For someone who makes his living forecasting which startups will succeed and which will fail, venture capitalist Vinod Khosla is surprisingly averse to spelling out such predictions.

"I'm a big believer that I can't predict anything. I can't get anything right," Khosla said Thursday at BoxWorks, Box's annual user and developer conference. The two-day event, which took place in San Francisco, had more than 5,000 attendees.  

The Sun Microsystems co-founder wasn't being self-deprecating. In general, he thinks most predictions are useless. 

"When change is happening, it's very hard to predict what the change is. It's not hard to know that the time is ripe for change--that something big will happen--but it's not clear what will happen," Khosla explained. 

However, when trying to be inventive during such a rapidly moving and complex time, certain steps can do improve your odds for success. Here's Khosla's advice on what exactly those steps involve:  

Just do it. 

I hate pontificators and lecturers and people who study stuff. I love people who do stuff. Do stuff. Get it right, get it wrong. Amplify the right, change the wrong quickly. I have no respect for these big forecasts and McKinsey studies. They just don't work--especially as the rate of change increases.

So my advice to everybody whether you're a big company or small--go do stuff, try stuff, learn from it. Don't try and predict stuff.

Learn how to program.

You can't live in France without learning French effectively. It's silly for Yale or Harvard to have you learn a second language like French when you live in a computer age and don't know computer science. So, in terms of the language you need besides your primary language--whether it's English, or Hindi or wherever you live--your next language should be computer science.

It's absolutely essential. No question about it. No reason to waste time reading Jane Austen when you can use that time to learn computer programming.

Take control and embrace luck. 

As an entrepreneur, you control about 30 percent of your success. If there are 10 important factors, you may have control over three of them. Three or four your competitors control. And the rest is just plain luck.

If you can survive longer, you give yourself more of an opportunity to get lucky.