Guess Ecole Poly is good enough. Is ENS more like the Harvard/Oxford, and EP more like the MIT/Cambridge of France? This would be my layman reading of the situation, charitably not invoking politics.
EP and ENS are both excellent schools. I am a mathematician and mathematics is more traditionally associated with ENS than with EP, but I've met excellent mathematicians, at all level of career (from PhD student to full professor) from both schools.
My impression is that all the French Grandes Ecoles are trying to expand their range of activities and subjects, reducing the level of diversification.
Well I guess "more traditionally associated" is one way to put it, but... 11 out of France's 12 fields medal came from people who studied at ENS. The only other country to even have 11 or more fields medal is the USA at 13. They literally exude math competency.
Having been in the selection in grande ecoles - and having been in one of those myself - EP is definitely a school of engineers and ens a school of researchers/professors. All of our teachers were from ens. Not that there aren't researchers in so or my own grande ecole but I have seen people refuse to go to ep to attempt and the next year. But those schools really are top notch, each in their own way.
Isn't a bit of the appeal of elite institutions that they put elite people in close proximity to each other? Adding a lot of padding* in between them would make it harder for them to find and get to know each other.
*I am fully aware that in this context I am considered padding
42 doesn't have anything close to those schools regarding general science curriculum. No serious maths, no chemistry or physics , no engineering. It's a very good school if you want to learn to code in 3 years, but let's not make it anything else...
I doubt it. Laboratories which teach and work in English have a huge advantage in modern academia because they are able to draw in talent from a far larger pool.
The majority of international academics working in France are from other Romance language countries. I presume the other Romance language countries have the same problem as France.
Having done research in both the U.S. and France, the latter is far better simply because of the general life quality.
Language is not really that big of a factor. The U.S. definitely provides better research ressources but when you look at your life as a researcher as a whole, France is the better option.
French is an extremely easy language to pick for someone who already speaks English.
In fact, English took from French over 30% of its vocabulary. French was the official language in England of the highly educated until mere centuries ago.
Most researchers in France already speak English.
The huge advantage of the US/UK is one: money, you could become rich and apply your ideas to real products(in a bigger market) much much more easily than in France.
And much less pressure to commercialize, there is a lot of fundamental and other research that is not immediately aimed at generating a cash-cow somewhere.
France already has an excellent research system, and the phrase "create its own MIT" is ambiguous at best: This is mainly about branding to increase international visibility of the institutions.
CentraleSupelec is France's best attempt to match MIT (more than 1000 students each year). It was formed from the merger of Centrale Paris and Supelec schools.
And it's a unique school, not like this merger: it's a reunion of several schools that want to keep their independance. I think it'll be harder for them to act in coordination to achieve their goals.
What! The land that gave the world Fourier, Poincare, Galois, Laplace, Pascal, Lagrange, Cauchy, Descartes, Fermat, Hadamard, Borel ... they had their MIT a long time earlier.
Which is meaningless in the present. Just ask Greece how valuable to your present circumstances basking in the very, very distant past is.
Australia has none of that immense legacy, due to the age of their nation. And yet Australians are already richer than the French at the median, along with having a far higher GDP per capita. The impressive history of France doesn't matter very much in the scope of most practical matters related to the present.
> Which is meaningless in the present. Just ask Greece how valuable to your present circumstances basking in the very, very distant past is.
Slightly off-topic, but for some reason I've run into a seemingly disproportionate (in terms of world population) number of good computer scientists who were born in Greece.
A Greek professor told me that when he started his PhD (in the US) his Greek advisor told him as a Greek in America his two options were to become a control theorist or open a diner.
> And yet Australians are already richer than the French at the median, along with having a far higher GDP per capita.
I agree with your main point, caveat that I don't know anything about the current state of French science because I don't know any French.
With this aside though, Australia has an order of magnitude more land than France, and half as many people, so individual Australians should have an absurd amount of raw resource available to them. In a sense, Australia should have a GDP per capita 20x France's purely on raw resources.
Comparing material wealth and academic success isn't really fair. Academics depends on small communities, and is a highly cultural endeavor. Wealth depends on resources and stability, which is less so.
Not really, because raw land is valuable for three things: agriculture, real estate, and natural resource extraction, and the first two are virtually worthless in deserts which is most of Australia’s land.
> France has its own MIT, it's called École Polytechnique
Is EP’s productivity (in terms of companies founded by, landmark inventions/discoveries made by or patents awarded to) comparable to MIT? I thought it was more geared towards producing future politicians.
This is a good measure: divide number of Nobel winners by total student population. By that standard, EP seems to be top ranked. The problem is that the last EP student who won the prize was enrolled in 1960s. Many other winners were all the wY back in 1800s. It seems that EP was one of the best institute before WW II but then something happened and things went downhill. This could be great case study.
My answer is mostly with respect to CS which I know best, although it does apply to other fields.
This is in no way arrogance. France has a history of long-time academics that are totally inactive but take up a faculty line for decades. This means a lot of the research is just well behind the times.
On top of it all most of the research gets published in 3rd tier or worse places. Often it gets published in no-name workshops or journals in French.
It's so bad a student cannot get a non-pity PhD at MIT by publishing in the venues where most of the papers from EP (and to a somewhat lesser extent INRIA) go to.
If you think EP is anywhere like say.. any of the top 10 institutions in the US/Canada/UK you're ignoring the facts. They're even falling behind China now and will in a decade or two be behind institutions in India too if they keep going like this.
For example. EP has basically no presence in AI or robotics at any major conference. Most faculty members in EP don't publish 1 paper every year in a reputable conference or journal.
INRIA is more international, they suffer somewhat less from these problems. They're also half the size of MIT at best and heavily skewed toward fields that aren't hot anymore.
In case someone doesn't believe my observations from conferences.
Just look below. Neither INRIA nor EP are in the top 30 institutions with publications at ICML and only INRIA is in the top 50 (at 14, not bad!) at NIPS. EP isn't in the top 50 institutions with papers at NIPS.
Out of curiosity, why is this true? Why does it seem like so many relevant tech companies are American, yet I see tons of talented European developers and conferences all the time? Are Euro nations just hostile to startups in a way the US isn't?
This is a cynical take, but, roughly speaking, in Europe, there are workers rights, and in the US, employment is at-will.
The US systems tilts the scales toward stronger employees. So, the more marketable your skills are, the more incentive you have to move to SV.
A slightly less offensive take on the situation is that, if you stay in Europe, you won’t be destitute, but you won’t get rich either.
However, note that there are many articles on HN that attempt to quantify this effect by looking at economic mobility. The consensus is that Europe is becoming more of a meritocracy and in the US, economic status is increasingly determined at birth.
> Why does it seem like so many relevant tech companies are American
The ones that are relevant to you.
To me all but very few of the relevant tech companies are irrelevant, and the miracle of Silicon Valley has not been replicated inside the USA either. So the reason those companies are American is because SV is in the United States, it has little to do with America as a whole, if it did then Silicon Valley would not be an exception within the United States.
Euro nations as a rule are not hostile to start-ups in any way that is different from say Texas or Michigan.
Then, finally, there are a large number of relevant EU tech start-ups. They are working from a slight disadvantage (need to market/sell immediately in a whole bunch of languages if they want to escape their home-territory) but plenty of them do well enough that they end up listed on stock-exchanges all over the planet.
That they are not relevant to you is another matter, but IT companies account for a huge percentage of EU employment and they're all doing stuff that they feel is relevant.
So it looks as if your argument mostly revolves around what you think is relevant and what isn't.
The only American company I can't seem to get around is Google, and they're a one-off even by US standards.
Apple, FB, MS, Twitter and so on can all be done without (at least: I could do without them).
On the hardware front that leaves Intel and AMD, neither of which could be classed as start-ups today any more than you could class Siemens or Sony as one.
You're right they don't seem relevant to me, but I can acknowledge some things are not relevant to me but are still important. Take Baidu for instance, I get I'll never likely use it, but that doesn't mean I don't acknowledge its importance. Looking at software vendors in the EU, there doesn't seem to be any companies that offer a service I would ever want. There are a few large ERP vendors, and some Anti-virus firms, and what?
Seriously, this isn't meant to be mean, I just don't know any of them. There seems to be an unlimited array of talented euro engineers on twitter talking about specialized type systems and exotic data models and theoretical comp sci and then a net nothing on the software front. Do they all just leave and work in SV?
This is one of those cases where I would love someone to come in and set me right. I'd much rather a world where there is a huge amount of cool things happening that I'm unaware of, rather than just MS, Apple, and Google sucking all the engineers from their countries to Silicon Valley.
I can't tell you it's the same cause, but you can see this phenomenon in many industries, such as entertainment. After WWII, the U.S. was the only wealthy country left untouched and it dominated the world. Around half of world GDP was produced in the U.S. (an astounding number); all the leading scholars and scientists had moved to the U.S. (for example, think of the language and the location of pre- and post-WWII physics); the U.S. also became the world's dominant military and political power, as well as the leading cultural power (from democracy and human rights to Hollywood films to jazz to Elvis). As a result, if you wanted to be a world leader in almost anything, you had to do it in the U.S.
Also, the international language of (seemingly) everything became English, from science to, importantly in this context, computer code. When a Finnish graduate student created his new operating system, the code and his announcement were in English. Assuming you are a native English-speaker, imagine if programming languages, HTML, and network traffic were in French - how many fewer English-speakers would step over that hurdle and learn to code, to hack? How many teenagers would give it a try out of curiosity if they have to learn another language to do it? Imagine you are a French musician; what is the chance that a song you record in French will be heard in the U.S. or worldwide? Ever hear a French song in the U.S.? See French musicians on major tours in the U.S.? Can Americans name one French musician? But can people in France name American musicians? Do English-speaking U.S. musicians tour France? All the time, AFAIK.
Those things persist today, U.S. businesses still have an economic advantage; they can sell to an internal market of 320 million people, all in one language - and at ~$19 trillion annual GDP the largest internal market in the world by far (not counting the EU, which is a complicated topic). Then they can sell to another 100 million+ in other wealthy, English-speaking countries, and then to the rest of the world accustomed to hearing pop songs in English and watching movies in English. A French business has similar access to 60 million, plus part of Canada.
IME, Americans often don't realize how the modern world was built around them.
The history of silicon valley is fascinating. Back in the WW2 days, ampex (some would argue they were the first SV tech company) was founded in CA specifically to avoid the corrupt regulatory machine the GE and friends had created on the east coast.
Eventually old tech figured out what was going on, and almost got them shut down for patents or something. Fortunately, by the time that happened, it was clear ampex was crucial to the war effort.
At that point, the war machine somehow managed to override the politicians, so the US had significant radio communications advantages over its enemies.
This helped the war effort, but it also helped create Silicon Valley’s traditional deep-seated mistrust of Washington, and Washington’s hands-off attitude about CA tech companies.
Eventually, self-imposed regulatory capture strangled east coast tech to death, so all that’s left are universities and (now) NY’s bootstrapping startup scene (and, of course, the military tech complex near DC).
The sad part is that the current generation of Silicon Valley giants are consciously trying to dismantle the whole thing by lobbying DC, and writing regulatory capture bills like FOSTA and the CLOUD act (and many earlier examples).
Google led the charge at this years ago. After watching the MS anti-trust thing, they hoped they could preempt such “misunderstandings” with piles of lobbyist cash. (Source: their cofounders, years ago. They didn’t use the word “dismantle”, in fairness.)
It will he interesting to see what country will have enough common sense to replicate SV over the next 50 years.
In the article, it sounds like France is trying to create a low-regulation academic and economic zone for this new university push.
China keeps bouncing between “no law enforcement” and “extreme, protectionist regulation”, which sounds fundamentally unstable to me.
Also, there are probably a dozen other geographical regions playing in this space.
Anyway, to answer your question, Europe’s industrial base has historically been even more protectionist (can’t fire people, product specs are legislated, etc) than the US East coast, which is why that region won the mainframe and minicomputer wars.
I think there is a broader theme, where older governments do more and more things to protect the wealthy and borrow money to provide social services until the economy simply collapses.
France has a problem with long lived institutions - they get filled with "union tenured" staff and new work suffers. I have heard that the French are well aware of this problem and when they establish an institution they set the budget and longevity for it. At the end of that time it is shut down and all the released people must find new jobs based on their merits. A long process to get rid of dead wood, which in the USA would be long gone.
They are missing an important structural point: MIT isn't really a school, it's a huge government research lab with a small school attached (look at the budget for which education is about 14% of revenues and 16% of expenditures). Research is primary (whereas, as some other comments have noted, it has atrophied in the French Grandes Écoles). They need another INRIA in the mix -- though not INRIA itself; that excellent institution would not survive merging with these others.
FWIW I am an MIT graduate who has lived and worked in research in the USA and France
I'll take any EP hacker over any MIT hacker. They are just better. Anyone heard of OCAML, compilers, netsec or security engineering? That's where they are better.
50 comments
[ 2.7 ms ] story [ 116 ms ] threadMy impression is that all the French Grandes Ecoles are trying to expand their range of activities and subjects, reducing the level of diversification.
ENS and l'X (Ecole Polytechnique) in particular hold special statuses (each in their respective way) vis-a-vis the French state.
Isn't a bit of the appeal of elite institutions that they put elite people in close proximity to each other? Adding a lot of padding* in between them would make it harder for them to find and get to know each other.
*I am fully aware that in this context I am considered padding
The majority of international academics working in France are from other Romance language countries. I presume the other Romance language countries have the same problem as France.
Language is not really that big of a factor. The U.S. definitely provides better research ressources but when you look at your life as a researcher as a whole, France is the better option.
In fact, English took from French over 30% of its vocabulary. French was the official language in England of the highly educated until mere centuries ago.
Most researchers in France already speak English.
The huge advantage of the US/UK is one: money, you could become rich and apply your ideas to real products(in a bigger market) much much more easily than in France.
The huge advantage of France: quality of life.
And it's a unique school, not like this merger: it's a reunion of several schools that want to keep their independance. I think it'll be harder for them to act in coordination to achieve their goals.
Australia has none of that immense legacy, due to the age of their nation. And yet Australians are already richer than the French at the median, along with having a far higher GDP per capita. The impressive history of France doesn't matter very much in the scope of most practical matters related to the present.
Slightly off-topic, but for some reason I've run into a seemingly disproportionate (in terms of world population) number of good computer scientists who were born in Greece.
I agree with your main point, caveat that I don't know anything about the current state of French science because I don't know any French.
With this aside though, Australia has an order of magnitude more land than France, and half as many people, so individual Australians should have an absurd amount of raw resource available to them. In a sense, Australia should have a GDP per capita 20x France's purely on raw resources.
Comparing material wealth and academic success isn't really fair. Academics depends on small communities, and is a highly cultural endeavor. Wealth depends on resources and stability, which is less so.
France has its own MIT, it's called École Polytechnique. There's also INRIA and other institutions...
Is EP’s productivity (in terms of companies founded by, landmark inventions/discoveries made by or patents awarded to) comparable to MIT? I thought it was more geared towards producing future politicians.
https://en.m.wikipedia.org/wiki/List_of_École_normale_supéri...
This is in no way arrogance. France has a history of long-time academics that are totally inactive but take up a faculty line for decades. This means a lot of the research is just well behind the times.
On top of it all most of the research gets published in 3rd tier or worse places. Often it gets published in no-name workshops or journals in French.
It's so bad a student cannot get a non-pity PhD at MIT by publishing in the venues where most of the papers from EP (and to a somewhat lesser extent INRIA) go to.
If you think EP is anywhere like say.. any of the top 10 institutions in the US/Canada/UK you're ignoring the facts. They're even falling behind China now and will in a decade or two be behind institutions in India too if they keep going like this.
For example. EP has basically no presence in AI or robotics at any major conference. Most faculty members in EP don't publish 1 paper every year in a reputable conference or journal.
INRIA is more international, they suffer somewhat less from these problems. They're also half the size of MIT at best and heavily skewed toward fields that aren't hot anymore.
Just look below. Neither INRIA nor EP are in the top 30 institutions with publications at ICML and only INRIA is in the top 50 (at 14, not bad!) at NIPS. EP isn't in the top 50 institutions with papers at NIPS.
https://medium.com/@karpathy/icml-accepted-papers-institutio...
https://medium.com/machine-learning-in-practice/nips-accepte...
EP isn't a player in academic CS basically. (the numbers work out to be the same if you attend CVPR or ICRA or any number of other conferences).
France is doing just fine in terms of research in many fields.
Those that would like France to produce an 'MIT' would also like Europe to produce the next Facebook and neither is going to happen.
But that doesn't mean things are bad, they're just different and no less valuable.
What are the advantages to starting a startup in France vs joining an established company?
The US systems tilts the scales toward stronger employees. So, the more marketable your skills are, the more incentive you have to move to SV.
A slightly less offensive take on the situation is that, if you stay in Europe, you won’t be destitute, but you won’t get rich either.
However, note that there are many articles on HN that attempt to quantify this effect by looking at economic mobility. The consensus is that Europe is becoming more of a meritocracy and in the US, economic status is increasingly determined at birth.
Poor Americans have been hammered over the last two decades - and have now put their faith in Trump. Let's hope he can fix the system.
> This is a cynical take, but, roughly speaking, in Europe, there are workers rights, and in the US, employment is at-will.
What's counter-intuitive to me is you would think in a system with more workers rights the workers would take more risks - but the opposite is true.
I'm also surprised that in countries in Europe with huge youth unemployment we don't see outsized open source contributions.
The ones that are relevant to you.
To me all but very few of the relevant tech companies are irrelevant, and the miracle of Silicon Valley has not been replicated inside the USA either. So the reason those companies are American is because SV is in the United States, it has little to do with America as a whole, if it did then Silicon Valley would not be an exception within the United States.
Euro nations as a rule are not hostile to start-ups in any way that is different from say Texas or Michigan.
Then, finally, there are a large number of relevant EU tech start-ups. They are working from a slight disadvantage (need to market/sell immediately in a whole bunch of languages if they want to escape their home-territory) but plenty of them do well enough that they end up listed on stock-exchanges all over the planet.
That they are not relevant to you is another matter, but IT companies account for a huge percentage of EU employment and they're all doing stuff that they feel is relevant.
So it looks as if your argument mostly revolves around what you think is relevant and what isn't.
The only American company I can't seem to get around is Google, and they're a one-off even by US standards.
Apple, FB, MS, Twitter and so on can all be done without (at least: I could do without them).
On the hardware front that leaves Intel and AMD, neither of which could be classed as start-ups today any more than you could class Siemens or Sony as one.
Seriously, this isn't meant to be mean, I just don't know any of them. There seems to be an unlimited array of talented euro engineers on twitter talking about specialized type systems and exotic data models and theoretical comp sci and then a net nothing on the software front. Do they all just leave and work in SV?
This is one of those cases where I would love someone to come in and set me right. I'd much rather a world where there is a huge amount of cool things happening that I'm unaware of, rather than just MS, Apple, and Google sucking all the engineers from their countries to Silicon Valley.
Also, the international language of (seemingly) everything became English, from science to, importantly in this context, computer code. When a Finnish graduate student created his new operating system, the code and his announcement were in English. Assuming you are a native English-speaker, imagine if programming languages, HTML, and network traffic were in French - how many fewer English-speakers would step over that hurdle and learn to code, to hack? How many teenagers would give it a try out of curiosity if they have to learn another language to do it? Imagine you are a French musician; what is the chance that a song you record in French will be heard in the U.S. or worldwide? Ever hear a French song in the U.S.? See French musicians on major tours in the U.S.? Can Americans name one French musician? But can people in France name American musicians? Do English-speaking U.S. musicians tour France? All the time, AFAIK.
Those things persist today, U.S. businesses still have an economic advantage; they can sell to an internal market of 320 million people, all in one language - and at ~$19 trillion annual GDP the largest internal market in the world by far (not counting the EU, which is a complicated topic). Then they can sell to another 100 million+ in other wealthy, English-speaking countries, and then to the rest of the world accustomed to hearing pop songs in English and watching movies in English. A French business has similar access to 60 million, plus part of Canada.
IME, Americans often don't realize how the modern world was built around them.
Eventually old tech figured out what was going on, and almost got them shut down for patents or something. Fortunately, by the time that happened, it was clear ampex was crucial to the war effort.
At that point, the war machine somehow managed to override the politicians, so the US had significant radio communications advantages over its enemies.
This helped the war effort, but it also helped create Silicon Valley’s traditional deep-seated mistrust of Washington, and Washington’s hands-off attitude about CA tech companies.
Eventually, self-imposed regulatory capture strangled east coast tech to death, so all that’s left are universities and (now) NY’s bootstrapping startup scene (and, of course, the military tech complex near DC).
The sad part is that the current generation of Silicon Valley giants are consciously trying to dismantle the whole thing by lobbying DC, and writing regulatory capture bills like FOSTA and the CLOUD act (and many earlier examples).
Google led the charge at this years ago. After watching the MS anti-trust thing, they hoped they could preempt such “misunderstandings” with piles of lobbyist cash. (Source: their cofounders, years ago. They didn’t use the word “dismantle”, in fairness.)
It will he interesting to see what country will have enough common sense to replicate SV over the next 50 years.
In the article, it sounds like France is trying to create a low-regulation academic and economic zone for this new university push.
China keeps bouncing between “no law enforcement” and “extreme, protectionist regulation”, which sounds fundamentally unstable to me.
Also, there are probably a dozen other geographical regions playing in this space.
Anyway, to answer your question, Europe’s industrial base has historically been even more protectionist (can’t fire people, product specs are legislated, etc) than the US East coast, which is why that region won the mainframe and minicomputer wars.
I think there is a broader theme, where older governments do more and more things to protect the wealthy and borrow money to provide social services until the economy simply collapses.
FWIW I am an MIT graduate who has lived and worked in research in the USA and France