The biggest mistake when I tried to start a company was thinking it was about finding a problem: the real problem is not having a worldview.
What does it mean to have a worldview? It means being able to partition companies, information, and people into "good" and "bad" piles. It means being a polarizing filter. It is building a collection of scissors.
It faces the contrarian tradeoff: the additional entropy--ideally investable information--of a polarizing worldview must compensate for its errors.
Two examples are Peter Thiel and Balaji Srinivasan, where Peter falls more in the strong-state worldview (Palantir, Anduril), and Balaji falls more in the decentralized tech progressive worldview; their unique perspectives filter the portfolio companies.
I'm not there yet. I can't articulate a vision of the future, which made finding problems hard and finding problems I cared about even harder. Worldview generate problems naturally--the diff between the predicted future and the present--but without one, I found myself grasping.
Unlike finding a problem as a discrete event, worldviews compound and develop with time.
So the next time I set out to "start a company"--"solve a problem someone will pay me for"--I'm going to have a worldview that helps me clearly articulate a vision for the future, and work backwards to concrete problems.