My 2c: "Upvotes" is too noisy a metric, and will give too much importance to HN ranking. I strongly suggest tweaking the HN ranking model for "Apply HN" posts to be as random as possible.
Funny, typing "livingsocial layoffs" in google gives me these suggestions "livingsocial layoffs 2015", "livingsocial layoffs 2014", "livingsocial layoffs 2013"
What I found most heartening about the article was that it spent several paragraphs describing why they initially had to use Flash, and how they managed to get rid of it, but not a single sentence saying _why_ they decided to get rid of Flash. We've come a long way! (Also, thanks Steve Jobs!!!)
Easily the most amazing "How to Get Tenure" article that I've ever read, even though (or maybe because) it hardly talks about how to get tenure! Passion and success are so strongly correlated, and Matt Might is such a great example of this, as are most successful researchers and entrepreneurs.
They got in trouble with lawsuits because of the recent Uber employee vs contractor verdict. In addition, they were bleeding cash in customer acquisition so that didn't help too. The lawsuit was the trigger but the bleeding of cash was the fundamental reason.
It's fascinating to be able to see generational power shifts in such short cycles of time. I still vividly remember the days when M$ could never do anything wrong, and only a few years later, an upstart Google became the next M$, and Facebook the next Google, and so on. Old timers may also remember the IBM to M$ power shift, which played out over multiple decades.
I'm a data scientist, and it relieved me no end that I got this one right: http://i.imgur.com/V5oJ4i4.png I would have had second thoughts about my career choice if I got this wrong :)
The correct approach for any data modeling problem is to think in terms of entropy. Each subsequent approach should minimize entropy, until you reach diminishing returns.
What? A tool that helps discover inefficiencies in Hive/Presto/Dremel query/pipeline/scripts.
Why?
It is so much easier to just add new machines to your cluster, than to optimize your code and fix inefficiencies. But the latter option typically results in millions of $$$s in savings.
This would have sounded credible if Apple didn't have an iAds business that mines user clicks and offers retargetting ads to paying customers.
Instead, now, this sounds like Microsoft's Scroogle campaign, where it accused Google of doing all the things that Bing also did, but wasn't successful at doing.
The most important skill, in order to be data-driven, is to ask the right questions. If you're looking to get to product market fit, the questions you should be asking are very different from the ones you should be asking if you're looking to grow a "good" product. In both cases, data can help you reach your goal, but only if you ask the right questions.
If an early-stage startup tries growth hacking before it reaches product market fit, it will likely end in disaster.
I'm curious to know: is this sentiment widely shared? Is git really that much better than hg?