Steve Lohr's New York Times article "Can this man reprogram Microsoft" doesn't offer much that hasn't already been endlessly reported about Ray Ozzie's background and mission to bring the services economy to Microsoft. He does bring up Google's discussions with thin-client system maker Wyse Technology:
"The discussions are focused on a $200 Google-branded machine that would likely be marketed in cooperation with telecommunications companies in markets like China and India, where home PC's are less common, said John Kish, chief executive of Wyse. 'Google is on a path to developing a stack of software in competition with the Microsoft desktop, and one that is much more network-centric, more an Internet service,' Mr. Kish said. 'And this fits right into that.' "
Sun has been trying push its vision of thin client networked computers for years. Oracle's Larry Ellison promised of a $199 networked computer in 1999 and by 2003 exited from business. Several companies tried their hand at Web surfing and email thin clients in the late 1990s, but failed. Wyse practically invented the Windows thin client business, along with Citrix in the 1990s, and now has Linux clients, but it's a corporate sell. There's also MIT's $100 Linux-based, Wi-Fi enabled laptop.
Certainly, times have changed, and in the rapidly growing world of broadband Internet a computer designed to consume those services is much more tenable than in the previous decade. David Berlind speculates on a Google PC, and Dana Gardner thinks that Oracle should resurrect its networked computer concept. The question Lohr alludes to in bringing up a Google/Wyse connection (why not Google's pal Sun? or Cisco?) is whether a Google software stack could significantly erode Microsoft desktop dominance. So far desktop Linux hasn't gained much traction versus Windows or the Mac.
Google's brand seems to be magic these days, and there's plenty of robust open source software and Google's own secret sauce and infrastructure prowess to come up with a consumer Internet computing appliance. Intel Inside or AMD Inside could be joined by Google Inside as an alternative to Microsoft, Mac and the branded Linux distributions. The 'GoogleNet' stack would include a consumer appliance version of Google's Linux, maybe a Google browser, lots of drivers, some hosted productivity apps along with the other Google services, but users could replace those services and applications with whatever they chose. Google would makes its money selling ads on its components and possibly charging subscriptions fees for net access and intrastructure support, and be able to extend its brand. Of course, Google would have to find partners as Lohr suggests to serve the world with a hardware device and telcommunication services, unless the company decides to make 'GoogleNet' global.
If Google were to get behind a consumer Internet-centric appliance, Microsoft and Yahoo would not be far behind. The new model for developing, distributing, consuming and paying for software has far fewer barriers to entry than previous generations. This time around any one company garnering a 90 percent market share is a very unlikely end result.