A friend of mine swears by a strategy he calls "smashing ideas" to generate ideas for research: Take two fundamental ideas from the literature (the more fundamental the better) and try to smash them together into a new idea. Says a remarkable amount of research (and successful papers) can be viewed in this light. The strategy is obviously a great creative tool, generally. Though, for business I still reckon the best primary mindset to have is identifying real world problems and thinking of solutions to these problems (perhaps smashing comes in at this point). In research and art, creativity pays off more because real world benefit as voted by paying customers is not as important.
I propose the ycombinator turing test. Can you generate an application which will get accepted by ycombinator? This may require generating additional content to establish team credibility including twitter posts, blog posts, etc.
How do you know who the people are? Presumably, the app would create blogs, facebook pages, linkedin, twitter, etc. Passing a phone interview would be harder, though not impossible. You might need really good voice recognition. Speech generation probably wouldn't be insurmountable.
In person, this might be trickier. This might require a robot mannequin. Alternatively, you could hire an actor with wireless ear plugs. Presumably, the computer would tell the actor what to do.
While the idea proposed is important, I believe pg cares more about the the team, which I think is reflected in the application. So you could very well generate a random idea and be accepted as long as you have an awesome group of people.
"CloudPerfect is a creative new way of getting you upvoted on Quora. Between trolling /b/ and marriage, people believe they can't learn how to paint. This is where CloudPerfect comes in, with lots of coffee and an army of monkeys, CloudPerfect will revolutionize your social experience."
There was a recent XKCD[1] on pairs of words that sound good together and quite a few of those domains were available. So I don't know what it'll be yet, but my next project is going to live at lasernarwhal.com.
Lol. My surprises of the day were toogeo and talkik. I think I have a knack for finding these. I should charge. :) Ooh, talkic is even better. It's like talk and topic!
Good Grief - never announce in a public forum, even one as trusted as HN that you are interested in a domain. Nefarious individuals (Domain Squatters) will grab it - then good luck ransoming it away from them.
I've grabbed it via 1&1 - if you're going to do something useful with either of them - let me know so I can transfer it to you gratis - just create something interesting. :-)
Wow, I had always been nervous about that, but I assumed I was being paranoid:
"Network Solutions may reserve domain names that are searched on our Web site for up to 4 days. During this period, these domain names will only be available to register at networksolutions.com" http://about-networksolutions.com/customer-protection-measur...
I knew somebody would check my work. Notabot.com is the only domain in that list that is available only as a premium domain name ($1,888.00). All the rest are up for grabs.
I had an idea for a similar joke website a few months ago, but their algorithm is much smarter than mine. Looking at their code, it looks like they are essentially randomly combining sentence fragments together in order:
var A = ["A new and improved", "A better, faster", "An awesome", "A vibrant", "An innovative", "A faster", "A rapid", "A quick and dirty"];
var y = ["Searchable", "Virtualized", "Cloud-based", "Mobile", "Application-aware", "Viral", "Friendlier", "Online", "Web-based", "Streamlined", "Performance-oriented", "Next-gen", "Revenue-shared", "Professional"];
I attempted to mine Crunchbase profiles and then build "descriptions" of a startup based on a markov chain, but I found that there was too much noise in most of the Crunchbase descriptions to produce text that sounded at all plausible. Their method is a lot smarter (and easier).
"Pure Javascript, HTML and CSS. No server side at all. Simple and easy."
Surprisingly surprising - the average framework X programmer would automatically use framework X. What a great example of using the right tools for the right job.
Great website. Can't be viewed in landscape mode on android. The socks and twitter link obscure the text. At first I thought that was part of the joke. Whatever. Gave me a laugh this morning and that is worth its weight in gold.
Can you make it so we can submit silly/lousy ideas we've had or heard from others? Or maybe ideas that were rejected from YC? That would be a fun collection to look at.
But seriously: sometimes randomness breaks assumptions and generates great ideas. See also:
http://www.rtqe.net/ObliqueStrategies/
Added: Sites with working online decks: http://www.joshharrison.net/oblique-strategies/ http://music.hyperreal.org/artists/brian_eno/oblique/oblique...