Tarpit Ideas: The Sequel

Dalton and Michael say tarpits are a small subset of ideas that repeatedly attract founders and often fail. They are unusually appealing and familiar, and friends may praise them, as with food discovery. An idea is not a tarpit just because someone dislikes it or because it is bad.

Many tarpit ideas have been tried repeatedly without a meaningful change in conditions. New technology can change that: the speakers say they are open to classic ideas becoming viable when founders explain what is now possible. They distinguish common ideas like “Copilot for X” from tarpits, since not enough people have tried those ideas yet.

The speakers say each wave of technology makes a different set of problems addressable. They describe funding an AI-generated podcast company that would not have seemed viable before LLMs and voice generation. Founders need to explain how the technology makes their specific problem solvable now.

One pattern assumes people will adopt a better way of doing things, such as an app for coordinating plans with friends; changing behavior across a group is difficult. Another pattern promises quick riches through a temporary arbitrage, illustrated by people inspired by WallStreetBets who wanted to manage friends’ investments.

The speakers say the tarpit list changes as technology changes. They suggest skepticism when an idea seems easy, receives unusually positive feedback, or seems like it should already exist. Their practical advice is to talk to users and get a customer rather than relying on a video to judge an idea.