The Australian Government has kicked off a $43B National Broadband Network (NBN) to give everyone in a Australia a 100Mbit/s fibre connection. It’s the biggest infrastructure project in our history, and represents about $2100 per person of unfunded government debt. The theory is that it will make us more productive, help education, health and business.
Funny thing is I am quite content with my 1Mbit/s DSL, in fact I could live with 128kbit/s for my web surfing, email, and occasional tarball download. I find email to be the most useful thing on the Internet, and that works fine over dial up. My kids soak up the extra bandwidth for movies, but that is hardly making the country more productive (probably the opposite). The best thing about DSL is that it’s always on, rather than the speed. Maybe I am atypical, but I have never said, “I wish I had faster Internet” while using DSL.
Reminds me a bit of Windows and MS Office. Once a certain level of performance/GUI was reached (Windows 95 and MS Office 97) everything that came afterwards was (expensive) fluff. Well once I obtained always on connectivity via DSL, I reached that “good enough” point.
I think we can do better. Here are some ideas I have for spending AUD$43B over 8 years:
- EV conversions: Lets convert every small car in Australia to be a 100km range EV. Now there are 20M people here, so maybe 5M small cars. That gives us $8,600 per car. A current 100km conversion in quantity 1 costs about $25,000, but in quantity 5M we can expect some big discounts for volume, so $8,600 should do it. Actually for $8,600 each we could probably build new EVs, however recycling a petrol car saves a lot of energy embodied in the manufacturing process. A 100km range EV would cover 90% of the populations driving needs (it does for our family). Pleasant side effects would be the creation of a new (export) industry, lower greenhouse emissions (if charged from green electricity), and radically reduced dependence on foreign oil.
- PV solar or Wind: Lets put PV solar on every house in Australia. I am guessing there are about 8M houses (2-ish people per residence), this means $5,400 per residence. A 1kW PV system costs about $12,000 today (although we currently get an $8,000 rebate). However it’s reasonable to assume at least 50% plus quantity discount for 5M so $5,400 should do it easily. That’s a total of 5GW of PV solar. That’s about twice the current peak electricity generation capacity of the state of South Australia where I live. Actually that’s probably pessimistic, industry standards are drifting down to USD$2/watt, so $43B would give us 15GW of PV solar (at 1AUD = 0.7USD). With wind power we could do even better, $43B would buy us perhaps 30GW at such a high level of investment. Now 30GW ($1USD/watt) at a wind power activity factor of 30% is 30E9(0.3)(24 hours/day)(365 day/year)/(1E3 W/kW) = 78BkWh/year. In 2005 Australia consumed 220BkWh, so thats a big chunk of our power. With some reasonable electricity consumption measures we could probably live on one third of our current consumption.
- Mesh Potato: The Mesh Potato is a Wifi mesh router with VOIP. You place one on your roof, and it self-forms a telephone network by talking to other Mesh Potatoes on nearby houses. It doesn’t need cell phone towers of land lines. Or phone companies. Lets say in very high volume we can install a Mesh Potato with a solar panel and battery to power it for $100 per house. With $43B we could install 430M mesh potatoes. That’s too many for Australia (only 8M houses), so we could put one on every house in Australia and North America. Or globally its one for every 28 people on the planet. Even better – distribute Mesh Potato networks to the poorest 1B of the world, which makes in 1 phone for every three people. This would build a global telephone network so we could all make free phone calls to each other. As a side effect it would build a free Internet backbone that is independent of land lines, governments, cell phone towers (it uses unlicensed spectrum), and telephone companies. Obviously some scaling and number of mesh hop problems but for $43B I am sure they can be solved!