A few things that you can't do in the official app, but you can do in the Teslabot: (1) voice control - send a voice messages/commands to Teslabot, (2) find out whether your car can go to a destination such as San Francisco with the current charge - Teslabot figures out location of the car, distance to the destination, battery range and answers that question, and also the bot is available from your laptop/desktop.
Agree with the author that academia sometimes focusses on a very narrow band of research topics, which might or might not improve end user experience. Also, agree with the author that personalized news has very interesting NLP+Machine Learning problems.
However, I am not convinced that personalized news is what users wants, and whether users want to discover popular news and articles by serendipity, socially, rather than personalization by algorithms. Further, I think personalized news has very limited opportunity for generating significant revenue.
Netflix focusses on movie/tv-series subscription model (streaming+dvds), and they are laser focussed on improving consumer experience in this area. On the other hand, Apple, Google, Amazon focus on other businesses (hardware, advertising, and e-commerce). That's why I think Netflix will survive, and that's why startups survive and thrive: Focus!
The problem is for avid tv-series watcher (I admit I am one) netflix is a very bad solution. The obvious problem is that they do not have the fresh stuff. I wouldn't mind some delay, but when it's a season worth of delay, it stops being useful. This, however, is not the worst part - I also watch some old ones. But even there - netflix has way less than full collection, some seasons are missing, some episodes are missing from otherwise complete season, etc. And there's no way for me to track which episodes I already watched, make a list of my favorites to re-watch, or do any other useful things with them. The should at least do something like next-episode.net, but they could probably do even better. But they don't seem to be focusing there, unfortunately...
This is where torrents come in handy. It's amazing how easy it is to grab a whole season of a show in 720p (if it's just aired) or 1080p (if the Blu-ray set is out). Yesterday I torrented season 3 of Downton Abbey - it won't even air (in the US) until January! I use Anigrate[0] to keep track of which shows I've watched.
I was just about to comment about that: browsing for content on Amazon is a terrible experience, because they use mostly the same layout as the rest of their products. Netflix has the advantage of focussing and a single type of products and (hopefully) give a better experience.
(1) Ability to save and tag tweets with urls (kind of Delicious)
(2) Url content inline with the corresponding tweet
(3) Search over all the tweets you could have seen. In other words, you can search over all the tweets of the people you follow.
“Being the richest man in the cemetery doesn’t matter to me … Going to bed at night saying we’ve done something wonderful… that’s what matters to me.” -- Steve Jobs. RIP.
With Silk, Amazon will be able to see what webpages Kindle Fire readers are reading, and they can mine that data to figure out which web pages are read together. Based on the pages that are read together, Amazon can build interesting web page recommendation system using collaborative filtering techniques.
I use Pig to write many of my Hadoop Mapreduce jobs and test my Pig script on a small dataset using Pig Grunt. Pig 0.9 has a lot more debug information as well.
Check out another similar clojure library called "MR-Kluj" that you can use to write Hadoop MapReduce jobs in Clojure: https://github.com/cheddar/mr-kluj
