1. I want to sing in the shower (hang the iPad on the outer side of the shower door on full volume)
2. I want to check Facebook reminders from my bed (Ami's birthday is tonight. OMG!)
3. I want to read a book (and the headlines) while commuting.
4. I need to Google what "Para Fuera" means. NOW!
5. Cuddle in bed and watch a movie at the same time.
The rising sun model in software services says that if you need to provide service and operations 24 hours a day then all you have to do is follow the sun. For example, an office in the US would work for 9 hours, and then an office in Japan would work for 10 hours and then an office in Spain (5 hours).
Sun light can also be used to save electricity costs. It could be used to heat water, or it could be converted into AC power (electricity). The problem is that it is difficult to make this process effective. Water heats even if I don't need to take a shower and converting sun energy and delivering it to neighbors far away is not an efficient process.
Converting sun light into computing power, however, could be more effective. A household in the US could send his compute task (such as face recognition) through a fiber optic channel to China (where it's currently shiny). The sun powered computer in China would detect the face and send the answer back to the US. Furthermore, sun light is everywhere and most shiny countries already have rooftops covered with solar panels. So there is no need to concentrate computer farms near an abundant energy source. Compute tasks could be distributed into multiple small computers powered by the sun.
In the World Summit of Cloud Computing there was a discussion about how to make sure that a handful of cloud computing providers (Amazon / Google / Microsoft) would not charge premium prices for their computing services. The concept is that an "open cloud" would pressure the prices down (similar to how open source software pushed Microsoft Windows prices down). Individuals would donate residual computing power for free, and in exchange would get free computing power from others.
Powering these personal computers with green energy would solve two conceptual problems:
1) From an ecological standpoint, a commercial cloud vendor can compute more with less energy since it is more energy efficient. So individuals not using their computer should let it sleep instead of using power in an inefficient manner.
2) The price that individuals pay for electricity is much higher than the price paid by commercial cloud vendors which get electricity for a substantially lower fee.
The 3rd problem is a little trickier... Personal computers as we know them today would eventually vanish in favor of dedicated devices (book readers, phones, game consoles, televisions). These computers would be optimized for one purpose and would be inefficient in general purpose computations. We need dedicated computer chips that can work in higher ambient temperatures, low DC power and perform small portions of a larger distributed compute tasks.
I went to the World Summit of Cloud Computing (http://events.myreg.co.il/IGT2009/ ) yesterday, and met many wonderful and inspiring people. The problem was that I could not remember their names. Yes, they each wore a name tag, but they could pretty easily tell if I took a sneak peek at their name tag, and besides half of the tags were rotated backwards.
So I figured, that in the spirit of cloud computing, I would invent a software service that would provide face recognition directly to my glasses (well, I don't wear glasses, but I think they could look good on me). So each time I get close to someone, his name would be reflected on my glasses, and perhaps also his latest facebook and twitter status. Real-time face recognition requires quite a lot of computing power, so my glasses would actually send the picture to the internet ("the compute cloud") and get a response with the name of the person in that picture.
The problem is that as an individual I would be consuming more energy that I would have before the invention of the cloud. And although I can afford paying a few cents per face, planet earth cannot. In 50 years, we won't have much cheep energy resources left - fuel (fossil), gas, atomic energy (uranium) would become rare and expensive. And the price I pay for the face recognition today does not reflect the hidden costs of limited energy resources. A time traveller coming from the future would be willing to pay a good price for that energy, and as a result I would remove my glasses and politely ask people for their name.
I love YouTube because it has great content. What worries me is that YouTube doesn't have much competition (Hulu? Netflix? Vimeo?) Even Microsoft shutdown it's SoapBox website due to hiking costs and lack of popularity. What does a new company need in order to compete with YouTube and stay profitable?
I decided to take on that challenge and called my new fictitious company ChooChoo. Here are some ideas that could help ChooChoo compete with YouTube.
Reduce network bandwidth
Internet bandwidth costs money… If I live in Tel Aviv, and the closest ChooChoo video server is in Budapest… then someone needs to pay for the video download costs. The only way ChooChoo can reduce this cost is by installing video server(s) in Tel Aviv.
Fortunately, most Internet Service Provider (ISP) offer(*) an Infrastructure as a Service (IaaS) portal much like Amazon EC2. ChooChoo can quickly deploy one or more video caching servers in any city. The most popular videos (**) would be copied once per city, and streamed for free to any user in that area. Yeay!
(*) Actually most ISPs don’t offer IaaS yet, but they should.
Being an ISP is all about Last Mile
Filtering copyright protected video content
Some users may decide to upload videos although they are not the video owners (which can cause some legal problems for ChooChoo). Fortunately, the latest (camera) cell phone models can mark (*) each video with an invisible owner signature. Take a video, click on a button, and it’s uploaded to ChooChoo. You cannot upload a video that does not have your signature on it.
(*) Actually current cell phone models don’t have this feature, but they should!
Charging money for video content
Good content costs money. Advertising can subsidize popular low cost productions (talk shows, reality shows). A pay per view payment model could work for live sports events. But other content (Films, TV series) could be easily downloaded illegally (with eMule / Torrent). In fact, some even claim, that viewers would never pay for content.
Fortunately, most Internet Service Provider announced (*) their support of a new hardware device (router) that discourages video piracy. This router enables fast upload speed for files that are marked as user generated content. All other type of files would be uploaded much slower. This would render most video sharing applications (that require an uplink) useless for illegal video sharing. Users working from home, would be able to lift this limitation by allowing the ISP to block file sharing applications entirely.
(*) Actually there isn’t such a router, but there should be one.
Otherwise, it's hard to compete against free videos.
What do you think? Would ISPs allow different video content delivery networks to compete over the last mile ?
Google Wave could open new revenue streams for Google in the future. But I wonder how it supports the existing Google Search business model. Google generates most of its revenues from ads that appear alongside it's search results. In order to make more money it needs to attract more visitors to www.google.com. Google also wants to increase the chances a visitor would actually click an ad. That's why Google keeps improving the Google search experience, and that's why Google gets better in displaying more relevant ads.
How do you think Google Search would benefit from Google Wave ? Below are some of my random thoughts. I'd like to hear yours.
Google Wave would have a better search experience.
Google Wave enables real time search updates.
