Visit Your Local PBS Station PBS Home PBS Home Programs A-Z TV Schedules Watch Video Donate Shop PBS Search PBS
I, Cringely - The Survival of the Nerdiest with Robert X. Cringely
Search I,Cringely:

The Pulpit
The Pulpit

<< [ Dethroning King Gillette ]   |  Small is Beautiful  |   [ What Carly Will Be Missing ] >>

Weekly Column

Small is Beautiful: Compared to the Mac Mini, AMD's PIC is Half the Computer at Less Than Half the Price. What's Wrong With That?

Status: [CLOSED]
By Robert X. Cringely

After two weeks of writing about Apple's Mac Mini, I have tiny PCs on my brain. This time, it is AMD's Personal Internet Communicator -- a $185 PC that probably ought not to exist at all, but I'm glad it does. The PIC's stated objective is bringing computing to 50 percent of the world's people by 2015, and to do that, AMD is selling the little bugger through third world phone companies and ISPs. I think, with a few modifications, they should sell it here.

I'm not sure why AMD is doing the PIC, since profit margins on this type of device are infinitesimal, but most people trace the project to AMD CEO Hector Ruiz, who supposedly wants to bring the Internet back to the Mexican village where he was born. AMD founder Jerry Sanders probably rolled over in his Mercedes when he heard about it, but bringing computing to the third world is laudable, and a little gizmo like this appeals to a lot of us in the first world, too.

The PIC is not a canonical computer, it's an embedded Internet appliance somewhat along the lines of the 3Com Audrey I wrote about a few weeks ago, though without a built-in display. The PIC's strengths and weaknesses are precisely the same, and depend only on what kind of user you are. It can't accept new applications (the OS is Windows CE). It keeps user changes and personal settings on a separate disk partition so that the main OS partition can be updated at any time back to factory settings from a hidden 'factory reload' partition. It has no legacy interfaces at all (just VGA, RJ11 modem, AC'97 audio ports, and four USB 1.1 ports). It has no fan or even any passive ventilation. It has a 366 Mhz AMD Geode processor, 128-megs of SDRAM, and a 10-gig Seagate hard drive. It is ugly. But balancing that it is cheap, cheap, cheap.

The basic model has most of what my Mom would need to be happy, though. You get a browser with typical extensions such as Real Player and Flash (no Java -- a shame), some 'productivity tools' -- Microsoft-compatible word processor and spreadsheet -- Microsoft Media Player, and that's about it. The most focused purpose is Internet access.

The main thing wrong with the PIC is its business model, which generally assumes that the end customer can't afford to buy such systems, so AMD sells to ISPs in each country who then lease/rent the PIC for a few dollars per month. That's not bad per se, but it keeps PICs out of hands like mine because I live in the wrong country and can't even get a friend to buy one for me overseas, which is just crazy.

Think of the PIC as a cheaper, dumber Mac Mini. Most of the right bits are there and the price is right. Yes, there should be a Linux model, there should be Ethernet, and that xBox (literally) hard drive is too small. But even without Linux, given a bit more effort on AMD's part, this little guy could be used to replace fading K-12 PCs all over America at prices that schools can actually afford. The power savings alone are such that an eight watt PIC will pay for itself in under two years.

But will any company but AMD ever build PICs? I think they should, and here's why. There is an interesting transition taking place in the ultra-low-end computer market right now as consumers are starting to use mobile phones to perform functions that might previously have been done with handheld computers like the iPaq. As a result, handheld sales are actually dropping, which in the PC market means the niche is already dead. Microsoft is trying to follow this trend by putting its software in phones, but for the hardware OEMs the course to follow is not so clear. The logical thing to do, it seems to me, is to split the niche into its two component parts -- mobile communication and cheap computing. Phones get the nod for mobility, but HP and Dell could easily pick up the cheap computing segment by selling many sub-varieties of PIC. It is ideal for home automation, for becoming a car video server to end drowning in Dora the Explorer DVDs, for acting as a home Internet gateway, for hosting the inevitable VoIP home PBX -- each a 100 million unit market, and each totally untapped by the big OEMs.

Will it happen? I don't know. The cost could be driven down (or the profit driven up) by eliminating Microsoft from the product completely, but that's unlikely to be driven by AMD, which these days does pretty much whatever Redmond tells it to. And the HPs and Dells are probably too married to their existing business model and can't escape its gravity. The irony is that there is only one influence I can see that just might drive these other companies to do the right thing, and that's the example of the Mac Mini.

We'll see.

Just in case $185 and eight watts is too much money and power for your taste, another alternative is the Gumstix line of Waysmall computers that start at $139. These little PCs, which are literally the size of a pack of chewing gum, come in 200- and 400-Mhz models using Intel XScale processors that are becoming popular in mobile phones. They run Linux and inherently do, well, nothing at all, but that's what makes them so cool. Here is a little machine with about the same computational power as AMD's PIC, without the hard drive or VGA port, but with an MMC memory card slot and optional Bluetooth networking.

I sat down for about 10 minutes thinking of applications to run on these little guys, and here is what I came up with. The fact that the best idea is sneaky is either commentary on our times or -- as my mother might posit -- on my (lack of) character. You be the judge.

How about a USB-powered Linux microcluster that uses a USB minihub to provide both network and power? The benchmarks aren't very impressive unless you start thinking in terms of performance per cubic foot.

Now do the same kind of microcluster, but put the computers on battery power and use Bluetooth for clustering. Fill a box with gumstix and build a nanosupercomputer. Invite your Gumstix buddies to a party and -- if it is a big enough party and people stand close enough together -- cure cancer.

Build a wireless running computer or cycling computer.

As a Bluetooth device, I think you could package a board into either a headset or a handset that actually talks to Bluetooth-capable mobile phones and applies heavy encryption to scramble calls. Drug dealers and terrorists would love them, except of course, they'd probably be conspicuous by their encryption. So instead find a way to use the board to do audio steganography, somehow encrypting a real voice conversation into a completely fake conversation. Spy agencies would love/hate them.

Forget about that last one. I never wrote it.

Comments from the Tribe

Status: [CLOSED] read all comments (0)