I'm curious, where exactly do you think his conclusions are wrong?
I lived and worked in Beijing for over 5yrs and I think his conclusions are in general pretty accurate. Which is fairly impressive for spending only 3wks there.
The number of issues surrounding the idea (customs, legal issues, etc.) are obvious, and the discussions around them reminds me of the questions thrown at many of the other major sharing economy start-ups initially (Uber, AirBnB, et al.), but with a good team and some experimentation they will figure them out, as other major startups have done and continue to do.
People who have spent a good chunk of time in countries outside the US, already know people have been informally doing this en masse for a _long_ time. I have family in in West Africa who would be using this years ago if available, and I have lived in Asia, and other countries where every time I travel there are friends and friends of friends who ask me to buy and carry goods on their behalf, in both directions.
To me it seems Shypmate is just creating a formal system for these actions to take place, and if successful, will make it more efficient, useful, and available to more people while extracting some value for themselves (as any startup seeking profitability should).
Good luck to the team. I think there is a lot of potential here.
Beijing Love Story (北京爱情故事). Actually 39 episodes.
Don't have a great reason why I chose it other than that it was really popular at the time, and I felt it would cover general day-to-day language well, which it did.
I can speak to the power of shadowing for improving spoken language ability. If you check out techniques discussed on the How To Learn Any Language forum [1] where many of the foremost polyglots and hyperpolyglots hangout, you'll see that Shadowing and L-R are basically accepted as fundamental techniques.
It takes real determination to correctly shadow though, like patio11 said you need to reproduce the sample _exactly_. This can cause you to spend an hour trying to perfect <5min of speech. But it's extremely rewarding.
For example, when I was learning Mandarin I chose a 20 episode (1hr each) tv show as my shadowing material. It took close to half a year of working on it almost every night to finish. But afterwards I had "somehow" developed a deep sense of what words/phrases/idioms/etc _felt right_ to say when speaking and how to say them in that oh-so-close-to-native like way. The topic has been extensively discussed on the learn any language forums. But suffice to say, if you want to improve your spoken ability, doing a lot of shadowing would help.
Seems I might be the only one confused about your wording, do you mean the top 10 articles on the front page, or top 10 comments within an article?
In answer to both questions, I do both. I usually go through the articles from #1 to ~#120, reading articles I find interesting, and usually reading all the HN comments for that article too (you can learn a ton from reading more than just the top couple of comments).
Sounds like this approach would suck up a lot of time, but HN submission topics are often pretty narrow, and my interests are somewhat narrow too, so I usually only end up reading only 4-5 HN articles or less a day.
This is great. It's crazy how huge the Who is Hiring thread has become since it began on HN. Browsing it has gone from a quick scan through to a time consuming process, even using Ctrl + F and other shortcuts.
I've been looking forward to a tool that would make going through the posts easier, so very happy to see this!
Seems like DataRobot is building their team directly from Kaggle's leaderboard [1]. Yet another top Kaggle competitor the article doesn't mention who joined DataRobot is Xavier Conort, who was ranked #1 for a very long time.
I'm very curious to see how they go about automating preprocessing and identifying algorithms to use, as I thought this step was part of the hard-to-automate magic that separates your data scientists from data analyst/statistician/etc.?
As a reference for anyone reading, South Korea has a population of a bit less than 50 million. 20 million credit card details stolen means that just under half of the country's population was affected..
There is something to be said about the drawbacks of living in such a small but highly networked society (esp. in S. Korea where the entire country has a particular 'small town' feel) but I do not have the energy to organize and put my thoughts to keyboard at the moment.
I would say Coursera's "Computing for Data Analysis" and "Data Analysis" courses, taught by JHU's Robert Peng and Jeff Leek respectively, were great introductions to the field using R. Both courses are over now, but you might be able to find archived content on the Coursera sites or some of the vids on YouTube.
The courses also pointed you to many additional resources that should do a lot to supplement your learning.
After that, you should have a good foundation to self-direct your learning by studying relevant texts (like the two you already picked up) and finding data sets you can play with to just see what you can do and push your skills further.
"Nothing really replaces the experience of having a wall of books you have read and curated". I strongly second that.
As a long-time avid reader, over the past decade its been an interesting change going from a wholly physical library to mostly digital, thanks to my Galaxy Note and Kindle.
But something I truly miss is having friends come over and curiously rifle through the books on my shelf and the conversations that were started as they asked about each one (as well as the opportunity to show-off a bit).
Almost as nice an experience was, sneaking a peek at the cover of the book a guy sitting next to me on the subway/bus was reading and going on to discover that it is another good read.
So far, none of the digital bookstores/libraries have been able to sufficiently reproduce this experience for me.. though I have hope they, or something else, will be able to in the near future.
Until then, I will keep a physical copy of any favorites on my shelf for that next curious mind.
Just to create an analogy for people not familiar with Chinese web, imagine if there were 2 Youtube's in the US who were both far ahead of the competition but neck-and-neck with each other... and then one-day one acquired the other.
The deal is that big. Youku/Tudou is now THE video site of China's 1.6billion+ population.
The title of the article got me quite frightened for a moment, until I read on and realized the "gun" is more-or-less an audio playback speaker with minor delay.
Early Skype users who made calls to people using their speakers instead of headsets/headphones know how this feels.. disorienting yes, but easily ignored after you get used to it.
I never imagined the effect could be implemented as a directed energy weapon. Just makes me think it would be that much easier to re-make as a DIY project, smaller and cheaper..
Long time CJK (Chinese, Japanese, Korean) language learner here..
I'm wondering, you say it's the app you always wanted but couldn't find, but I'm sure if you looked during your Jpns studies you MUST have heard of Skritter (http://www.skritter.com/) before? They have what I would say is the best app for learning to write Hanzi or Kanji.. curious how would you say your app compares with them?
I will be checking out your app either way, great to have more options!
One of my absolute favorite short-stories. First came across it when I was on a SciFi reading rampage around the age of 8 or 9yrs old.. it completely struck me then, and still does each time I come across it again.
I owe it in part to SciFi stories and books like this for shattering any chance I had of narrow-minded thinking in life, and bringing perspective to Humans as a species and our place in the cosmos.
Probably up there amongst the most useful Python libraries IMO. Hasn't it been available for free online for a long time now though?
Anyway, in case anyone reading this missed it, the Stanford NLP class taught by Chris Manning and Dan Jurafsky starting next week (Jan 23rd) will allow programming assignments to be submitted using Python and NLTK, which is really good news.
So now's a good time to get familiar with the NLTK, or for a refresher for those of us already acquainted with it.
This is almost exactly the same path I've taken to learn Python/Django. Right now I'm at where you were at the end of Week2 (stopped at Chpt8 in DjangoBook for same reasons), and have been trying to narrow down a good next step to help me fill-in some gaps in setting up a Django site I had in mind.
'Django by Example' seems like it's what I've been looking for, surprised I missed it when gathering resources.
Good luck building your app and thanks for sharing, will be keeping watch of your blog!
I lived and worked in Beijing for over 5yrs and I think his conclusions are in general pretty accurate. Which is fairly impressive for spending only 3wks there.