This is excellent advice, but be careful about following it too literally: you could end up iterating on Artix or FamSpam.
You can draw a straight line from the abandonment of Artix (pg's first startup, making online art galleries), through Viaweb (startup #2, acquired by Yahoo!), to the founding of Y Combinator and the creation of Hacker News. And, as his recent keynote (http://news.ycombinator.com/item?id=487686) reminded us, Chris Wanstrath's decision to abandon FamSpam helped bring GitHub to the world.
It's hard to imagine any amount of 'tuning' leading Artix or FamSpam to fame and glory; in both cases, the way forward was to start anew. In startups, rapid iteration is a key to success---and sometimes you have to iterate at a higher level of abstraction than a single startup.
Totally agree with your last bit. "sometimes you have to iterate at a higher level of abstraction".
I think, anyone who wants to start a business should look at things on higher level. They shouldn't limit themselves on idea-level. Instead, they should look at business and life as one. That way, probably the advice of "never give up" sounds more relevant.
If having your own business is one of the goals in life, then, probably the only time you are pronounced failed (and should give up) is when you die (end of your runway). Otherwise, running out of cash, work on failed idea and things like that, for me, is just a temporary obstacles and you still have chances to keep playing with the knobs.
Probably this link below illustrate the point better.
http://tinyurl.com/bjh8wd
When we start looking at things in this high-level, i think the case of FamSpam to GitHub sounds more like another dial on the knob.
You can draw a straight line from the abandonment of Artix (pg's first startup, making online art galleries), through Viaweb (startup #2, acquired by Yahoo!), to the founding of Y Combinator and the creation of Hacker News. And, as his recent keynote (http://news.ycombinator.com/item?id=487686) reminded us, Chris Wanstrath's decision to abandon FamSpam helped bring GitHub to the world.
It's hard to imagine any amount of 'tuning' leading Artix or FamSpam to fame and glory; in both cases, the way forward was to start anew. In startups, rapid iteration is a key to success---and sometimes you have to iterate at a higher level of abstraction than a single startup.