Teaching Your Computer to See

Google is doing it again, not satisfied with self driving cars now they’re trying to teach your iPhone to see. I found this article on Forbes.com, called “Teaching the iPhone to Drive.” Basically what the article is saying is that camera technology is getting to where it can see better than the human eye however, a computer cannot process the data without human interaction, it still needs us to tell it what it is seeing. According to the Author of this article however that could soon change, and we’re already seeing steps in that direction.

The author calls it the “visual singularity” the point where computers can see better than we can, and according to him it’s fast approaching. To an extent its already here in specialized formats, one thing he mentions is the LIDAR system in Google’s autonomous cars. However, that system is extremely expensive and too large to be practical except for what they are using it for. Also, there are license plate and facial recognition systems, but those are specialized, they are good at what they do but nothing else. What’s holding the technology back is that being able to see and processing what you see is a ton of information and right now computers can’t handle it. This is where the guys at Google come into it. 

Basically they figured out that it was the internet with its mass of online texts and digital sounds that taught computers how to read and understand sounds (Siri), so they networked a bunch of processors(16,000), and set them loose on YouTube. The machines ran for over a week and “looked” at millions of images. The results? The network taught itself to recognize cats. Why cats? Have you ever seen how man cat videos there are on YouTube? Now taken out of perspective it doesn’t sound all that impressive, but it’s a huge improvement over anything else that has come along. So, to make a long article less long, how does this affect us? One thing the article mentions is Medicine. According to the author coupled with advanced diagnostic programs that are already being developed, doctors may be out of the job. What other ways do you guys think this could be used? Should it be used, computers have been known to have errors? How will it affect us if/when computers are literally able to do everything we can, and do it better?

The Final Blow to Internet Privacy?

The internet isn’t well known for being private. Major companies like Google and Facebook have been tracking users all over the place for years, as well as mining data from all corners of the internet to sell to advertisers. But even Google and Facebook don’t know everything about your web activity. It’s still kind of private-ish. However, thanks to AT&T and a number

of other internet service providers, the last bits of privacy we currently enjoy are about to be dissolved.

Earlier this year, all of the major ISPs in the United States (AT&T, Verizon, Comcast, Cablevision, and Time Warner Cable) announced that they will be rolling out a “six strikes” plan to crackdown on copyright infringement. Basically, the ISPs will penalize anyone they catch pirating stuff online, and the penalties will become more severe with each strike.  But what is disconcerting about the initiative is how they plan on catching pirates. Reports from earlier this year have revealed that they will be utilizing deep packet inspections on a massive scale, which means that they will be keeping track of absolutely everything that you do online. (A real-life corollary to a DPI would pretty much be someone following you around 24/7 and filming everything you do, everywhere you go.) It appears that they also have databases to store the information they collect; however, no information has been released on how much of the data they will store or for how long.

They may catch a few pirates this way and deter a lot of future ones. On the other hand, most serious infringers (at least the ones with brains) will probably just start using VPNs and keep pirating away. If the people they’re trying to catch are just going to find a way around it, the result of the ISPs’ program will still be massive-scale surveillance, but only of their innocent, law-abiding customers who aren’t doing anything wrong. So what do you think? Will this initiative be effective? If so, is it worth sacrificing what’s left of our privacy for?


The Beauty of Google Documents

Hello gang, with Milestone III coming up I thought it appropriate to use my blog post as an opportunity to pay homage to Google Docs. Google Docs is a free Java-based web application that allows you to create documents and upload them to the cloud for storage online. The service is initially free up to a memory cap of 10 GB, after which monthly payments are necessary. This in and of itself is uninteresting – there are many services that provide this kind of data storage, such as Dropbox.

Google Docs is unique for being a web 2.0 technology for its ability for the document creator to share his/her work and collaborate in real time with others. This means multiple people can be working on the same document in real time and have their additions appear as they write them. This feature is a boon for students, as anyone with even a basic computer can access the internet and write text. I have had many a study session with other students by uploading our professor’s study guide, and having everyone fill in the information they know best. It also works wonderfully for projects (such as our upcoming Milestone).

But how did Google Docs come to be? Its origins are in fact two different products, Google Spreadsheets and Writely. Google spreadsheets was a simplified version of the current Google Docs, limited to creating data spreadsheets. However, Google’s purchase of Upstartle in March of 2006, the startup that created Writely, was the major jump into creating the present product. Writely carried the feature that defines Google Docs today – collaborative text editing. Four years later, in March of 2010, Google purchased DocVerse, allowing full compatibility with Microsoft office. Last year, offline viewing was made possible by a web app that automatically uploads your content once connected. The present day Google Docs is a clean, multi-functional product that is fully compatible with almost all common file types.

What have your experiences been with this product, or similar collaborative text editors? What do you think could be done to further improve it? Personally, I think adding speech to text would make it the perfect product. What are your thoughts?