AI summary by TwinMind
Building Confidence In Yourself and Your Ideas
Y Combinator
21m 10s
Opening Anecdote: Low-Effort Validation
- YC company contacted 25 companies (mostly startups) over 2 weeks
- Product didn't exist, just an idea
- None wanted to buy, decided to pivot immediately
- Founders found this "very low effort" and "funny"
- Vision in founders' heads didn't align with reality
- Instinct to validate with startups even though not their problem
Common Anti-Pattern: Superficial Validation
- Version: Cold email 1000 people on LinkedIn
- Get only 3 responses, conclude "no one wants this"
- This approach is insufficient and superficial
- People trained as PMs struggle with this transition
- Trained for user research/interviews vs selling products
- Coming from big companies solving other people's problems
- Consulting instinct vs founder instinct
- Should copy "employee one" not "employee 10,000"
Building Conviction vs Quick Pivoting
- Quick pivots prevent building muscle of conviction
- High-quality rep concept: need to complete full pivot cycle
- Low-effort example: built product, spammed 10 people, threw away
- Not a high-quality rep - didn't learn anything, wasted time
- Most important customer is yourself
- Build conviction in your own mind that idea is worth working on
- Not about making investors/YC partners happy
- Convince yourself customer is worth helping
- Low conviction founders won't succeed even with other skills
- High conviction can overcome deficits in programming/sales/fundraising
- Example: Justin Kan as high-conviction co-founder
- Every good startup team needs at least one high-conviction person
Definition and Importance of Conviction
- Not believing religiously, but not getting blown off course
- Example: Hard fundraising causing pivot = low conviction
- If investors cause you to give up, you lack conviction
- If every customer conversation changes your product = low conviction
- Historical perspective: Many big companies had investors who thought idea was horrible
- Investor rejection is neutral indicator, not necessarily bad or good
- Problem: Using others' opinions as only key indicator for decisions
Fear and Misinformation Driving Bad Decisions
- Quality of thinking decreases 90% from job to startup
- Fear creates fake information/assumptions
- Example: Founder claimed 75% of batch already launched (false)
- Batch hadn't even started yet
- Taking anonymous online advice as gospel
- Another example: Alum said need $10K MRR to raise
- Different expectations for companies entering with products vs ideas
- Rational investors judge differently - why don't founders?
- False expectations create fear → fear creates bad thinking
Customer Perspective Technique
- Founders forget they've been on buying side of software decisions
- Ask: "Which sales tactics would have worked on you?"
- "How often do you respond to LinkedIn cold emails?" (Answer: Never)
- Founders realize they're experts at understanding their customer
- Should use tactics that would have worked on themselves
- Realization: "That's why we funded you" - they knew something about the topic
- Pivoting moves them into areas they know nothing about
YC Standard Deal Impact
- Before: Founders worried about demo day
- Now: Half million dollars removes artificial pressure
- "No gun to your head except the one you're holding"
- Low conviction founders: money isn't the problem
- Need to "put the gun down"
Random Walk vs Directed Progress
- Random walk: use random number generator for direction
- Never get anywhere with random direction changes
- Like emailing 25 startups repeatedly
- Rowboat analogy: changing direction constantly = lost at sea
- Need vector with forward progress in specific direction
- Random walk leads to running out of energy after 1.5-2 years
- No learning, tired for no good reason
- Hard to articulate why it was good use of time
- Creates bad memories of startups vs additive learning experiences
MVP Definition and Reality
- Many claim to launch MVP that "nobody wanted"
- MVP lacks agreed-upon definition despite frequent use
- If nobody uses your MVP, it's probably not viable (missing the V)
- Should be able to make it work for at least one person
- Can cheat and do things that don't scale for MVP
- Developer tools example: Founders building tools they don't use themselves
- "It's not good enough for you to use but you're selling to strangers?"
- Minimum bar: Be willing to use your own product
- Many companies never actually build true MVP
- Focus on making one customer happy before scaling plans
Key Principles Summary
- Takes longer than expected - don't set unrealistic expectations
- Fear makes smart people think poorly
- Solve your own problem, be your first customer
- Complete high-quality reps with good form
- Build conviction through proper validation cycles
