Started SimpleDirect with 5 co-founders in 2022.
Ended with 0 in 2024.
Revenue went up. Here's why PG is wrong for 2025.
The Co-Founder Gospel
Paul Graham preaches it. Every YC partner repeats it. VCs demand it.
"Solo founders are a red flag. You need:
- A technical co-founder
- A business co-founder
- Complementary skills
- Shared emotional burden"
Where this comes from: 15 years of YC data. Startups with co-founders had higher success rates.
Why it made sense (2005-2020):
- Building software required full-time technical expertise
- Scaling required specialized roles
- Division of labor increased total output
- Emotional support mattered (it's genuinely hard)
Who benefits from this advice: VCs. They want optionality. Easier to back a team than bet on one person.
But here's the thing.
The math changed.
My SimpleDirect Experiment
2022 (5 co-founders):
- Revenue: ~$12K MRR
- Team: Me + 4 co-founders (technical, marketing, sales, operations)
- Equity split: 20% each
- Monthly burn: $18K
- Margin: 33%
- Decision speed: Glacial (every choice needed consensus)
2023 (2 co-founders):
- Revenue: ~$28K MRR
- Team: Me + 1 technical co-founder
- Equity split: 50/50
- Monthly burn: $12K
- Margin: 57%
- Decision speed: Slow (two-person decisions)
2024 (0 co-founders):
- Revenue: ~$35K MRR
- Team: Me + AI tools + 2 contractors in India
- Equity split: 100% mine
- Monthly burn: $8K
- Margin: 77%
- Decision speed: Instant
The pattern: Fewer people = more profit = more freedom.
What AI Actually Replaced
1. The Technical Co-Founder
Used to need someone who could code full-time. Now I have:
- Cursor: Writes code faster than I can review it
- GitHub Copilot: Handles boilerplate and debugging
- Claude: Architects complex systems and reviews code
- ChatGPT: Solves specific technical problems instantly
Cost comparison:
- Senior developer: $200K/year + 15-25% equity
- AI stack: $600/year
ROI: 333x cost savings, 10x speed increase.
What AI can't do: Make product decisions. But that was never the technical co-founder's job anyway.
2. The Business Co-Founder
Used to need someone for sales, marketing, operations. Now I have:
- Make/Zapier: Automates entire workflows
- ConvertKit: Handles email marketing and automation
- Offshore specialists: $10-20/hour for specific tasks
- AI assistants: Research, writing, analysis
Cost comparison:
- Business co-founder: $150K/year + 15-25% equity
- Automation + contractors: $3K/month
ROI: 50x cost savings, zero equity dilution.
What contractors can't do: Long-term strategic thinking. But that was always my job anyway.
3. The Emotional Support
This was the hardest to replace. Building alone is genuinely difficult.
What works better than co-founders:
- Systems over emotions: Sunday Night Test catches problems early
- Metrics over opinions: Data doesn't lie or have bad days
- Community over co-founders: Twitter, podcasts, founder groups
- Profit over validation: $35K MRR feels better than any pep talk
The breakthrough: I don't need someone to share the burden. I need systems that eliminate the burden.
The Equity Math
Here's what nobody talks about: Co-founders are expensive. Really expensive.
My 5-co-founder structure (2022):
- My equity: 20%
- Revenue per person: $2.4K MRR
- My share of $35K MRR (current): $7K/month
My solo structure (2024):
- My equity: 100%
- Revenue per person: $35K MRR
- My share of $35K MRR: $35K/month
Difference: $28K/month = $336K/year
That's not a rounding error. That's life-changing money.
The 10-year calculation:
- With co-founders: $840K (20% of $4.2M assumed exit)
- Solo: $4.2M (100% of same exit)
Co-founder cost: $3.36M over 10 years.
Are co-founders worth $3.36M? Sometimes. Usually not.
When You Still Need Co-Founders
AI didn't make co-founders obsolete for everyone. Here's when you still need them:
1. You're Raising VC Money
VCs want teams, not solo founders. If you're burning $300K/month, you need specialized roles.
But: If you're profitable, why raise VC money?
2. You're Building Hardware
Physical products need multiple expertise areas. Industrial design, manufacturing, supply chain, regulatory.
AI can help. But can't replace domain expertise.
3. You Have Massive Scale Ambitions
Want to build the next Google? You need a team from day one.
But: Most "massive scale" ambitions are ego, not strategy.
4. You're in a Regulated Industry
Healthcare, finance, legal. Domain expertise matters. Compliance is complex.
But: Most SaaS businesses aren't in regulated industries.
5. You Genuinely Can't Learn the Skills
Some people can't code. Some can't sell. Some can't write.
But: AI made learning faster. Cursor teaches you to code while you code.
The New Playbook
Year 1: Validate Solo
- Use AI tools for everything possible
- Hire contractors for specific tasks
- Keep 100% equity
- Prove the concept works
Year 2-3: Scale with Systems
- Build automation instead of hiring
- Use offshore talent for execution
- Maintain high margins
- Stay profitable
Year 4+: Strategic Hiring Only
- Hire when bottlenecks can't be automated
- Give equity sparingly (2-5%, not 20-50%)
- Maintain control and decision speed
Never: Default to Co-Founders
- Don't split equity because "that's what you do"
- Don't hire because you're lonely
- Don't give up control because it's conventional
The Contrarian Cases
Daniel Vassallo: Left $500K AWS job in 2019. Built 15+ income streams solo. Makes $50K+/month. Zero co-founders.
Pieter Levels: Built Nomad List, PhotoAI, 20+ other products. $2M+ revenue. Zero co-founders.
Justin Welsh: $2M+ course business. LinkedIn + Twitter + newsletter. Zero co-founders.
The pattern: AI-first builders who kept 100% equity, optimized for freedom over scale.
These aren't exceptions anymore. They're the new normal.
What Changed in 2024
AI crossed the capability threshold.
2022: AI could help with simple tasks 2023: AI could handle complex workflows
2024: AI could replace entire job functions
Want the full playbook? I wrote a free 350+ page book on building without VC.
Read the free book·Online, free
Geographic arbitrage became mainstream.
High-quality talent in India, Philippines, Eastern Europe costs 80% less than US/UK.
Automation tools matured.
Make, Zapier, and specialized SaaS handle most business operations without code.
Capital requirements dropped.
Used to need $500K to build software. Now need $5K for AI tools and contractors.
The combination is lethal for traditional co-founder models.
Paul Graham Is Wrong (For 2025)
PG's essay "What I Wish I'd Known About Startup School" says solo founders are "at a significant disadvantage."
He's right for 2005-2020. Software was hard to build. Teams had advantages.
He's wrong for 2025. AI made software easy to build. Teams became overhead.
The new advantage: Speed and low burn rate, not complementary skills.
The new metric: Revenue per person, not total revenue.
The new optimization: Profit and freedom, not growth and scale.
PG optimizes for venture outcomes (1% chance of $1B exit).
I optimize for freedom outcomes (90% chance of profitable, sustainable business).
Different games. Different rules.
Recommendation
If you're pre-revenue: Start solo. Use AI. Prove it works. Then decide if you need help.
If you're profitable: Question every hire. Can AI + contractors solve this? Usually yes.
If you're considering co-founders: Run the equity math. Is this person worth $1M+ over 10 years?
If you already have co-founders: Don't fire them. But don't default to hiring more.
The default should be solo + AI + systems.
Co-founders should be the exception, not the rule.
The Real Question
Old question: "Who should be your co-founder?"
New question: "Why do you need a co-founder?"
If the answer is "because everyone says I should," that's not a good reason.
If the answer is "to split the workload," AI probably handles it cheaper.
If the answer is "for emotional support," community works better.
If the answer is "because I can't do X," learn X or hire contractors.
The only good answer: "Because this specific person brings irreplaceable value that's worth millions."
What I'd Do Differently
If I started SimpleDirect today, here's my approach:
Day 1-30:
- Solo founder with AI tools
- $5K budget for validation
- No equity given away
Month 1-6:
- Contractors for specific tasks ($2K/month)
- Systems for everything else
- Still 100% equity
Month 6-12:
- Hire 1-2 key people if absolutely necessary
- Give 2-5% equity max, not 20-50%
- Maintain control and speed
Year 2+:
- Scale with automation and offshore talent
- High margins, low burn rate
- Freedom to experiment and pivot
Result: Higher profit, more freedom, less dilution.
The Uncomfortable Truth
Most founders want co-founders for the wrong reasons:
- Fear: "I can't do this alone"
- Validation: "Smart people think this is worth their time"
- Loneliness: "Building is hard and lonely"
- Convention: "That's what successful companies do"
None of these are good reasons to give away 20-50% of your company.
AI solved the capability problem. Systems solved the complexity problem. Community solved the loneliness problem.
The only remaining problem is your mindset.
Action Steps
This week:
- List every task your co-founder(s) handle
- Research AI tools that could replace 80% of those tasks
- Calculate the 10-year equity cost of your current structure
This month:
- Test AI tools for 2-3 high-value tasks
- Find offshore contractors for remaining tasks
- Run the Sunday Night Test on your co-founder relationships
This quarter:
- Build systems that reduce dependence on any one person
- Document processes so they can be automated or outsourced
- Have honest conversations about equity and responsibilities
The goal isn't to fire your co-founders. It's to build a business that doesn't require them.
Why This Matters
Because equity is your most expensive currency.
Because co-founders introduce complexity, not just capability.
Because AI changed the game and nobody updated the playbooks.
Because keeping 100% of a profitable business beats 20% of a venture-backed exit.