amit_m

It is a credible source with a left-wing focus.

An Introduction to Statistical Learning: https://www.statlearning.com

It is a free, easy introductory text by leading researchers that covers all the classics and includes many "Lab" sections with code.

It has a chapter on deep learning but doesn't cover any of the recent advances. You will need other sources for that.

The Palestinians in the west bank and Gaza have their own healthcare systems. This was part of the Oslo accords. Plus, any Israeli medic who goes to Gaza is likely to be slaughtered. They wouldn’t trust vaccines coming from Israel anyway. Decades of war and demonization have done their part.

Israel is vaccinating the Palestinian residents of east Jerusalem since they are part of the Israeli medical system.

Needless to say that the Arab citizens of Israel (20% of Israel’s population, many of whom identify as Palestinian) have full access to vaccines.

I don't own a BMW for the signaling. I have no one to impress and it is quite an old station-wagon, so not a head turner. My wife doesn't know shit about cars but refuses to drive the Toyota on her longer commutes, because it's not comfortable for her.

So I would prefer a Toyota that handles nice over a BMW, but it would not be a Toyota. Toyota optimizes their cars for reliability and cost whereas BMW optimizes its cars for high performance and comfort. Specifically, their cars, like other German brands, are well suited for safe high-speed cruising on the autobahn, making them incredibly solid driving machines.

I happen to own a Toyota Camry and a BMW 5 series (not M5 or any other specialty model). Even for day-to-day errands, the BMW is a pleasure to drive. There is a lot of feedback and sharp responses from the steering wheel and throttle.

The Camry on the other hand is one of the least pleasing cars to drive. The suspension feels “floaty” and the steering wheel is completely dead. It also feels completely outdated and cheap and the UI is bad.

Waze started out as a community-based map building startup. They had their own maps.

In the mathematical sciences, working on old problems using classical tools is typically considered career suicide.

I can only imagine how depressing it must be to do academic research on age-old philosophical questions - an endless literature survey, at the end of which there is no light, because even if you "solve" the problem, many others will not find your solution satisfactory.

Wow, $60k CAD is really low. I don't think the smaller market is enough to explain it. In Israel the market for most startups is practically nonexistent, yet typical salaries are >$100k with Google/MSFT senior devs making more than $200k CAD.

Many of the Israeli startups are structured as dual US-Israeli companies with sales and marketing in the bay area and R&D in Tel-Aviv. Canadians could pull this stunt a lot more easily, since the geographical distance timezone and culture are much closer to the bay area.

From his FAQ: "About My Areas of Non-Expertise (e.g., Applied Cryptography)

I believe that sound practice of Cryptography has to be based on firm theoretical foundations. Still, this necessary condition is not sufficient, and expertise in applied (or practical) aspects of cryptography requires more than understanding of the theoretical foundations of Cryptography. For example, although I consider myself an expert on the theoretical foundations of Cryptography, I have very little knowledge of the applied (or practical) aspects of cryptography. In particular, please do not ask me to evaluate the security of any specific construct.

About Consulting

My above statement of limited knowledge of the applied (or practical) aspects of cryptography means that I'm not the right person to ask for consulting regarding the latter aspect. On top of this, I am not interested at all in consulting to or being involved in any commercial enterprise. My main reason is my unwillingness to make any commitment to spend time on anything other than my research and personal interests. Unfortunately, for lack of time, I properly pursue only a tiny part of my research and personal interests, and I have no intentions of making this part even smaller."

Nowadays there's no clear difference between statistics and machine learning. But the latter gets you more grant money. :-/

One of us is confused. In that table Berlin is behind in every category.

Sorry, I am not trying to "push" any evidence nor am I an expert in startup ecosystem measurement. If you don't like the "Global Startup Ecosystem Ranking 2015" data set you can link to a different data set of your choosing.

Where is the data supporting the statement "Berlin beats it in every category"?

Obviously Beijing has huge market opportunities that companies in other countries can only dream on. They will probably get to #2 in the near future, but I doubt that they are there yet. Otherwise, I can't explain the (relatively) small number of internet companies making a global impact coming out of China, nor can I explain how someone like Lee Ka-Shing (Horizons Ventures), born in China and living in Hong-Kong would make 24 investments in Israeli startups and just 1 in a Chinese startup. [1]

[1] http://horizonsventures.com/portfolio/

There's a very simple reason for leaving: the cost of living is high and the typical salaries are low. But this isn't true for the Tech industry. An experienced software engineer makes around $100k/year, 5 times the median salary.

(source: have been living in Israel my whole life, have friends who emigrated)

From a legal perspective, anyone who has at least one jewish grandparent can obtain immediate citizenship and be eligible to free healthcare, subsidized university education, etc. Historically, this was created in response to the Nazi Nuremberg laws in order to serve Israel's purpose of protecting jews from persecution. Also, on the emotional level, many jews living in the US and other places feel they have some kind of religious or cultural connection to Israel.

For people of non-jewish descent it is very hard to obtain citizenship, work visas are usually given for up to 5 years, after which you need to leave. In my opinion, this is a major weakness of the local startup ecosystem.

"largest startup hub outside the US": http://www.businessinsider.com/the-20-hottest-startup-hubs-i...

"highest concentration of startups per capita": http://www.forbes.com/sites/ilyapozin/2014/09/26/the-ten-sta...

There are also more investment dollars per capita: http://www.geektime.com/2016/01/11/annual-report-2015-startu...

For the number of NASDAQ-listed companies I queried the NASDAQ site: http://www.nasdaq.com/screening/companies-by-region.aspx?reg... http://www.nasdaq.com/screening/companies-by-region.aspx?reg...

Israel is one of the largest startup hubs in the world, with the highest concentration of startups per capita. It is typically regarded as the largest startup hub outside the US. It also has R&D offices of all the major tech companies (typically following acquisitions): Google, Facebook, Microsoft, Apple, Amazon, IBM, HP, Oracle, ebay, Yahoo, AOL, etc.

Let's take NASDAQ-listed companies for example. Israel (pop. 8 million) has 89 companies listed in the NASDAQ. Spain (pop. 48 million) has only 8. Yes, that's 66 times more per capita.

Why is it the norm that employees even need to pay cash to buy their options? Why don't companies set the exercise price to zero? Or if one must have a non-zero exercise price - why isn't there a way to do something like a "cashless exercise"? (i.e. buy your vested options by selling a small fraction of them back to the company).

Hong Kong's subway is profitable and is the best subway I've ever ridden. (fast, clean, easy to navigate)

What is the problem you are solving?

Because it is difficult to do business in the US without a US subsidiary and bank account. You can't even run a Kickstarter campaign as a foreigner.

Very quick, simple UI - I like it! Where do you pull the event data from?

Btw, you should look into adding a custom icon when bookmarking on an iphone. At the moment it just shows a screenshot of the page.

Or you can simply take a spoonful of very finely ground coffee and add boiling water for a delicious unfiltered coffee.

https://en.m.wikipedia.org/wiki/Turkish_coffee

Can anyone explain the version numbers? PyPy 2.6.1 compared to 15.11, which will eventually become 4.0.0...?

No, the helmet is not supposed to split in half. It is supposed to crush, absorbing the impact's energy.

And you would most likely not suffer serious question injury from such a fall. The horizontal speed is not so important if you hit your head on a flat surface. The vertical speed is essentially the same as if you've fainted standing still.

There is a lot of this post-hoc analysis, the main line of thought being that if the helmet broke then maybe your head would've been the one to break instead had it not been for the helmet.

This is completely false and not based on any science. Bicycle helmets are meant to crush, not to break. When they break, they absorb very little of the impact's energy.

The waste produced by a nuclear power plant is orders of magnitude less than what a coal power plant produces. Even if we only look at radioactive waste - coal power plants produce more of it, only instead of being stored in a special underground site it is dispersed into the atmosphere giving us cancer.

So yes, nuclear waste is a serious issue, but let's not forget that the alternatives have major environmental costs as well.

(not from the US)

For cars we don't really need nuclear power, rooftop photovoltaics are almost as cheap as coal and getting cheaper fast. Their main drawback is intermittency, but cars can charge overnight or in the office parking when the production of electricity exceeds the demand.

Furthermore, electric cars can be used as energy reservoirs during high demand periods.

Why go through the blurring, rescaling and JPEG compression?Simply take the first few DCT coefficients of the image, then quantize and compress them. In fact, one can make this scheme output the gaussian-blurred image by adding gaussian-weighted coefficient decay.

Opening paragraph, the honest version:

"When we launched Google+, we were scared of Facebook winning over the internet and knew how powerful the network effect is. So we decided on a douchebag move - abuse all of our monopolistic powers and superior engineering in order to shove a Facebook killer down our users' throat at maximum speed, integrating it with each and every google service in existence (whether it made any sense or not) and killing social features that actually work in exchange to experimental G+ social features that might work eventually. According to our analytics, it didn't work, so... Never mind."

All of these datasets seem to be some kind of unweighted graph with no additional information (except community information in some cases).

Does anyone know where one can find richer network data sets? i.e. a graph in which the vertices have some attributes.