Showing posts with label YouTube. Show all posts
Showing posts with label YouTube. Show all posts

Friday, October 17, 2014

Net Neutrality: A Counter Viewpoint


Net neutrality to me is obvious. I don't want the Internet to go the way of cable television. Expanding capacity is how you respond to an increase in traffic load.

But it is okay to listen to a counter viewpoint.

The Right Way to Fix the Internet
Letting go of an obsession with net neutrality could free technologists to make online services even better. ..... the Internet never has been entirely neutral. Wireless networks, for example, have been built for many years with features that help identify users whose weak connections are impairing the network with slow traffic and incessant requests for dropped packets to be resent. Carriers’ technology assures that such users’ access is rapidly constrained, so that one person’s bad connection doesn’t create a traffic jam for everyone. ..... It costs more to get online in the United States than just about anywhere else in the developed world ..... U.S. service is sometimes twice as expensive as what’s available in Europe—and slower, too. ...... the Internet arose in an ad hoc fashion; there is no Internet constitution to cite. ..... their equivalent of the Federalist Papers: a 1981 article by computer scientists Jerome Saltzer, David Reed, and David Clark. The authors’ ambitions for that paper (“End-to-End Arguments in System Design”) had been modest: to lay out technical reasons why tasks such as error correction should be performed at the edges, or end points, of the network—where the users are—rather than at the core. In other words, ISPs should operate “dumb pipes” that merely pass traffic along. This paper took on a remarkable second life as the Internet grew. In his 2000 book Code, a discussion of how to regulate the Internet, Harvard law professor Lawrence Lessig said the lack of centralized control embodied in the 1981 end-to-end principle was “one of the most important reasons that the Internet produced the innovation and growth that it has enjoyed.” ...... “unavoidable vagueness” about the dividing line between allowable network-management decisions and impermissible bias. .. The line remains as blurry as ever, which is one reason the debate over net neutrality is so intense. ......... if profit-hungry companies are left unfettered to choose how to handle various types of traffic, they “will continue to change the internal structure of the Internet in ways that are good for them, but not necessarily for the rest of us.” ........ codifying too many overarching principles for the Internet makes many engineers uncomfortable. In their view, the network is a constant work in progress, requiring endless pragmatism. Its backbone is constantly being torn apart and rebuilt. The best means of connecting various networks with one another are always in flux. ......... “You can’t change congestion by passing net neutrality or doing that kind of thing,” says Tom Leighton, cofounder and chief executive of Akamai Technologies. .. To keep traffic humming online, Leighton says, “you’re going to need technology.” ........ A central tenet of net neutrality is that “best efforts” should be applied equally when transmitting every packet moving through the Internet, regardless of who the sender, recipient, or carriers might be. But that principle merely freezes the setup of the Internet as it existed nearly a quarter-century ago, says Michael Katz, an economist at the University of California, Berkeley, who has worked for the FCC and consulted for Verizon. “You can say that every bit is a bit,” Katz adds, “but every bitstream isn’t the same bitstream.” Video and voice transmissions are highly vulnerable to errors, delays, and packet loss. Data transmissions can survive rougher handling. If some consumers want their Internet connections to deliver ultrahigh-resolution movies with perfect fidelity, those people would be better served, Katz argues, by more flexible arrangements that might indeed prioritize video. Efficiency might be more desirable than a strict adherence to equity for all bits. ......... For many years, high-volume sites run by Facebook, YouTube, Apple, and the like have been negotiating arrangements with many companies that ferry data to your Internet service provider—backbone operators, transit providers, and content delivery networks—to ensure that the most popular content is distributed as smoothly as possible. Often, this means paying a company such as Akamai to stash copies of highly in-demand content on multiple servers all over the world, so that a stampede for World Cup highlights creates as little strain as possible on the overall Internet..................... Netflix last year was accounting for as much as one-third of all U.S. Internet traffic on Friday evenings. .... In the short term, Netflix resolved the problem by paying for more of the peering points that carriers such as Comcast and Verizon required. More strategically, Netflix is arranging to put its servers in Internet service providers’ facilities, providing them with easier access to its content. ....... the Netflix fight shouldn’t distract regulators who are trying to figure out the best way to keep the Internet open. They should be focusing, he says, on making sure that everyday customers are getting high-speed Internet as cheaply and reliably as possible, and that small-time publishers of Internet content can distribute their work. .... A tiny video startup doesn’t generate enough volume to force Comcast to install extra peering points. ........ “zero rating,” in which consumers are allowed to try certain applications without incurring any bandwidth-usage charges. The app providers usually pay the wireless carriers to offer that access as a way of building up their market share in a hurry... In much of Africa, people with limited usage plans can enjoy free access to Facebook or Wikipedia this way. ......... In the United States, T-Mobile lets customers tap into a half-dozen music sites, such as Pandora and Spotify, without incurring usage charges. ...... When Tim Wu talked about net neutrality a decade ago, he framed it as a way of ensuring maximum competition on the Internet. But in the current debate, that rationale is in danger of being coƶpted into a protectionist defense of the status quo. If there’s anything the Internet’s evolution has taught us, it’s that innovation comes rapidly, and in unexpected ways. We need a net neutrality strategy that prevents the big Internet service providers from abusing their power—but still allows them to optimize the Internet for the next wave of innovation and efficiency.


Monday, March 03, 2014

Deep Learning

A.I. Artificial Intelligence (album)
A.I. Artificial Intelligence (album) (Photo credit: Wikipedia)
"..... deep learning, a relatively new field of artificial intelligence research that aims to achieve tasks like recognizing faces in video or words in human speech ..... "

Is Google Cornering the Market on Deep Learning?
Companies like Google expect deep learning to help them create new types of products that can understand and learn from the images, text, and video clogging the Web..... Not everyone is happy about the arrival of the proverbial Google Bus in one of academia’s rarefied precincts..... a cultural “boundary between academia and Silicon Valley” had been crossed ..... deep learning experts were in such demand that they command the same types of seven-figure salaries as some first-year NFL quarterbacks..... Of the three computer scientists considered among the originators of deep-learning—Hinton, LeCun, and Bengio—only Bengio has so far stayed put in the ivory tower. “I just didn’t think earning 10 times more will make me happier,” he says. “As an academic I can choose what to work on and consider very long-term goals.” ..... in December, DeepMind published a paper showing that its software could do that by learning how to play seven Atari2600 games using as inputs only the information visible on a video screen, such as the score. For three of the games, the classics Breakout, Enduro, and Pong, the computer ended up playing better than an expert human. ..... might be particularly useful in helping robots learn to navigate the human world
Have you sometimes wondered, especially if you are someone who takes, uploads and publicly shares a ton of photos, that maybe noone else is seeing all those photos? What if your thing is video not photo? Then definitely even less people are watching the videos. What if there are important nuggets in them? What if it is a problem that no one is watching your videos?

Deep Learning
Deep-learning software attempts to mimic the activity in layers of neurons in the neocortex, the wrinkly 80 percent of the brain where thinking occurs. The software learns, in a very real sense, to recognize patterns in digital representations of sounds, images, and other data. ..... computer scientists can now model many more layers of virtual neurons than ever before ..... remarkable advances in speech and image recognition. ..... Last June, a Google deep-learning system that had been shown 10 million images from YouTube videos proved almost twice as good as any previous image recognition effort at identifying objects such as cats. Google also used the technology to cut the error rate on speech recognition in its latest Android mobile software. ...... a demonstration of speech software that transcribed his spoken words into English text with an error rate of 7 percent, translated them into Chinese-language text, and then simulated his own voice uttering them in Mandarin. ...... image recognition, search, and natural-language understanding ...... machine intelligence is starting to transform everything from communications and computing to medicine, manufacturing, and transportation. .... “deep learning has reignited some of the grand challenges in artificial intelligence.” ..... software that is familiar with the attributes of, say, an edge or a sound ...... This is much the same way a child learns what a dog is by noticing the details of head shape, behavior, and the like in furry, barking animals that other people call dogs. ...... In 2006, Hinton developed a more efficient way to teach individual layers of neurons. The first layer learns primitive features, like an edge in an image or the tiniest unit of speech sound. It does this by finding combinations of digitized pixels or sound waves that occur more often than they should by chance. Once that layer accurately recognizes those features, they’re fed to the next layer, which trains itself to recognize more complex features, like a corner or a combination of speech sounds. The process is repeated in successive layers until the system can reliably recognize phonemes or objects. ...... At least 80 percent of the recent advances in AI can be attributed to the availability of more computer power ...... Until last year, Google’s Android software used a method that misunderstood many words. But in preparation for a new release of Android last July, Dean and his team helped replace part of the speech system with one based on deep learning. Because the multiple layers of neurons allow for more precise training on the many variants of a sound, the system can recognize scraps of sound more reliably, especially in noisy environments such as subway platforms. Since it’s likelier to understand what was actually uttered, the result it returns is likelier to be accurate as well. Almost overnight, the number of errors fell by up to 25 percent—results so good that many reviewers now deem Android’s voice search smarter than Apple’s more famous Siri voice assistant. ..... Some critics say deep learning and AI in general ignore too much of the brain’s biology in favor of brute-force computing. ....... deep learning fails to account for the concept of time .... human learning depends on our ability to recall sequences of patterns: when you watch a video of a cat doing something funny, it’s the motion that matters, not a series of still images like those Google used in its experiment. “Google’s attitude is: lots of data makes up for everything” ...... deep-learning models can use phoneme data from English to more quickly train systems to recognize the spoken sounds in other languages ....... more sophisticated image recognition could make Google’s self-driving cars much better ....... Kurzweil will tap into the Knowledge Graph, Google’s catalogue of some 700 million topics, locations, people, and more, plus billions of relationships among them. It was introduced last year as a way to provide searchers with answers to their queries, not just links. ....... apply deep-learning algorithms to help computers deal with the “soft boundaries and ambiguities in language.” ..... sensors throughout a city might feed deep-learning systems that could, for instance, predict where traffic jams might occur.
It is possible to imagine a city that has zero traffic jams. If all cars are smart, driverless cars, and all traffic is machine coordinated, it is possible to get rid of traffic jams.
Enhanced by Zemanta

Monday, July 22, 2013

Google: 25% Of North American Internet Traffic

Image representing Google as depicted in Crunc...
Image via CrunchBase
That number is just mind boggling. I think that puts the onus on Google to take the country into the gigabit Internet era all on its own.

Google Now Serves 25 Percent of North American Internet Traffic
Three years ago, the company’s services accounted for about 6 percent of the internet’s traffic. .... more than 62 percent of the smartphones, laptops, video streamers, and other devices that tap into the internet from throughout North America connect to Google at least once a day. ..... The lion’s share of it comes from YouTube. But Google traffic involving search, analytics, web apps, and advertising is far from insignificant. .... Google is big and getting massive. .... To handle its growth, Google has been on a building binge. It now has data centers on four continents. All this work has been getting a lot of attention. But the tech titan is also hip-deep in another type of build-out, one that’s largely gone under the radar. ..... Google has added thousands of servers — called Google Global Cache servers — to ISPs around the world. These servers store the most popular content from Google’s network — a YouTube video that’s going viral right now or apps from the Android marketplace, for example — then serve it directly from the ISP’s data center, rather than streaming it all the way from Google’s data center. These servers were in a handful of North American ISPs three years ago. Today, they’re in 80 percent of them ..... the world’s leader in infrastructure magic
Enhanced by Zemanta

Monday, April 01, 2013

Business Insider Is Doing Something Right

English: Arianna Huffington attending the prem...
English: Arianna Huffington attending the premiere of The Union at the 2011 Tribeca Film Festival (Photo credit: Wikipedia)
The online publication Business Insider is doing something right and it reminds me of The Huffington Post. The Post did commenting really, really well, and it was and is very good at linking to others, but a paragraph of quote first. Business Insider does that too. It often links to others. But what it does really good is taking you to read other related stories. You read one, and the next thing you know you are reading five others. And it is also smart about getting page hits from pictures and infographics. It sure "gets" digital. It feels lightweight but deals with plenty of heavy topics.

Timothy Thomas: Why China Is Reading Your Email
Henry Blodget Is Quietly Planning a Stunning Return to Wall Street
Exclusive: Snapdeal raises $50 million from eBay and existing investors
YouTube co-founder Chad Hurley announces MixBit video collaboration site
E-Commerce Companies Bypass the Middlemen

NYC Subways Deploy A Touch-Screen Network, Complete With Apps
Washington Post seeks blogger for Style section
Why did Apple hire Adobe CTO Kevin Lynch?
Enhanced by Zemanta