[yyyuuu]

A large part of today's population already resides in big cities around the world. Would it be more plausible to research on how we can improve the situation in the current cities, instead of starting from scratch.

It's like Software, where you come up with a large code base that is ridden with problems, but somehow works most of the times. You can either dive into that and improve it or you start writing your own system from scratch.

I am sure YC would have thought about this. Just want to know why they choose starting from scratch.

[rtkwe]

To change an existing city you have a lot of baggage in the form of existing structures and landowners who you have to either convince that the change is beneficial to them or force the changes and fight the inevitable legal battles. That's just assuming that your plan can be implemented with the existing structures, if not it's going to require purchasing the existing structures and rebuilding/tearing them down.

That's not to stay starting over is easy. Trying to create a city out of whole cloth with sufficient jobs and infrastructure to attract and retain people isn't an easy problem in the slightest. It does avoid the NIMBY fight of trying to change an existing city though.