Google-funded startup builds search engine + API for predicting the future | Hacker News Reader

Google-funded startup builds search engine + API for predicting the future

technologyreview.com

27 pointsdstorrs16 comments

i wonder what happens when journalists start paying for access to recorded future and put those predictions on the internet. i suppose it will create prediction bubbles that are disjoint from reality. wikipedia cyclic-reinforcement all over again.

They already do it - just look at all the Apple blogs. The only question is whether they'll pay for predictions vs. pull it out of their ass for pure profit.

If it really could predict the future, why don't they trade stocks using it?

Could be a sign that stocks are not part of the future =)

OT: but I really think someone perverted should create a derivative market for people's Farmville "wealth" and other video game scores/assets.

I haven't played video games since Street Fighter II, and can't make a better informed argument, but if the scores are public for some huge games, I think it would be fun to bet on them (even if for "virtual currency".) Say, offer the games in "baskets", ranging from Farmville to MindCraft to WW. Bet on today's top score, biggest losers, etc. I say basket because individual score picks might be gamed.

Something like this might encourage more economic and financial research into video games, helping make the dismal science more attractive to "kids".

Good point - if you could create an accurate tool for predicting the stock market, how much money could you make? Answer = all of it.

If it was accurate, we'd never hear about it. It'd be top secret.

You don't even have to be that accurate. If you're right 51% of the time (consistently), all you have to do is make enough transactions to make unlimited money.

It's not that simple. You need to take into account maximum drawdown and slippage. If all your wrong trades (49%) will happen in a row, you will be toast ;)

they have financial firms as customers

maybe the stock market isn't predictable?

In theory, this shouldn't be able to predict future financial information or stock prices, because the stock prices should already reflect any information that would be used in the prediction model (efficient market hypothesis). Of course, that's just a theory. Thousands of traders think they can beat the market with tools, models, and prediction engines of their own (though few actually do) So, in the end, this site may attract that type of crowd, people who think they are gleaning information to help them "win" - in actuality, I'd be skeptical about its ability to do that though.

Jane Street is an example of a company that builds models and predicts stock prices (short term only) successfully.

http://vimeo.com/14317442

They made $2 billion last year, I think.

Take a look at Urban Survival (http://urbansurvival.com/week.htm) and the peoplenomics and halfpasthuman sites referred to there.

George Ure says some weird things, but I find the discussions and predictions there fascinating.

All I can say is, I don't know what's going on.

In times like these I think Knuth's earth shaking announcement is a good rebuttal: http://river-valley.tv/tug-2010/an-earthshaking-announcement

This is to figure out what Apple plans to do before they do it, right?

YAHS (Yet another Hari Seldon).

Reply on news.ycombinator.com