AI summary by TwinMind
Tarpit Ideas: The Sequel
Y Combinator
13m 50s
Tarpit Ideas Definition and Common Misconceptions
- Tarpit ideas: startup concepts that repeatedly attract many founders but consistently fail
- Analogy: like tar pools that look like water from afar - dinosaurs approach to drink, get stuck, die
- Common misconception: people interpret "tarpit" as simply meaning "bad idea"
- Actual definition: small subset of ideas that are insanely attractive to founders but historically unsuccessful
- Key characteristic: if other founders don't constantly come up with same idea, it's not a tarpit
- Tarpit ideas get excessive praise from friends ("that's a great idea!") vs normal startup ideas that get honest negative feedback
Historical Context and Technology Changes
- Video recorded ~1 year ago, received significant feedback since
- Technology changes can transform tarpit ideas into viable businesses
- Ideas that were tarpits for 100 years can become great businesses when technology shifts
- Current era: LLMs/AI creating new possibilities for previously failed concepts
- Dalton now open to funding classic tarpit ideas if founders can explain how new technology enables success
- Problem: many founders ignore that others have tried and failed before ("history started when I was 18")
AI-Era Common Ideas (Not Tarpits but Frequent)
- "Co-pilot for X" pattern extremely common in applications
- Examples: co-pilot for real estate agents, investment bankers, lawyers, surgeons, dentists, dog walkers
- Not technically tarpits because not enough people have tried them yet
- Result of people applying first principles to AI startup ideas
- YC funding AI-generated content companies since 2016-2017, quality improving significantly
- Example: AI-generated podcasts now indistinguishable from real ones
Two Main Flavors of Tarpit Ideas
Flavor 1: Utopian Coordination Problems
- Belief that world should work differently/better than it does
- Requires simultaneous adoption by many people to work
- Example: coordinating with friends for event planning and hangouts
- Problem: hard to change human behavior despite good intentions
- Everyone supports the concept but won't actually use the product
Flavor 2: Get-Rich-Quick Arbitrage Schemes
- Based on secret knowledge or limited-time opportunities
- Founders think they can fast-forward through normal startup timeline
- Often tied to trending topics (Wall Street Bets during pandemic)
- Example: founders doing well in stock market wanting to create investment advisory for friends
- Mercenary approach rather than solving real problems
Key Takeaways and Advice
- Tarpit list is not static - changes with technology evolution
- Ideas mocked today might not be tarpits in 6 months due to tech advances
- Warning signs: too much positive feedback, "why doesn't this already exist?", seems too easy
- Most important advice: talk to users before asking if idea is tarpit
- Many founders skip user research, expecting investors to validate ideas instead
- Success requires understanding why current technology enables previously impossible solutions
