In my experience, completely related on topic great ideas can still kill your startup, its an easy thing to bury your head in code and produce feature after feature without looking up and evaluating what that feature gave you apart from more code to manage. Microsoft can afford to make massive bloated software that does everything and the kitchen sink "decently" startups can rarely afford to do anything more than one thing very very well.
reminded me of a quote i tweeted recently & then saw @bmull include today: "As long as I listen to my customers, I never need to have another original idea" -@nielr1 in Do More Faster
"If I had asked people what they wanted, they would have said faster horses."
of course you can't listen to what everybody says, but as a team, take all of the generated ideas and determine which are the best. i think it removes a lot of the assumptions that are present going in.
lots of memorable quotes. stuff that makes complete sense, but until somebody phrases it a certain way, it just never clicks...ya know?
 - http://www.readwriteweb.com/archives/most_promising_company_...
plenty know about the groupon story, but who remembers when livingsocial was emerging by having people share their top 5 lists on facebook? what about scoutmob which began as a wifi provider?
both of those companies noticed that they were sitting on an large amount of consumer data, and they both figured they could use it to pivot their model towards the emerging group discount space.
But the point remains - don't be fooled into thinking that the grass is always greener. Unless it really is.
So they went in the opposite direction, again from the starting point of regular blogging, toward micro blogging (less energy, lower fidelity, whereas audio blogs are higher energy, higher fidelity).
Since it was based on fundamentally different (in fact, opposite) assumptions than the business they had raised money for, they offered all investors their money back and created a new legal entity to keep the paperwork clean.
That is not the same as being distracted by a flashy idea, where you have neither good evidence that your current idea is bad nor concrete evidence that the upcoming idea is an improvement.
The article is also talking about jumping to ideas unrelated to your current focus, which aren't able to leverage what you've already learned and built. Twitter was invented entirely in response to what they learned from their experiment with audio blogging.
I have recently been building projects, when in the middle, I suddenly get an idea for another project, I would down all tools to kick start something new with the belief I could multitask and do both.
The truth is, you can't! Give your initial project 110% and get it out of the door before you try anything new. That way all of your products will have the quality they deserve.
If you have that problem, too, you should check out a program I wrote called Notepad Consolidator (https://github.com/jaredsohn/NotepadConsolidator) that writes the contents of all open untitled notepads to an autonamed file.
Definitely beats notepads :)
The important thing is to make sure you're pursing this other idea because it's actually something worthwhile not just as an escape from your startup not doing so great... Basically, think about things rationally and not emotionally.
All categorical statements are bad, including this one.