Y Combinator Pushes The Envelope
tomstechblog.com
tomstechblog.com
Maybe the details of the implementation could change my mind, but in the end this would probably be useful for little other than a marketing gimmick.
Is now really the best time for the online dating world? The best times seemed to be shortly after the big sites we have now had just came into existence. Now there social stigma of actually using them as such has become overpowering. Bars though, don't really have this problem.
> I’m not saying its ever going to be 100% accurate but if we can determine that x guy and y girl live happily ever after 85% of the time if they share z traits that still gives us the ability to match people far more successfully than we do now.
Go ahead and even begin to try to find data to back up that big assumption. No metric with such high correlation exists. This is my point.
The position you seem to have taken right now is that you believe no metric exists and no matter who says there is one you won't believe it.
eHarmony has shown several reporters verifiable numbers proving 90 member every day. I've never heard anyone dispute that claim. The only dispute is what percentage of their members that actually represents. But lets say its a pathetically small percentage, 20% or so.
The fact that they've achieved even limited success seems to indicate that the idea has merit and that refinement of that idea could yield better results.
That's a theory backed up by mathematics. In the end relationships are just two people choosing to join together and stick with each other which means a discipline such as game theory which has been shown to effectively predict the future actions of people could be a prototype for predicting relationship success. Because staying in a relationship with someone is just a behavior.
It really is. It isn't cheesy. It's inspiring and exciting.
I wrote the post in question and I actually registered with Hacker News for the first time just to reply to you.
As far as your comment goes, I apologize in advance for being harsh, but get over yourself.
Being aloof might be cool in high school but if your over the age of 18 you’re old enough to realize that the world is too harsh to turn down any opportunity to be in awe of it. The reason I chose to share my observation is because it made me feel good about the world we live in and I thought it might make others feel the same way. Quite frankly, I wish more people would do the same and I wish you wouldn’t discourage them from doing so.
So yes, "Technology is cool" and we could all use as many reminders of that as we can get our hands on.
I'm not aloof toward technology or cool uses of it. Awe is an important emotion and something we could all use more of. I guess I was just not very piqued by your post, because the ideas in it are fairly generic, except for the dating one, which has been done in Japan for a while now.
However, this is am amazing time to be alive, and I agree with you, the rate at which cool things happen is increasing.
HN is attracting its share of "1 quip snipers", trolls that criticise but not really bother to expand why & generally make a nuisance of themselves ~ http://www.paulgraham.com/trolls.html but we're not all like this.
"... But if you look outside of that self imposed box you see there's so very much left to do in technology ... Now I've found that dating someone gives me a tremendous opportunity in bars because rather than be a part of the dating rat race I actually get to observe it as an outsider which is really enlightening ..."
This was the standout part of the post that struck a chord with me. It's the reason I posted it. A hard problem at the moment I see at various levels is the "human" matching problem. I often wonder the reasons why this problem is tackled so badly in software? Is it because the same thing that makes nerds good at computing also degrade their ability to undertand "human matching"?
It's not just dating where solutions to this problem can be applied. It could also be applied to anywhere you have a requirement for recruiting individuals to interact with other people. Part of the reason why the problem domain is not well understood by nerds could also be covered by the "homo-logicus", "homo-sapiens" divide Alan Cooper describes in "Inmates are running the Asylum". Funny enough this idea also partially explains Trolls. Nerds posting insults like jocks giving wedgies.
The big problem with the matching humans from what I can gather talking to people who try pure software solutions, is the lack of ground-truth. What appears on a site just doesn't match real life. Nothing beats the opportunity to evaluate a human in real-life. Any augmentation to help filter this process may have pay-off in a lot of other areas.
The big problem with the matching humans from what I can gather talking to people who try pure software solutions, is the lack of ground-truth. What appears on a site just doesn't match real life. Nothing beats the opportunity to evaluate a human in real-life. Any augmentation to help filter this process may have pay-off in a lot of other areas.
I think you’re right that people lie but I think statistics can, to a certain extent, help with that. To give an example, let’s take a guy who loves Gilbert and Sullivan musicals but is ashamed of it so he says he hates theater productions. But then in response to other questions he reveals more truthful answers.
Eventually, a statistical measurement will determine that Women who love musicals match best with guys who claim not to as long as they fit those other traits. So the computer never manages to decode the lie but it still, through the magic of statistics, manages to make the match just the same.
I’m sure we’re years and years away from actually getting to the point of having an effective system like this but just the fact that we seem to have the power to do it is pretty amazing.