Start out with a problem that you want to solve. With limited resources (money, time, talent), your team is in a race to find product-market fit that’s either profitable or has potential to be before you run out of resources. The more cycles (ie experiments) that you can go through to try to get to product-market fit, the higher chance your startup will be successful. Having a team of hackers and / or developers helps speed up cycles times gives you a huge advantage and is why Y Combinator values these types of teams so much.

The process of figuring out an enduring solution to the problem is through this cycle:

Form hypothesis - Form an educated hypothesis on how to solve the problem. Set metrics that you want to hit to prove or disprove, as much as possible.
Test - Do the research, create mockups, and test them out on users, as quickly and cheaply as you can.
Draw conclusion - Does this prove or disprove your hypothesis? Where does this lead you? Be honest with yourself.
Repeat. Repeat steps 1-3 until you land on something that potential users you’re talking to desperately want.
Then start to build (but don't stop using this cycle).

This framework applies to each aspect of product, marketing, and sales. Be open minded when trying to figure out the solution. The solution probably isn’t what your first idea or hunch. It also may not be a highly profitable, technologically scalable solution. You just can’t force it to be.