cdavid

25 ms·

211.
by cdavid 8y ago
Financially speaking, the odds are incredibly against you as far as equity goes. One advantage. of working in startup is what Paul Buchheit calls working above your correct level, a kind of Peter principle in reverse ( https://tri
212.
by cdavid 8y ago
I don't know about NTT, but 5-6 millions JPY for manager-level roles in general is terrible, even outside Tokyo. W/ decent engineering skills, you should be able to get at least the double after enough experience in Tokyo. If you
213.
by cdavid 8y ago
It is not unreasonable at all. If you have 15+ years experience in domains in demand, 15-20 millions total is totally achievable, even staying in the Japanese companies. What is true is that many companies screw software engineers over, but
214.
by cdavid 8y ago
If you have a good resume and can justify good software engineering skills in demand, it is not that difficult to get to 10 millions JPY or more, even outside the big foreign companies (Google, Nvidia, etc.). Indeed in Tokyo starts paying 9
215.
by cdavid 8y ago
I am interested: do you have a reference that shows this through examples ? In my grad school years, I was always dissatisfied by matrix calculus notations
216.
by cdavid 8y ago
Interesting, I almost always got the exact opposite experience. How do you calculate non trivial matrix derivatives using index notation ?
217.
by cdavid 8y ago
That never happened to me, and I have been living there for 8 years now. But I would agree that Japan is generally racist, as in most Japanese believe there are fundamental differences between people due to what they call race. I am not sur
218.
by cdavid 8y ago
There is value when e.g. you need more than simply textual output (think plots, etc.).
219.
by cdavid 8y ago
Tokyo proper (23 wards) is much less dense than Paris, not more. The densest wards are around 20k / km2, while the least dense "arrondissement" in Paris is 17k / km2, with most of them quite above 50k / km2 and more
220.
by cdavid 8y ago
As someone joked to me, Google has too many compiler people, and this may have explained some design choices. While I am sure pytorch competition contributed to this change, having a graph-based API was a terrible idea whose time would have
221.
by cdavid 8y ago
The company I was working for before had no link whatsoever with Japan. But you're right that the wording there is unambiguously referring to the sponsoring company's salary. I applied a bit more than one year ago, I wonder if th
222.
by cdavid 8y ago
I guess by local talent you mean Japanese ? It is hard to attract them as a small company, even today. I have some anecdotal feedback on why this would be the case, but I don't know how prevalent it is. Otherwise, there are plenty of f
223.
by cdavid 8y ago
Rakuten is not the best you can aim for in Japan, at least if you are willing to live in Tokyo. It is a decent first experience, especially if you are young, but it is still very stiff compared to international companies, and salaries are l
224.
by cdavid 8y ago
Which clause are you referring to ? When I applied to the visa last year, I had to justify my salary, at that time in the UK, with proof. Checking my application, I got 40 points for having > 10M JPY in the 35-40 year range.
225.
by cdavid 8y ago
You need to make 10M JPY in your job before coming to Japan.
226.
by cdavid 8y ago
Hm, I thought I was careful of getting the numbers for the 23 wards (I live there too ;) ), but looks like I made a mistake indeed. With the correct numbers, Tokyo "23 wards" becomes 40 % higher than NY metropolis, and approximate
227.
by cdavid 8y ago
If we are talking total comp, this is somewhat the range I have observed as well. There are Japanese companies where you can expect 80k+ with only a few years of relevant experience. If you have some specialties in demand and are good at it
228.
by cdavid 8y ago
Japanese cities are actually not especially high density. For example, Tokyo's density is lower than Queens.
229.
by cdavid 8y ago
One "simple" thing that could be done is to make it easier to build python statically, and improve the freezing toolchain. When I used to care about the problem, I looked into it ( https://cournape.wordpress.com/201
230.
by cdavid 9y ago
Indeed, I am a bit puzzled by the recent articles claiming Japan is housing heaven. There are a few things to keep in mind w.r.t. housing in Japan, and especially in Tokyo: 1. Buying houses is still very expensive. One of the reason it is &
231.
by cdavid 9y ago
The writing "天ぷら" is common enough I think. I just got tempura udon a day ago, and the menu was using the kanji.
232.
by cdavid 9y ago
I was curious as well, as it smells made up and rather unlikely to be true given current wealth distribution. Where you are born (e.g. developed vs developing country) is a very strong causal factor in how much wealth you get during your li
233.
by cdavid 9y ago
(warning, I am not a numerical computation specialist) Krylov subspaces are a (the ?) common framework to compute eigenvalues and singular values, and it works for sparse matrices as well. IOW, it is extremely useful for most problems that
234.
by cdavid 9y ago
I believe ABC, one of python's influence, used tagged pointers, but I can't find a quote on this. Not using tagged pointers was a conscientious decision from Guido (see e.g. https://github.com/dropbox/pyston&#
235.
by cdavid 9y ago
Many things could be improved in NumPy: * make it easier to implement and deploy custom dtypes, fix the time-related dtype * support for ragged arrays * consolidate internals, especially around ufuncs I also think some non trivia
236.
by cdavid 9y ago
no, it is not. The OP statement is about the probability of the event "{X in Q}" (equal to 0), with X a random variable uniformly distributed on a bounded interval. That even contains many points (infinitely many actually), but ha
237.
by cdavid 9y ago
I find that result fairly intuitive, when you understand how measure theory came up to be. A much more surprising result is that most irrational are normal numbers, but we know almost no normal number (morally speaking, a normal number is a
238.
by cdavid 9y ago
That's a different statement: OP is alluding to the fact the measure of Q is 0 when using the "standard" sigma algebra on the real line, while you are saying that the measure of a number of 0. [edit] strictly speaking, you wo
239.
by cdavid 10y ago
I suspect this is only in the bay (or maybe other "high value" areas) ? I never heard such figures in the UK for example in those companies (well, for FB/Google anyway, never heard of Linkedin/netflix).
240.
by cdavid 10y ago
Any recommendation to improve 4) ?

More ›