Everyone assumes AI would have made college easier and my path to entrepreneurship faster.
They're right. And that's exactly why I'd be worse off today.
AI in 2019 wouldn't have made me successful. It would have made me generic faster.
The Counterfactual That Haunts Me
What actually happened (without AI):
- Dropped out of college in 2019
- Had to make a risky, all-or-nothing bet on entrepreneurship
- Couldn't balance studies + startups + visa constraints
- Was forced into complete commitment
- No backup plan, no safety net
What would have happened (with AI):
- Probably wouldn't have dropped out
- Would have automated the parts of school I hated
- Could have done both—gotten a degree AND tinkered with startups
- No forcing function to commit fully
- Multiple options, hedged bets
The obvious take: "AI would have given you more options!"
The deeper truth: More options at 18-19 = more dilution, less commitment, no forcing function for real growth.
I probably would have gotten both—the degree and some startup experience. But I'm not sure where my business would be today.
Why This Seems Better But Isn't
The surface-level logic makes sense:
With AI in college, I could have:
- Automated homework and assignments I found tedious
- Used AI to write papers while focusing on building
- Gotten better grades with less time investment
- Graduated with a degree while running side projects
- Had multiple career options after graduation
This sounds objectively better. More optionality. Less risk. Better outcomes.
But here's what everyone misses about optionality at 19:
More options = less commitment Less commitment = less learning Less learning = generic outcomes
The constraint of dropping out wasn't a bug. It was a feature.
The Generic Founder Problem
With AI at 18-19, here's exactly what would have happened:
Phase 1: Information Consumption (Months 1-6)
- No principles yet, no developed worldview
- Would have consumed whatever startup information was available
- Y Combinator playbooks, Twitter founder culture, TechCrunch wisdom
- AI would have helped me absorb and synthesize this faster than anyone
Phase 2: Generic Execution (Months 7-18)
- Would have copied what seemed to work
- Built apps that were slight variations of existing ideas
- "Uber for X" or "Airbnb for Y" type concepts
- AI would have made prototyping and testing these ideas incredibly fast
Phase 3: Conventional Path (Months 19-36)
- Maybe gotten into Y Combinator with a well-executed but unoriginal idea
- Maybe raised seed funding based on impressive execution speed
- Would have looked successful by all conventional metrics
Phase 4: Plateau and Failure (Years 3-5)
- Most of these scenarios wouldn't have worked out long-term
- Would have wasted years chasing conventional startup routes
- Would have had the best tools in the world but not been successful
The paradox: AI would have made me incredibly efficient at pursuing the wrong things.
The AI Acceleration Principle
AI accelerates you toward wherever you're already going:
- Going somewhere original? AI gets you there faster
- Going somewhere generic? AI makes you generic faster
- Going nowhere? AI just speeds up your aimlessness
At 18-19, I was going nowhere original.
I had no thesis, no contrarian view, no developed principles about business or life.
AI would have amplified my consumption of conventional wisdom and made me incredibly efficient at executing other people's ideas.
Instead of developing original thinking, I would have become a very fast follower.
What Dropping Out Actually Forced
The constraint of dropping out created several forcing functions:
1. Full Commitment (No Backup Plan)
- Couldn't just "try entrepreneurship out"
- Success or failure had real consequences
- Identity became tied to entrepreneurial success
- Had to figure it out, not just experiment
2. Real Stakes (Financial and Social)
- No degree meant fewer career options if business failed
- Social pressure and family disappointment were real
- Couldn't retreat to traditional career path easily
- Every decision mattered more
3. Identity Investment (You Become "The Dropout")
- Couldn't half-ass the entrepreneurial identity
- Had to commit to being different from peers
- Forced to develop contrarian thinking to justify the choice
- Created psychological investment in proving unconventional path right
4. Resource Constraints (Built Character)
- Couldn't rely on university resources or network
- Had to build everything from scratch
- Developed resourcefulness and independence
- Learned to create value without institutional support
AI would have removed all of these constraints. And with them, the forcing functions that drove real growth.
The False Confidence Trap
Here's what would have been most dangerous about having AI in college:
AI gives you incredible tactical ability before you develop strategic wisdom.
At 19 with AI, I could have:
- Built sophisticated applications in weeks
- Created impressive demos and prototypes
- Automated complex business processes
- Appeared technically competent beyond my years
This would have created false confidence in my own judgment.
I would have thought I was ready to build a company because I could execute quickly.
But execution without direction is just expensive motion.
The hard part of entrepreneurship isn't building things. It's knowing what to build.
AI would have made me incredibly good at the easy part while leaving me completely unprepared for the hard part.
When AI Helps vs When It Hurts
| Situation | AI Effect |
|---|---|
| Clear direction + developed principles | Massive leverage |
| Unclear direction + no principles | Faster toward generic |
| Technical execution gap | Fills it beautifully |
| Strategic thinking gap | Makes it worse (false confidence) |
| Learning known things | Incredible accelerant |
| Forming original views | Mostly useless |
At 19, I was in the bottom row of every category:
- Unclear direction ✓
- No developed principles ✓
- Strategic thinking gap ✓
- Need to form original views ✓
AI would have been actively harmful for my development.
The Specific Counterfactual Scenarios
Scenario 1: The YC Track
With AI, I probably would have:
- Built a polished SaaS tool for college students
- Applied to Y Combinator with impressive demo
- Gotten accepted based on execution quality
- Raised $500K seed round
- Hired 5-10 people to scale the "proven" concept
- Struggled with product-market fit for 18+ months
- Burned through funding on a fundamentally unoriginal idea
- Shut down or acqui-hired after 2-3 years
Outcome: Would look successful initially, fail quietly later
Scenario 2: The Side Hustle Track
With AI, I probably would have:
- Stayed in school while building multiple small projects
- Automated income through various AI-powered services
- Made $5-15K/month during college
- Graduated with good grades and multiple "businesses"
- Felt successful without building anything substantial
- Never developed the commitment necessary for big outcomes
Outcome: Comfortable mediocrity, no breakthrough learning
Scenario 3: The Consultant Track
With AI, I probably would have:
- Used AI to become incredibly efficient at academic work
- Started consulting services for other students
- Built a profitable but not scalable service business
- Graduated with strong grades and immediate income
- Continued consulting after graduation
- Never took risks necessary for venture-scale outcomes
Outcome: Good income, limited upside, no original value creation
In most scenarios I can imagine, I would have been worse off long-term despite appearing more successful short-term.
What I Actually Learned by Struggling
Without AI, dropping out forced me to learn:
1. Original Thinking Development
- Had to develop my own frameworks since I couldn't copy others efficiently
- Forced to understand principles behind successful patterns, not just execute them
- Created contrarian views through necessity, not choice
2. Real Customer Development
- Couldn't fake customer research with AI-generated insights
- Had to actually talk to potential customers and understand their problems
- Learned the difference between what people say and what they pay for
3. Resource Constraint Innovation
- Couldn't automate my way around fundamental resource limitations
- Developed creativity and resourcefulness through genuine scarcity
- Learned to build sustainable businesses, not just grow fast
4. Commitment and Persistence
- Couldn't quit when things got hard because there was no backup plan
- Developed tolerance for uncertainty and discomfort
- Learned to stick with problems long enough to find real solutions
5. Identity and Principles
- Had to develop strong sense of self and values to justify unconventional path
- Created personal framework for decision-making under uncertainty
- Built confidence based on proven ability to figure things out
AI would have short-circuited all of this learning.
The Paradox of Helpful Tools
The more helpful tools become, the less they help with the things that matter most.
AI is incredibly helpful for:
- Executing known strategies
- Optimizing existing processes
- Scaling proven concepts
- Learning established knowledge
AI is actively harmful for:
- Developing original thinking
- Building tolerance for uncertainty
- Creating forcing functions for growth
- Learning through struggle and constraint
At 19, I needed the second list. AI would have given me the first.
This is why I'd be worse off with AI in college despite being more "productive" in the short term.
What This Means for Current College Students
If you're in college with access to AI:
The temptation is to use AI to optimize your current path:
- Get better grades with less effort
- Build impressive projects quickly
- Automate away the parts of school you don't like
- Create optionality and reduce risk
The danger is that this optimization prevents the development you actually need:
- Struggling with difficult concepts that build mental models
- Failing at things and learning to persist through difficulties
- Developing original thinking through constraint and necessity
- Building tolerance for uncertainty and incomplete information
Questions to ask yourself:
- Am I using AI to get better at important things, or to avoid important things?
- Is AI helping me develop original thinking, or just execute other people's thinking faster?
- Am I building tolerance for difficulty, or automating it away?
- Is AI creating forcing functions for growth, or removing them?
The counterintuitive advice: Sometimes choose the harder path even when AI could make it easier.
The Constraint-Seeking Strategy
Instead of using AI to remove constraints, use constraints to direct AI:
Bad use of AI in college:
- Automate homework so you don't have to think about it
- Use AI to write papers without engaging with the material
- Build projects without understanding the underlying principles
- Optimize current path without questioning if it's the right path
Good use of AI in college:
- Set artificial constraints that force original thinking
- Use AI to test your ideas faster, not to generate them
- Automate execution so you can focus on strategy and direction
- Remove busy work so you can spend more time on hard problems
The principle: Use AI to amplify your original thinking, not to replace it.
Examples of constraint-seeking:
- "I will only build businesses that solve problems I personally have"
- "I will only use AI for implementation, not ideation"
- "I will commit to one project for 18+ months regardless of other opportunities"
- "I will drop out if my business reaches $X revenue" (create stakes)
The Delayed Gratification Test
Having AI in college would have failed the delayed gratification test:
Immediate gratification with AI:
- Better grades with less effort
- Impressive projects built quickly
- Multiple options and reduced risk
- Appearance of success and competence
Long-term gratification without AI:
- Original thinking developed through struggle
- Resilience built through constraint
- Commitment created through necessity
- Real competence earned through difficulty
The marshmallow experiment for founders: Can you choose constraint and difficulty when AI offers ease and optionality?
Most people can't. That's why most people don't build exceptional businesses.
The founders who succeed with AI are the ones who would have succeeded without it.
AI doesn't make you a better founder. It makes you a faster version of whatever kind of founder you already were.
The Uncomfortable Questions
For current students with AI access:
Are you using AI to become more capable, or to avoid becoming more capable?
If you're finding this useful, I send essays like this 2-3x per week.
·No spam
Is AI helping you develop contrarian views, or just execute popular ones faster?
Are you building original value, or just optimizing existing templates?
Would you be proud of your path if AI suddenly disappeared?
For founders building with AI:
Are you using AI to test your original ideas faster, or to avoid having original ideas?
Is AI amplifying your unique insights, or helping you copy what already works?
Are you building something only you could build, or something anyone with AI could build?
Would your business survive if your competitors got access to the same AI tools?
For investors and advisors:
Are you backing founders who would be exceptional with or without AI?
Do you prefer founders who use AI to optimize known strategies or to explore unknown territories?
How do you evaluate original thinking when AI can simulate it convincingly?
Are you investing in AI-amplified mediocrity or AI-amplified excellence?
What I'd Tell My 18-Year-Old Self
If I could go back to 2019 with AI available:
"Don't use AI to make college easier. Use it to make dropping out smarter."
Specifically:
- Use AI to validate business ideas faster, not to stay in school longer
- Let AI handle execution so you can focus on original problem identification
- Use constraints (like dropping out) to force real commitment and stakes
- Build something only you could build, not something anyone with AI could build
The meta-lesson: The most valuable thing about being young isn't access to tools. It's the ability to take risks that older people can't.
AI makes risk-taking feel less necessary because it provides more options.
But optionality is the enemy of exceptional outcomes.
The best thing I never had was an easy way out.
Why I'm Actually Grateful I Didn't Have AI
Looking back, not having AI in college was a gift:
It forced me to:
- Develop original thinking through constraint
- Build resilience through struggle
- Create commitment through necessity
- Learn principles through first-hand experience
It prevented me from:
- Optimizing the wrong path efficiently
- Building false confidence through easy wins
- Avoiding the hard work of developing unique insights
- Taking the conventional route just because AI made it faster
The result: I built something genuinely different because I had to, not because I chose to.