update, October 11, 2026: That was the schedule announced in August 2025. It is preserved as historical context; it is not a current cadence guarantee. The current free publication is at Founder Reality.
Daily thoughts: @TheGeorgePu on Twitter/X