Business

'Don't build a computer to predict stock moves'

Jason Collier
Image caption Jason Collier did not have the best experience dealing with banks just before the crash

Each week, we ask chief technology officers and other high-profile tech decision-makers three questions.

Answering this week is Jason Collier, chief technology officer at Scale Computing, which is based in Indianapolis, Indiana, and develops storage for small and medium-sized firms.

What's your biggest technology problem right now?

One of the reasons we created this [storage] is that it was something we needed ourselves.

This was because our biggest problem was data sprawl. We're getting in so much data, and nobody wants to delete any of their e-mails or anything, so you just get more and more of it.

And with the amount of data, it is difficult for mid-market companies that don't have the money to buy the data storage and the tools to manage it.

What's the next big tech thing in your industry?

The ability to "scale out" - this means instead of one big storage system, you have lots of little systems that plug in and add to each other. Adding nodes into a cluster, basically.

Traditional storage methods are "scale up" - you fill it up until you can't fill it any more, then you buy a new system and upgrade everything into that.

"Scale out" hasn't taken off yet because of its complexity, so making it simple is one of our priorities.

What's the biggest technology mistake you ever made - either at work or in your own life?

We made a product once that had a spam filter that used a predictive engine, so we thought, "We made the decision engine, why don't we use it to suggest whether you should buy, sell or hold stocks?"

So we built a supercomputer named SuperBruce for $60,000 and we got it to be as fast as the computer BMW uses to do its crash-test simulations.

We set up a hedge fund to attract investment and approached banks at the beginning of 2007. You can see this wasn't a good time to get into stocks.

The first two banks we approached were Bear Stearns and Lehman Brothers.

They said, "Give us nine months of stock data." So we did. Then they asked for nine more months and we got fed up of dealing with people in this industry.

Of course, by the end of 2007, it became very obvious that they didn't want to put money into us, because they were going under themselves. So we shut down the hedge fund.

So I wouldn't advise people to build a computer to predict stock market movements. But, at the same time, the data sprawl we ran into with SuperBruce convinced us to start this firm.

Related Internet links

The BBC is not responsible for the content of external Internet sites