Wednesday, May 14, 2008

Counting the days

Truemors reports that IBM has build a computer that can do 1 petaflop, which is 1e15 (1 with 15 zeros after it) calculations per second. According to the commenter on the same article, the human brain is estimated to have a capacity of about 2e16 calculations per second, or 20 petaflops. So this machine has a theoretical capacity of about 1/20th of a human brain. Big deal.

On the other hand, Moore's law tells us that processing power doubles about every 18 months. About four-and-a-half doublings gets you to 20x, which works out to just under seven years.

In my last post, I talked about John Gruber's calculations that we moved computing technology from desktop to smart phone in about 7 years. For rough numbers, lets call that $2100 down to $700. And the size went down much more than that.

Ray Kurzweil has done the complete math on this, but basically we can count the days until you can have a computer in your pocket that is as smart as you, for a few hundred dollars.

Tuesday, May 13, 2008

Not your father's phone

My 6-year-old daughter was practicing lines for a play she will be in, and one line was “Radio was at the peak of popularity in that day.” I tried to help her understand what that meant, and asked her if people listen to the radio today. She said “yes”. I asked what was more popular than radio – expecting to hear “TV”. Instead, she said “computer”.

John Gruber estimates that today’s iPhone is about as powerful as an 8 year-old computer. When my daughter is 14, and needs her own cell phone and computer, they might be the same thing - or at least the thing she carries in her pocket will be as powerful as what I have on my desk today.

It would never occur to me to go to a physical dictionary or phone book to look something up, and I rarely look at a paper map. She will have a device in her pocket with a processor fast enough to edit video, over 100GB of storage capacity, and a full-time, high-speed Internet connection to a much improved world-wide network filled with applications (that run in the web – not on her phone) designed from the ground up for mobile use.

If current trends are any indication, she will know where her friends are, what they are doing, and if they are available to talk or meet. She will always know where she is – and not in a “here is your address” kind of way; she is likely to have at her fingertips the intimate knowledge normally reserved for a native of her immediate surroundings – even if this is the first time she has been there. How far to the nearest bathroom, coffee shop, power outlet? Is the subway running on time, or should I just (electronically) summon a cab? What is the trade-off in time and cost of those two choices?

Maybe her phone will slurp down the menu and interior and exterior pictures of every restaurant or cafĂ© she walks past on a street, just in case she wants to come back and try one later. If she sees a cute boy in Starbucks, her social network system will figure out if they have any friends in common – and maybe get them an on-line introduction right away.

Speaking of Starbucks, she probably won’t have to order with a live person. When she walks in, Starbucks will recognize her phone, look up her typical order, and ask her phone if she wants one of her recent drinks, or maybe today’s special. She can just confirm what she wants, the phone will pay the bill, and maybe buzz her when it is ready.

For the older generation, this stuff will feel like a gizmo to figure out. For her it will feel like an automatic transmission – learning to do things the manual way will seem hard.