No, it's not that. Look, each day the world generates 2.5 billion gigabytes of data. Honestly, processing it is the simple part. The hard part now is moving it to and from the chip.
Can you explain it to me like I'm a child?
Imagine that you run some water through a generator, like a hydroelectric dam. Building and running a dam is the easy part, it's the miles and miles of waterways each side to move the water -- that's the problem.
So the issue is that there's too much water to move efficiently.
So how do we fix it?
That's the problem. We can't.
Then I'm just gonna have to take this $300 million somewhere else.
The DOE REPRESENTATIVE gets up to leave, just then, SCIENTIST #3 turns around in his chair to face the audience. Except it's not a scientist, but DON DRAPER from popular TV series Mad Men.
Wait. What if you didn't have to move the water at all?
There is a STUNNED SILENCE.
DON DRAPER picks up a WHITEBOARD MARKER, gets up, and begins scribbling on the wall.
What if we created something called "Data Centric Computing," that, rather than pumping information in and out of supercomputers, we put the processing power into the data itself.
But that's impossible.
Think about it. If we put computing power on every stage of the process, then we won't have to move the data at all. It'll be like putting a tiny hydroelectric dam on every river in the country. That's how you get the power you need.
Make it happen. I want one in the Lawrence Livermore National Laboratory and one in the Oak Ridge National Laboratory by 2017.
The DOE REPRESENTATIVE hands over a check for $300 million. The SCIENTISTS sit, open mouthed in admiration. DON DRAPER goes over to the cabinet and fixes himself a large glass of Woodford Reserve, no ice.