cdavid

17 ms·

181.
by cdavid 6y ago
So manufacturing is generally quite efficient. Toyota would be one of the powerhouse I had in mind, as would be most car companies. Note that Toyota is very much a global company and under the pressure of competition worldwide. But many la
182.
by cdavid 6y ago
It traditionally isn't. Japanese middle management practices would often make office space look good. Especially in software, the culture is quite antiquated, because software is traditionally seen as a cost center, and most companies
183.
by cdavid 6y ago
I am quite bearish about the impact of WFH, especially for software engineers, but if there is one country where I think the coronavirus can have a huge positive effect for the economy, it is Japan. Japanese companies are generally extremel
184.
by cdavid 6y ago
As the OP mentioned, many countries, including democratic ones, do have laws around hate speech. Wikipedia has a few words about it in English: https://en.wikipedia.org/wiki/Hate_speech_laws_in_France#The... You also h
185.
by cdavid 6y ago
I agree 100 %, and I started as a PhD in ML 15 years ago. DL, thanks to its "lego aspect", is getting ML commoditized fast for most applications out there. It hollows the gap between applying more or less directly existing models
186.
Big Tech Won’t Be the Same If Everyone Works from Home(bloomberg.com)
1 points by cdavid 6y ago|0 comments
187.
by cdavid 6y ago
The US are not exactly regulation-free. Actually, I remember seeing people in HN who mentioned legal costs for companies tend to be much higher than in (continental) Europe. The cost of capital is indeed one of the core issue. In Europe, al
188.
by cdavid 6y ago
Global and a few Japanese companies in Tokyo pay well compared to Europe. E.g. you can make 200+k$ total comp if you have skills in demand (ML, etc.), which is hard in Europe outside management consulting/finance. How that compares to
189.
by cdavid 6y ago
I would say that you don't need it if you don't feel hindered in your work. But it is solving a real need. Problems arise when you start to have larger teams who interact with models, when the models are applied in production at a
190.
by cdavid 7y ago
It is possible to have what I think you mean by automatic response level without living in the country. I myself sucked at English until University: I hated learning it, I could not understand it spoken, and I could not read well a mildly c
191.
by cdavid 7y ago
The topic is complex, but France was (and still is) a very centralized country, compared to the UK, Netherlands, Denmark, or what would become Germany. And you had fewer, bigger land owners. It is often argued that it is only through compet
192.
by cdavid 7y ago
Necessary but not sufficient and all that... In how Asia works, Studwell argued quite convincingly how land reforms in Japan, South Korea, etc. was a prelude to successful industrialization because it allowed creating financial surplus than
193.
by cdavid 7y ago
I see how you could interpret his post that way, but saying that Cowen does not actually read books is... not knowing the guy very well :) His perspective I believe is that a lot of those points have been studied, but need a fresh look, and
194.
by cdavid 7y ago
Google had little influence on scientific python. I agree python did not have technically much advantage over ruby, but: 1. Old history of using python for scripting linux, starting from mid-late 90ies, especially around clusters and HPC,
195.
by cdavid 7y ago
I think that's exactly his point, actually, although I thought that was quite well understood. E.g. "How Asia works" explained the difference between taiwan/japan/s korea on one hand, and philipines/indonesia&#
196.
by cdavid 7y ago
Most of the ML useful in industry is getting commoditized at a fast pace, and not enough people understand that. The focus on modeling for applying ML in a business environment is as incongruous as people believing programming is graph algo
197.
by cdavid 7y ago
true to a point, but difficulty to transition is real. I would be the first to agree that a lot of the elements in the remote manifesto are good things in absolute (written down processes, etc.). In fact, that's one of the first thing
198.
by cdavid 7y ago
Yeah, the advantages are obvious and real, especially for individual contributors. I am myself quite pessimistic about the reality of a massive change there because of the points I raised, and the ability to do commute < 30 minutes has a
199.
by cdavid 7y ago
> All 7 companies were process driven companies, with discipline. The processes were not overly complicated, nor bureaucratic in nature, but they were followed religiously. If the process wasn't working everyone still followed it, b
200.
by cdavid 7y ago
1. I am the first to recognize it is hard to quantify, it is just my experience. I should note that I have always worked in fairly diverse environments, at least culturally (e.g. in my current team, I have people from France, the US, Austra
201.
by cdavid 7y ago
Did you write anything about remote team management ? The "value proposition" for remote work for IC is obvious to me, but much less so from a management perspective.
202.
by cdavid 7y ago
> In a remote-only organization, communication is necessarily all done in a matter that's transparent and not transient like face-to-face communication. You claim doing so remotely is more difficult but do not substantiate this. You
203.
by cdavid 7y ago
Yes, performance management is really hard (remote or not): the reality is that it is actually quite difficult to measure the output of software engineers. There are definitely overly bad measures (LOC, etc.), but things like mentoring, qua
204.
by cdavid 7y ago
As a former software eng. who has been on the management side for a few years, I am quite ambivalent about remote work, and I find most arguments for it a bit naive. For example, the idea that the manager schedule is built around taylorism
205.
by cdavid 7y ago
Senior Machine Learning Engineers, Mercari | Tokyo, Japan, VISA | onsite | https://mercari.com Responsibilities As a Machine Learning Engineer at Mercari, you are going apply your machine learning skills to develop Mercari Consu
206.
by cdavid 7y ago
Since I have been in a position to do so, I tried to push for coding challenges, and base tech IV on the assignment itself. Your blog post about hiring has been my biggest inspiration on doing hiring. I hired 10+ people in 1.5 years this wa
207.
by cdavid 7y ago
Putting all the "French philosophers" in one bucket is not fair. Lacan is clearly a charlatan, Foucault, while controversial, is absolutely a deep thinker. There is some truth about their writing being too obtuse, but this is (was
208.
by cdavid 7y ago
I agree, and that's why I find wanikani so useful. After 7-8 years trying to learning kanji and always failing after 3-400, I have been using wanikani for ~ 6 months, and the system of building vocabulary on top of kanji on top of radi
209.
by cdavid 7y ago
The solution depends a lot of the problem at hand. Many people focus on the data size saying it is "trivial", but depending on what kind of data and the format, it may not be trivial. I have e.g. inherited a project where we have
210.
by cdavid 8y ago
I am aware of the dread of the corporate life in large Japanese companies, and that many pay really badly. The numbers I have quoted are from people who worked in well known Japanese companies (Canon, Nintendo, etc.) and non Japanese (Pivot

More ›