The real answer is build something you know obsessively well, are extremely passionate about, and something that is vaguely original. If you can't do that don't do a start up.
If you just have that much time and money to throw away fund a popular (existing) open source software project into a commercial enterprise.
The UK Universal Credit benefit system is terrible: thousands of people in the UK could be left penniless this Christmas, all because of an admin error with the Universal Credit benefit system. It could mean families up and down the UK going without the money they need to pay their rent or buy food – let alone celebrate Christmas.
The government is already under pressure to fix this fast. It’s splashed all over the news. That’s where the startup comes in to reduce this pain and make sure families on low incomes get what they need before Christmas comes round (rather than politicians just moan – which I hate but won’t help lessen the pain for people tomorrow).