Part of the problem with ConnectU is they didn't have an ability to execute their ideas themselves and when they recruited Zuckerberg he was like why am I executing their idea for them and letting them reap the profits? ConnectU didn't fail because they shared their idea. They failed because the executed poorly by not drawing up a contract with the person that was going to be working for them.

I personally believe that you should share ideas with people you trust to get feedback. Then build a v0.1 prototype that you can launch and start getting feedback and improving it.

Some things to consider withholding are long-term goals, tricky algorithms, novel monetization strategies, and markets you've researched and have confidence are under-served.

It essential you safely expose as much of the idea and the work you do as early as possible so you can learn. But don't let anyone tell you that ideas are worthless or that there aren't people who will profit from good ideas that spill out. There is a reason, after all, that we have a patent system. "Ideas are worthless, its all in the execution" is as much a lie as "execution doesn't matter."

I'm not sure if you are aware of "Lean Startup" methodology but strongly recommend you get aquatinted with some of the principles. Web 2.0 companies like Dropbox, IMVU, Aardvark, KISSMetrics etc have applied these techniques to be successful.

You don't necessarlity need to share your idea with anyone but SHOULD build it and launch as soon as possible, your goal in the first month should not be to get lots of users but validate your hypothesis and collect data on usage (like if your users like your offering and wanting to come back)

I posted this comment on a different thread but re-posting here:

------------

We launched a product recently http://sociaholic.com and applied lean startup principles.

Problem Hypothesis: The existing Twitter to Facebook integrations and apps don't preview links/videos when publishing to Facebook from Twitter, thus by limiting the desirability to click on a link on facebook pages.

Execution: Design to launch along with a video in less than 2 weeks, the goal was to validate our solution has a market. We update your facebook page from twitter in nice way (previewing links, converting @handles, etc)

Validation: Our first week: Over 200 active users out of ~250 signups, conversion rate (homepage to adding twitter account) ~30%. 2% dissatisfied customers (5 out 250 people uninstalled) The product has a much bigger vision and Customer Development is doing wonders and we plan to do this for every step we move forward.

Design to launch in 2 weeks is impressive. It makes me feel like I should set more stringent goals with my projects and reading.

http://sivers.org/zipit

That someone will eat you for lunch.

Another reason to talk about 'good ideas' is that sometimes they're not that good, you just don't know it. Finding out early because of conversations with others will help you focus your efforts on those things that really stand out.

Collaboration is essential for success, ideas can be born in a vacuum but it will need a lot of effort to bring one to market in a way that you can make a profit on them.

The true important thing is execution. And is VERY hard to get it right.