I’m sorry, but I don’t want there to be software predicting success or failure for startup companies. It reminds me vaguely of screening tests to predict an individual’s success or failure. Why bother to live it if the algorithm says it won’t work?
And I’ll bet I’d like them, too. I certainly don’t want them to fail when I see their picture and read about them. Two former Oxford students, one 21 years old and the other 27, have an interesting claim and, apparently, seed money that makes it at least worth following. Today’s New York Times has the story, A Start-Up Says It Can Predict Others’ Fate.
Kirill Makharinsky, 21, and Bob Goodson, 27, call their software a “start-up predictor,” and they say their company, YouNoodle.com, might give an edge to venture capitalists and other investors trying to decide whether to sink money into an early-stage company.
“We don’t want to replace investors,” Mr. Goodson said. “We simply believe that industries of comparable size have utilized artificial intelligence to inform decision-making.”
“Give us some information, and we’ll give you some idea of what the company will be worth in five years,” he said.
Needless to say, though, like them or not, I’m skeptical. I’ve been following artificial intelligence since the early 1980s when I was paid for that by Creative Strategies International. Why not algorithms to predict a good novel or film, or whether a team will succeed? Come to think of it, how has professional sports lived without it?
I generally root for startups to succeed. But artificial intelligence predicting startup success? I’m finding that hard to swallow.
Then again, focusing well and carefully on factors affecting probabilities is what makes grand masters in bridge and card counters in Las Vegas.
Through the years I’ve met with dozens of venture capitalists and done business with half a dozen, and I don’t think I’ve met one who couldn’t talk about general factors predicting successes and failures, or one who wouldn’t be able to point to some exceptions to the rules.
And of course the NYTimes story can’t resist the obvious mirror-on-mirror question at the end: What does their system say about their own chances of success?
“So far, we haven’t run ourselves through it,” Mr. Goodson said, adding that the results could prove baffling. “If it says we’ll fail, and it’s right, that’s something of a paradox.”
What? You haven’t run yourselves through it? What’s wrong with that picture?
This entry was posted on Monday, February 18th, 2008 at 6:55 am and is filed under startup stories, venture capital. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.4 Responses to “Is it Wrong to Hope They Fail?”
Leave a Reply
February 18th, 2008 at 4:35 pm
“Why not algorithms to predict a good novel or film”
Epagogix has done this successfully with scripts in the film industry. If you are looking for more examples you should check out the book Super Crunchers. People in general seem to be resistant to the idea that a computer can rank something better than they can, but with large enough data sets it seems obvious that a computer will be able to do much better quantitative analysis than any human.
February 18th, 2008 at 5:09 pm
Thanks Jamie. I found Malcolm Gladwell’s New Yorker piece on Epagogix. A google search on epagogix turns up some additional viewpoints, fascinating, certainly not unanimous, but worth considering. I appreciate the addition.
For my sense of it, it’s not the computer that makes me skeptical, it’s the rules-based human inputs it needs to process. So many subtleties, so hard to put into systems.
And the other problem is just the grinding power of it, if and when it works, what that would do to creativity, not to mention availability of capital, and entrepreneurship. Can you see the pitch presentation ending up with data inputs, number crunching, and the calculated conclusion? Chilling, no?
February 19th, 2008 at 12:34 pm
I agree that if pitches ended up like that it would certainly be chilling, I don’t think many would want to live in such a world. Certainly you can never judge an idea solely from its beginnings, as many great ideas have been discovered progressively, or even by accident.
That being said, it seems that what YouNoodle is trying to do is to have some sort of predictor of future success based on team factors. (If you check out the VentureBeat article you will find a more balanced explanation of what they are actually trying to do).
February 19th, 2008 at 12:48 pm
Jamie, is that you noticing me squirming, trying not to go back on my black-or-white approach on this issue? And you keep coming up with evidence, too, making it harder for me to retreat and entrench. I may end up believing in it yet. Thanks again.