Even that experimental wireless speed is still nowhere near as fast or reliable as current fibre speeds. It isn't even anywhere near as fast as NBN's proposed 100Mb/sec, and would be many times more expensive to roll out.
Compare the experimental speeds mentioned in that article of up to 16Mb/Sec (about the same as current DSL speeds of up to 24Mb/sec), to the experimental speeds of fibre optics which reach 2,000,000 Mb/s. This capacity of fibre optics to carry up to 2000 Gb/sec is 125,000 times faster than the up to 16Mb/sec touted in the above article.
The NBN, by the way, is currently operating at speeds of 100Mb/sec - 6.25 times faster than the 16Mb/sec in that article. Telstra's current LTE system, while touted as "up to 48Mb/sec" by comparison is only really 8Mb/sec.
Also, don't ignore the issue of trying to get 20 million+ users onto a radio spectrum which simply won't support that much data, not to mention the issues we'd have of needing another 50x as many wireless base towers than we currently have to achieve that speed This is much more expensive than the money being spent on the NBN. By the way the government is only funding about $27 billion, with industry paying a large portion of the remaining $10 billion.
For more info, see "Why not wireless?" on the NBN myths blog, or the NBN section of the Whirlpool industry forums.
