Hacker News new | past | comments | ask | show | jobs | submit login
Scientists plan huge European AI hub to compete with US (theguardian.com)
213 points by charlysl on April 24, 2018 | hide | past | favorite | 190 comments

I believe that a contributing factor to the pay gap in Europe is the fact that we have so many languages, that we are too segregated to create the same level of competitiveness as seen in the US. This means that each country is fairly protective of its people, and even though EU has free movement of labour, the language and culture barrier is strong, and internationals have a harder time to get hired, unless the speak the local language.

So the end results is that talent is spread across, and no single country has ability to become the place to be.

It also means that despite being a single market you can’t realistically launch a product across the EU. A SV startup can immediately address the whole US (if they don’t have a physical presence), but an EU startup has to go country by country. That means growth is slower, and talent is less competed for.

After brexit I’d like to see business & gov across the EU standardize on english as lingua franca to enable faster growth.

> After brexit I’d like to see business & gov across the EU standardize on english as lingua franca to enable faster growth.

I wonder whether this will become easier now when the UK leaves, because the English language won't be tied to the competition between the major states.

This is definitely a shame for London, which I think was emerging as the centre of English-speaking Europe, I expect that role now will be more spread across a number of northern European cities, like Stockholm, Copenhagen, Amsterdam etc.

As someone living in a European country with several languages and internal political problems related to them, i can say that in my opinion this wont happen for a very long time and im not even sure its needed. Kids in these multi language countries are already growing up with english as a second or third language. They are growing up on the internet and are used to reading stuff in english, so the need to standardize on a language, in my opinion, is not even needed. In a generation or 2 the majority of people will speak it anyway.

The problem is that half of the EU has no common language (and that ratio is worse once you step outside the EU and broaden it to Europe as a whole).

Certainly Scandinavia etc. have tremendous English adoption, that doesn't do much for the two dozen nations in Europe that don't. If you're Swedish or Danish and you are fluent in English, the problem is that still doesn't help you with the other half of the EU or Europe that doesn't understand English.

France is near 40%, and Italy is near 35% on English speaking. Even Germany is just barely over 1/2.

Spain, Portugal, Czech, Slovakia, Ukraine, Bulgaria are near or sub 25%. Russia is closer to 5%.

France + Spain + Italy is about $6 trillion worth of GDP, and maybe 1/3 of their combined population is English speaking.

That's what the parent post was speaking about - for various reasons these percentages are increasing, especially among the younger people, and the expectation is that in a generation or two the situation would be quite different from the current one that you've described.

Well, in CEE (EU bit) 80% people that are <35 y/o certainly speak solid English - I've been and spent some time in all of them.

The older generation does struggle with English, though.

London will still be the centre of English-speaking Europe. It's the biggest irony of brexit: London is vehemently opposed to it, but likely not going to be all that affected by it.

Dunno, most people I know are trying to escape London, miserable weather, low pay, massive increase in rent, uncertain future. Despite having most AI startups after US. At FB London as a SWEng you are capped at most at 110k; in the US you'd earn 4-5x more for the same job and on more interesting projects as well.

Further anecdote to support that claim: in my old job an admin who worked in the London office for over seven years, and had numerous raises and promotions in that time, was on the same salary as her newly hired graduate counterpart in the states. Relatively rural area in the states too, it wasn't NYC where the cost of living may have made that acceptable.

Our relative salaries with the states are god awful.

Nowadays even with Germany. UK is going backwards fast...

110k$/year? I really find it difficult to believe it, even if it was 110k£/year.

I worked for Facebook London office and the pay level compared to our US counterpart was something that my more talented peers were talking about. The idea of a ‘cap’ feels alien to Facebook in general, but I can confirm that to get more than that, you would probably have to be an exceptional developper -- think: get recognised at a conference, have your work #1 on HN for 24 hours, have regular conversations with iOS or Android core devs, etc.

> have your work #1 on HN for 24 hours

Done. Still didn't go there and they keep spamming me ;-) London is not attractive at all these days.

I joined at the same time as one guy who was working remotely full time, officially based in London. But he was the co-founder and the lead developper of the most used technology at Facebook, so…

>After brexit I’d like to see business & gov across the EU standardize on english as lingua franca to enable faster growth.

Do you imply that brexit might make it easier for that to happen? That sounds counter-intuitive to me, once the UK leaves I believe that the only country in the EU who has English as an official language will be Ireland and even there it's cohabiting with Irish Gaelic.

Beyond that I've always been torn on this issue, on one hand having a lingua franca across the EU would be amazing and English is probably closest to achieving that, on the other hand English is effectively the language of American imperialism and its cultural hegemony. In an ideal world I'd prefer something less politically loaded and tied to a foreign superpower, like Spanish, Swedish or Romanian for instance (I'm picking random European languages who cannot be suspected of having any kind of cultural overreach in present day Europe).

Of course in this case maybe practicality trumps ideology and we should just accept our English speaking overlords for the sake of convenience.

> English is effectively the language of American imperialism

I know this is often said, but why? English did not originate in America, but in England and even there it originated as a mix of various other European languages (Latin, French, Nordic languages and a few more). If anything, it represents the colonial past of the United States.

I think in the EU the main problem is that every country is supposed to have an equal part in the community and choosing the language of one country in the union would contradict that. Simply said: If England leaves the EU, then nobody can say that one of the countries has a language advantage, as everyone has to learn English as a second language. Ireland does not count because they have Irish as the first official language.

Furthermore, adding to my first point, since many places around the word start to speak English for professional communication, it represents the US less as time goes by.

>I know this is often said, but why? English did not originate in America, but in England and even there it originated as a mix of various other European languages (Latin, French, Nordic languages and a few more). If anything, it represents the colonial past of the United States.

That's irrelevant and I doubt many people in America considers English the language of the colonial past. The reason English has become the lingua franca for businesses around the world is evidently because the USA became a superpower and managed to gain a huge worldwide influence in both economy, military and cultural sectors. Like most Europeans my age I've grown up eating a huge dose of American cultural goods.

Captain America is in theaters around the world. Katty Perry and Kanye West play on the radios in the middle east. You can watch American politics play out on your TV in Madrid while you sip on your Coca Cola. Or maybe you prefer to watch Game of Thrones while eating your Big Mac? Don't forget to post about it on Facebook, Twitter or Reddit using your iOS or Android phone. Many artists produce music in English instead of their native language in order to sound more modern and reach a wider audience. Yesterday I noticed that they didn't even bother translating the title of Stephen King's latest novel for the French version, French readers are expected to understand it. Can you imagine that happening with a German or Russian novel? American culture spreads over the world like no other, I don't think its hegemony is debatable at this point.

>Simply said: If England leaves the EU, then nobody can say that one of the countries has a language advantage, as everyone has to learn English as a second language.

So in order not to advantage any country in the EU we'd advantage the USA, UK and Australia by doing their job for them? That's cutting off the nose to spite the face, although unfortunately I can see that happening. I'm also not convinced that you can pretend that Ireland doesn't speak English solely because it also recognizes Irish as an official language. Wikipedia tells us that `Less than 10% of the population of the Republic of Ireland today speak Irish regularly outside of the education system and 38% of those over 15 years are classified as "Irish speakers"'. There might be more Spanish speakers in France than Irish speakers in Ireland.

>Furthermore, adding to my first point, since many places around the word start to speak English for professional communication, it represents the US less as time goes by.

I'm not convinced, the more the world speaks English the easier it is for the USA to spread their media and, indirectly, their message and propaganda. And the USA is incredibly good at that already. We call it "English" but let's be real, it's really USA'an that's taking over the world.

> The reason English has become the lingua franca for businesses around the world is evidently because the USA became a superpower

No. It was because of the globe-spanning, imperialist behemoth that was the British Empire whose scale and reach at its peak the USA has not yet matched. The empire really waned with the baby boomer generation, so it was not too not so long ago and certainly within living memory.

America's cultural hegemony with Hollywood (and pop music) came much later and built on it.

France's influence was also much greater than it is now, as can be seen by prominence of the French language and culture around the world today. Yet it's paltry compared compared to American influence. The British Empire undoubtedly gave America a head start but I don't think English would've become the de-facto lingua franca of the world based on British influence alone.

The "scale and reach" of the British Empire was certainly tremendous but I don't think it's really comparable to the modern "American Empire", their nature is wholly different. I think the influence of American culture on the common people of the world today is much greater than the British Empire ever accomplished. There's a Mc Donalds on the Champs-Élysées, a KFC on the Red Square and Starbucks in China.

>The reason English has become the lingua franca for businesses around the world is evidently because the USA became a superpower and managed to gain a huge worldwide influence in both economy, military and cultural sectors.

That's true but it was already well on its way in the 19th century because of the British Empire.

As an American, it's fascinating to read your perspective. But you do leave the impression that you're bitter about my country's comparative economic success. Is that an inaccurate impression?

Esperanto is surprisingly easy to learn and is meant to be a secondary language. Realistically, it's probably going to be English. Some say Mandarin in 20 years, but there are a lot of practical reasons why I find that hard to believe.

There's no chance that Mandarin will ever take the position of English and that statement is solely based on the peculiarities of the written language, not even the seeming lack of interesting content produced in Mandarin (comparatively tiny Korea and Japan are much bigger cultural exporters than China and Taiwan).

Chinese is only really spoken in any capacity as a second language in countries that used to use Chinese or Chinese characters for the written language (Korea, Japan, Vietnam) and countries where ethnic Chinese people currently live. If you did not grow up looking at Chinese characters and speaking a language that has a big lexical overlap with Chinese (Korean, Japanese, Vietnamese) then Chinese is extremely difficult to learn. The writing system is incredibly impractical since it requires learning thousands of symbols to read any normal piece of text and the lack of spaces makes parsing many basic sentencse very challenging until the learner is at a very high level (must know all characters in sentence, must know most words in sentence, must know general grammatical flow in sentence).

The coming of electronic dictionaries and Pinyin IMEs has made it much easier to learn Chinese than it used to be, but it is still ridiculously challenging. Many learners living in China simply skip learning how to read, which I don't think happens with many other languages. Chinese people are often functionally illiterate and even educated Chinese have trouble reading certain things (names, especially). There's a very well-known essay in the Sinologist community about all of these challenges: http://pinyin.info/readings/texts/moser.html

> Some say Mandarin in 20 years, but there are a lot of practical reasons why I find that hard to believe

English has been taught in European schools for decades and most European citizens still don't use it much or know it that well (the biggest reason is that it's not the official language of the countries).

Virtually nobody teaches Chinese in European schools now, so I find it very hard to believe that Mandarin will be anywhere close to the language of the internet in 20 years. It's really only the Asian countries (except India) where a larger percentage of their populations can speak or understand Mandarin, but that's always been the case anyway, so I don't think that's going to change much.

Except singapore and taiwan no other country has mandarin as its national language. Infact more countries have tamil and hindi (ex british colonies) than mandarin. Food for thought

I grew up with Esperanto as one of my birth languages - yet I'm writing this reply in English. So yes, I agree - English is the more practical contemporary language as you can communicate with more people with.

I'd be really surprised if Mandarin reached such status not only in 20 years, but in the next 100 years. While it certainly gets more popular due to China's growth as an economic power, its far behind English when it comes to number of non-native speakers.

On top of that, learning Mandarin is very difficult. It's a tonal language - as someone who has absolutely no ear for music, I find it challenging to understand the differences between different tones. The alphabet is another obstacle - even in China children at school learn words written in pinyin (system of writing Chinese characters in Latin alphabet) first. Even with China becoming an economic world leader, I don't see Mandarin becoming mainstream (having said that, I learn it and I enjoy it, although it takes me much, much longer than to learn any other European language)

> I believe that the only country in the EU who has English as an official language will be Ireland

And Malta. Though Malta is minuscule.

Historically, French was the court language of Europe and therefore the the language of diplomacy. (US Passports include English, Spanish, and French; I think they used to only have English and French.) Of course, that's not going to happen even though Brussels is primarily Francophone.

I didn't propose French because both France and Germany are often accused to have too much control and influence in the EU so I think there would be a significant push back if either of them tried to impose their language to the rest of the union.

Furthermore Spanish is probably easier than French for any kind of objective metric you could throw at them. In particular Castilian is easier to spell and has a simpler phonology. On top of that Spanish is one of the most widely spoken languages in the world. I think it would be a strong contender for "lingua franca", although that might be my romance language bias showing.

Judging by the almost physical pain I experience when listening to non-native speakers trying to express themselves in my language, I think making English the lingua franca is a fitting revenge on all those imperialists.

They may have raped our natural resources and women, but they can’t protect their irregular verbs and diphthongs.

> Judging by the almost physical pain I experience when listening to non-native speakers trying to express themselves in my language,

Just by curiosity, what is your mother tongue? In my case, hearing non-natives trying to speak my language fills me with a sudden outburst of joy! I just want to immediately become friends and help the person, and introduce them to my friends and family! (Maybe because it is quite rare to see foreigners learning catalan.)

> I’m German. We have a 70-years tradition, well deserved, of being embarrassed by our nationality and its outward signs.

I never understood the rationale of being proud, or embarrassed, about the actions of other persons. Guilt/pride by association just makes no sense to me; it seems like a complete non-sequitur.

Sure you're German, but neither you nor your contemporaries participated in the horrors before or during WW2.

Nationalism is, for better or worse, part of human nature. It’d be mighty hypocritical for humanity at large if people could just pick and choose from their history.

For Germany specifically, denying responsibility for the holocaust just wouldn’t work if you also wanted to field a team at the soccer World Cup. Given that choice, even the right wing at least goes through the motions.

Two well-worn formula encapsulating the idea are “not guilt, but responsibility”, and “it happened once, therefore it can happen again”.

But the “dirty secret” of German identity is that we are living a rather good life within this cocoon of professed guilt. It’s not just about accepting history, but also (I hope) the work for peace and European integration that the country has done over the decades. But the result is that I very rarely actually experience negative consequences of my nationality. I’m far more liable to get encouraging feedback like yours. In a way, it’s the most successful humblebrag of history.

Politically, Germany had it far easier to escape the various calls to arms during the Cold War and beyond, or to keep its military smaller than official NATO agreements require. If you get really lucky, your country might even be forced to give up its fearsome currency, accidentally handing you the keys to a continent’s economy in the process.

Which is why I have no earthly idea why countries such as Turkey and Japan resist acceptance of their historical atrocities. Seriously: stop fighting it, and enjoy the adulation of a repentant sinner.

> Nationalism is, for better or worse, part of human nature. It’d be mighty hypocritical for humanity at large if people could just pick and choose from their history.

It might be hypocritical, but nationalism not only regularly involves that, but goes beyond it to often involve picking and choosing from a combination of actual history and fictional history, and that seems to be at least as much part of human nature and nationalism itself.

> Nationalism is, for better or worse, part of human nature.

Nonsense. Humans (and other mammals) are born with genetic programming to recognize faces, drink milk and cry when hungry. Nationalism is entirely a cultural artifact.

I disagree. Caring for "your" people is very much part of human nature. Where you draw the line can change (examples are blood, race, hometown, city, region, nation, continent, family etc), but "us vs them" it is an overwhelmingly common pattern.

As JoeDaDude pointed out tribalism is not nationalism. Most mammals are social creatures and, unsurprisingly, tend to bond with their peers.

The idea that someone that you've never seen or met is "your" people because it has the same letters on a passport is a cultural artifact.

> "us vs them" it is an overwhelmingly common pattern

The fact that is common in various cultures does not imply that it is "human nature", but the opposite.

Prof. Sapolsky research on baboons "group culture" is a good example. There are various lectures from him on youtube.

I agree in principle as "nations" are a fairly recent development. Perhaps a better term would be "tribalism" as people like to belong and/or identify with a group - call it a tribe for lack of a better term.

Ideally, people would consider supra-national entities (such as the EU) as their tribe, but I wonder how far that will go. Even in a large country such as the USA, it is hard for everyone to identify as belonging to the same USA tribe, what with divisions of North-South, religion, race, etc.

I’m German. We have a 70-years tradition, well deserved, of being embarrassed by our nationality and its outward signs.

Edit: I should add that, in my initial comment, the “raping” of women and natural resources was not meant to apply to Germany. It was more of a literary device playing on the sweeping accusation of imperialism.

> I’m German. We have a 70-years tradition, well deserved, of being embarrassed by our nationality

But do you feel pain when non-natives try to speak german because they speak it incorrectly, or for some other reason?

I love german! I started learning it a few weeks ago, by reading the original Hilbert-Courant. I'll try not to slaughter your language by speaking it yet, though :)

Sure, Spain has never did anything imperialistic!

As someone from Australia, I actually take the opposite impression from that. These startups are faced to deal with being international by day one, making scaling out to more and more countries easier.

For US startups, on the other hand, the US is a large enough market that they might get ‘stuck’ there and need to build up more ‘inertia’ to scale out internationally

Getting stuck in a market that is 23% of world GDP is not so bad.

The fact of the massive advantage that the US gives to companies and researchers in the form of essentially a single regulatory zone and, importantly, a single language, is reflected in the fact that almost all of the technology you are using was designed in the US. Google, Apple, Facebook, Microsoft, Amazon, etc. for companies, C, Python, Lisp, etc. just in programming languages...

You only have to write documentation once for more potential customers or research collaborators to be able to read it than if you wrote in both French and German.

This is well studied. Note that in terms of developed countries the US is the largest by far.

Pst, Python is Dutch. Not retracting from your point though, it was built in English and found adoption in SV.

Wow, you’re right! Thanks for that, learned somehow today.

Having to be international by day one puts a high barrier to entry, requiring lots of effort in a growth phase that's resource constrained, and kills potential ideas before they can get traction, as it's more expensive (and slower) for them to do so.

On the other hand, California by itself is a large enough market so you can get a profitable (because of sufficient scale) proof of product-market fit and tackle internationalization with much, much more resources. Which seems the proper way to go; good internationalization doesn't really scale.

Its certainly an advantage the European startups don't seem to take up on. For example Duolingo is headquartered in the US.

I would guess that the main reason Duolingo is based in the U.S. is that it originated at a U.S. university (Carnegie Mellon).

Duolingo is purely an example - but you could ask why the research occurred in an American University rather than an EU one. Google Translate is another.

In theory there should be many more people willing to pay to learn another language / translate in Europe. Yet the US is far ahead.

Definitely agree re English being adopted as the lingua franca (that really sounds weird haha). Been saying this for years however there are too many fragile egos in European politics that feel that adopting English is an attack on their identity. There was even talk of removing English from EU legal texts altogether after brexit

Depends, in regulated industries you might have to adapt to product state by state in the US much more that you would have in the EU, where a lot or regulations are now more uniform.

> lingua franca

Hmm, I seem to recall we already had such a language in Europe since middle ages, it started with L and ended with n. It's still dominant in medicine. It could be funny to use it in tech EU-wide as well with "proper German pronunciation" ;-)

Does medicine still have universities teaching in Latin?

I haven't come across significant latin usage beyond naming conventions in medicine.

In Italy still teach it in school, and it's not the first time I've heard of mayors (beside linguistics) that have some compulsory Latin thrown in. But that depends a lot on the university.

National egos will probably prevent that. I simply can't imagine French people speaking anything but french in their home country. Anytime I visit the place (which is almost every weekend), I see how much they don't like speaking english, although their level is OK and they are perfectly understandable. Italy might be the same, Spain too.

Northern europe has much stronger usage of english and level is much higher (I recall old grannies responding in perfect english around Oslo, and that was almost 20 years ago).

I think us French are generally pretty terrible for this, there's definitely a strong anti-English bias (although oddly splicing random English words in a terrible French accent in everyday discussions is seen as fashionable by what seems to be a growing portion of the population).

On the other hand I've never had any issue speaking English in Germany or Portugal for instance, to the point where I kind of became lazy and expected people there to actually understand English and being surprised a couple of times to find somebody who didn't.

But you're right, I can certainly imagine a significant portion of the French people and institutions fighting back hard if the EU ever decides to move towards English as official language.

It's true for the northern countries. I remember being in Stockholm once and everybody was speaking a decent English, from little kids to grannies. They said it's because they don't have dubbing on TV and learn it faster and better that way. :)

> and even though EU has free movement of labour, the language and culture barrier is strong, and internationals have a harder time to get hired, unless they speak the local language.

Certainly not in all fields. In all the math labs that I know (from France, UK, Italy) more than half of the researchers are from other countries.

Of course, if you want to teach, it is much easier to get hired if you already speak the language.

(edit: format of the quote)

Is that really true? Everyone in science speaks English and DeepMind, based in the UK, is in fact one of the world's top AI research hubs. Their papers are consistently excellent and of course DeepMind's workforce comes from around the world.

That's the European AI hub, right there.

The issue here isn't the language barrier. The issue is one that goes largely unstated in the ELLIS letter (it's a call for action, not really a plan). Which is this ... remind me what's wrong with working with the Americans again?

DeepMind was founded by a Brit - the guy who once worked on Theme Park and other much loved British games. He built up an absolutely top class research team out of nowhere, based purely on VC funding, then sold to Google not only because of money but because Google also built a world class AI research team out in California, and it made sense to join forces so they could work together on things like TPUs.

And why shouldn't the European AI hub have joined forces with the US AI hub? What benefit could there be from preserving institutional and funding barriers between Google and DeepMind? The letter sort of dances around the edges of this question, but ultimately the answer is purely politics.

It says things like

"This weakens Europe"


"we want the best basic research to be performed in Europe, to enable Europe to shape how machine learning and modern AI change the world"


"many European companies whose future business crucially depends on AI are not perceived as competitive"

These are all political goals, worse, they are not logical thinking at all - not a good look for a bunch of scientists.

AI research is entirely open. Having a new academic research hub in the EU-that-isn't-the-UK won't change the access those European countries have to the underlying research papers or techniques. They can use AI just the same regardless of where the research is being done. Maybe China would try to close up their researchers, but America and the UK certainly won't.

So if European companies are "perceived as uncompetitive" or are not "shaping the world" due to their poor use of AI (this is a dubious theory to begin with) then the solution is for more European companies to download TensorFlow and Torch and get cracking. It isn't for money to be poured into academics who will take it, write a few papers and then go straight into industry anyway.

> remind me what's wrong with working with the Americans again?

Make that American companies and the problem, from an academic perspective, is that they recruit away students and researchers from academia. That works for a few years, but longer term, the argument goes, it depletes academia to the point that new talent can no longer be trained... by academia.

At that point, the companies wishing to keep recruiting will create their own training facilities and/or step up collaboration with academic institutions. That's where the American part comes in: if the existing talent has all gone to the US and those making the decisions have been there all along, they are likely to focus their training efforts on the US.

On top of that, you have the well-publicized problems with data collection by foreign entities and local tax avoidance, which are indeed political.

You make it sound like a zero sum game. If academics are being hired into industry, that frees up grant money to be spent on other areas where they aren't recruiting so aggressively. It doesn't actually 'deplete' academia except in the short run, unless the supply of students who want to do research becomes fully tapped out.

Also, people who are trained already can always go into academia or return to Europe. These movements aren't permanent by nature.

> If academics are being hired into industry, that frees up grant money to be spent on other areas

That's great for those other areas, but the open letter in question is about machine learning.

> It doesn't actually 'deplete' academia except in the short run, unless the supply of students who want to do research becomes fully tapped out.

... or the very limited supply of established researchers who can train them is tapped out. Also, "fully tapped out" is a squishy notion; you can always lower your recruitment standards, with obvious consequences.

But there is no need to get into hypotheticals here:


> people who are trained already can always go into academia or return to Europe

But why would they want to? They chose to move away for a reason, so something would have to change for them to change their mind. What would that be?

The authors of the open letter seem to think that a European AI hub could be that thing.

Academics consider anyone who has worked in industry “tainted”, such a person would never make it to tenure, even if they wanted to

Tenure is about the only carrot academia has to dangle

I don't see that in computer science, or in my subarea of research. Major industry labs are highly respected contributors to science, and this has a long tradition with Xerox PARC and Bell labs, and now with Microsoft Research, Google Research and DeepMind (to name few among many).

That's not true. I knew a lot of professors at my undergrad, which is a well-known university, that worked in industrial research positions after their PhD before returning to academia.

These are all political goals, worse, they are not logical thinking at all - not a good look for a bunch of scientists

Indeed, it’s almost on the level of Make Europe Great Again

Is this not preferable to one country "buying all the best players" and exerting their will on others?

I'm not sure that's the case because the United States is also filled with people that speak different languages.

A smaller labor supply implies that the supply and demand curves cross higher. Pay should be higher then.

Though the segregated markets could also imply smaller demand.

From an outsider, I can tell you what. It's not the langauges. It's the culture, it's the work/life balance. It's things like working reasonable hours, spending time with family & friends. Having long holidays, take time off from work to care for a newborn. It's things like GDPR.

America is brutal, America is a gladiator pit, winner take all. Money above all things, the religion is capitalism. Make money or die trying and roll over anyone in the way.

No, language doesn't matter: EU has lower wages because it tax (and redistribute) more, and I like it this way.

Our big problem in Europe is not basic research, it's systems research. We have a European platform for Big Data and AI - Hops. It's open-source. It has nice properties, but it is harder to bootstrap an ecosystem outside the valley:

(1) HDFS compatible FS with distributed metadata

(2) GPU support in HopsYARN

(3) UI to develop Keras/TensorFlow apps in jupyter, deploy on tensorflow serving. Spark support, too.

However, getting it 'out there' is still a challenge from Europe. Here is a talk on the platform at CERN last week:


More talks here:


> Our big problem in Europe is not basic research, it's systems research.

Hmm maybe we mean different things by 'systems research', but I'm in systems research and we're relatively strong in this area in Europe! Notable systems groups at Cambridge, Kent, Oxford, Glasgow, Manchester, EPFL, ETH, Linz, lots of smaller ones. Major influential systems projects started in Europe like Graal. Conferences which attract Americans to fly over like ECOOP and Curry On.

I agree - i would not include ECOOP and Curry On in systems research. They are prog language conferences. Systems research are the USENIX confs (NSDI, ATC, Fast), SOCC, SOSP, OSDI, Eurosys. Typically 10-30% of papers are european (often with American co-authors).

I think things have gotten worse, if anything, since Peter Druschel's article on the state of systems research in Europe a decade ago: https://hal.inria.fr/inria-00444781/document

In any case, the largest systems software company in Europe Nokia, went belly up due to a crap OS. Now we're left with SAP, who are just trying to protect last generations' software (a DBMS and its stack). Where are Europe's next big players? And why do they sell out so quick (MySQL, Elastic, JBoss, etc)?

That's absolutely true but if you read between the lines it's possible that's not what he means. Lots of people in Europe use the word "Europe" when they actually mean "EU". This is such a common verbal tic it largely passes unnoticed.

The systems groups you name are mostly in the UK (soon to be ex-EU) or Switzerland (not EU, doesn't want to join). Only Linz is in an EU country with no signs of that changing.

Substituting Europe for EU and the original statement might well be correct from a purely geopolitical perspective (obviously tons of research is done in the USA by people from Europe so we're only discussing geography here and not ethnicity).

Except ETH and EPFL, other names you mention do not seem to be very competitive in CS systems research with top departments in the US and even some in Asia. I am not in systems so please let us know if the ranking based on conference publications skews American or it is a fair reflection of reality.


Three in the top ten when I select programming languages which is my sub-field. That's really not bad for Europe. I said it was relatively good. Compare it to something like AI and it's good.

And this website does seem to exclude the major European conferences like ECOOP... which obviously doesn't help European institutions.

You noted about programming languages I see. Several names you mentioned are in the top 10 or 20 in the subfield.

The headline seems misleading. They're not planning one "huge hub" in one location. Instead, they're planning a bunch of mid-sized institutes spread across different countries, each with "hundreds of computer engineers, mathematicians and other scientists".

That approach is understandable politically, but it seems much less likely to have the intended effect. If the goal is to kick-start a knowledge cluster in Europe that can compete with the Bay Area, then it seems much more likely to succeed if the funding and talent is concentrated in a single site. "Hundreds of computer engineers" is equivalent to a mid-size remote office of a major tech company.

That's just an observation, though. I realize that the choice probably isn't this plan vs. one mega-site. More realistically it's this plan vs. nothing at all, or vs. competing piecemeal efforts by various EU member states.

I think it would go standard European way - build an institution by a top-down directive reflecting political will, shake each others hands, then a few years later cry that you have all this outdated infrastructure but no inventive people working for you nor any bleeding-edge results, as all relevant people are snatched up and overpaid by industry giants.

I think it's worth experimenting with a more diverse and distributed model because the 'one megahub' structure also has disadvantages, low spread beyond the borders of the tech industry being one example.

Especially countries like Germany with diverse industrial production spread out across the country might benefit more from having technology clusters spread out as well.

Even Silicon Valley companies like Google have at least partially taken on this approach by opening labs in Toronto or Paris, Amazon plans to create about a hundred jobs around Stuttgart and Thübingen.

I think a hundred researchers/engineers is well beyond a midsize company remote office, because the caliber of people and mission will be very different. DeepMind is less than a thousand people, and was probably on the order of a few hundred when it became famous.

For comparison, the Fraunhofer Society has 72 institutes in Germany, all focused on applied research. Collectively, these institutes employ ~25,000 scientists and engineers[1], which gives us an average of ~350 scientists and engineers per office. This seems very comparable to what's being proposed, both in mission and in scale of individual institute.

This seems to be an effective model for applied R&D, but I haven't seen anyone argue that a Fraunhofer Institute is sufficient to be a seed for a major tech cluster. The scale just isn't big enough.

[1] https://www.fraunhofer.de/en/about-fraunhofer/profile-struct...

the scientists say the proposed Ellis institute is essential to avoid brain drain to big tech firms

They don’t say tho’ the mechanism by which it will accomplish this. What’s in it for the individual researchers who choose this path over an industry job? ITER only exists because there is no well-funded private-sector fusion research. If Musk or Bezos or Gates entered that game it and CERN would evaporate overnight.

I’m no AI guru but I’m a pretty decent infra engineer who can wrangle storage and compute at scale, and they’ll need people like that too, why would someone like me choose this path? What’s the offer?

Genuinely curious.

Well, a nice office in a "culturally rich city with dozen of events, museums and world-class cuisine"?! Like it's done for all other industries. Some nice polished marketing and the fear of missing out if you don't go to the place where all the "exciting things happen".

Well, a nice office in a "culturally rich city with dozen of events, museums and world-class cuisine"?!

I have been thinking about this today, and I'd take a cut in pay of say 25% in exchange for a private office in which to do interesting work, and absolute job security followed by a final salary pension. The way it used to be.

But that actual deal would be a pay cut to 25%, crammed into a office too dingy for Civil Servants, at risk of being TUPE'd at the drop of a hat. No thanks.

Question for people knowledgeable on the topic: are there many examples of planned efforts like this actually working, where some group is able to attract top talent in some industry and keep it for the long term? The article mentions CERN; are there others?

I would guess that most concentrations of talent are an unpredictable side-effect of some environmental factor.

Answering in a wider context:

Silicon Valley: https://youtu.be/ZTC_RxWN_xo

It wasn't planned as such, but the outcome is a result of unlimited debt-based US government spending in WWII to win the electronics war. Only after creating very fertile soil did private enterprise enter the picture.

Also, anything war related, ever, including rockets, airplanes, anything that shoots and explodes, submarines, the Internet,... Part of it is that the public is perfectly willing to have "central planning" and pretty much unlimited tax-based spending for military purposes, but try to do get them to accept the same for civilian purposes...

You may also check out the economic history of Japan and South Korea, but also how European countries like the UK rose to power. LOTS of government planning.

Then there is the entire exploration of the world, done mostly by governments, and even when it was private entities it was with very heavy government support and involvement.

Did you know that Gregor Mendel, the monk who did the experiment with the peas that is seen as the basis for genetics, was not an underemployed monk working on his own, but part of a very organized business-government-church effort to support local industry? I think it was about how to get better wool producing sheep: https://youtu.be/D8m-ZEr9qV8?t=276 (MIT Prof. Eric S. Lander)

The "market" is not magic that does everything. It only works well under specific circumstances - and those are exactly the scenarios the books will pick. Read (economist) authors like Ha-Joon Chang (https://en.wikipedia.org/wiki/Ha-Joon_Chang).

>You may also check out the economic history of Japan and South Korea, but also how European countries like the UK rose to power. LOTS of government planning.

Yes, although in the case of Japan (MITI and the like), there are counterexamples like the Fifth Generation project. In addition, there were economists like Lester Thurow in the US arguing strongly at the time that the US had to follow Japan's lead in industrial planning to create cosortia like SEMATECH, which largely didn't pan out.

>Fifth Generation project

Does anyone know what eventually happened to it? I had read about it a bit in the media at the time, it seemed like an ambitious project. I remember they were going to use Prolog instead of Lisp (for whatever reason). Did any concrete benefits come out of the research, for example, that are widely applied in (any) industry? Interesting to know.

Based on what I remember, the Wikipedia article seems to be a pretty good writeup.[1] The bottom line seems to be that these sorts of efforts are usually not complete wastes in the sense that they bring a lot of smart people into the field. But it's hard to point to much that came directly out of this project that had much of an impact.

[1] https://en.wikipedia.org/wiki/Fifth_generation_computer

Thanks, will check it out.

I'm familiar with the European Molecular Biology Lab [1] (worked there), which is one of the best research centres in cell and molecular biology as well as bioinformatics in the world. The only real prerequisite for working there is excellence. It seems like this new AI proposal is modeled after the EMBL model, with one headquarter and several other locations around Europe. At least in the case of the EMBL, it worked incredibly well and created a great joint European atmosphere which really motivated everyone to do their best.

Obviously, there is currently much more money in AI in the industry, so I guess it will be hard for such a research centre to compete with industry. The wage differential between biotech industry and EMBL is much less than that of a hypothetical EAIL and the software industry.

[1] https://www.embl.org/

I don't disagree with your sentiment, but I think you've missed the question that GP was asking.

> are there many examples of planned efforts like this actually working, where some group is able to attract top talent in some industry and keep it for the long term?

If we're talking about EMBL in Heidelberg, we can say with confidence that (along with DKFZ) it's a huge drawcard for the University. But the GP asked if it had attracted talent in some industry and kept it for the long term. Can we really say that the Rhein-Neckar region has really attracted big private sector biotech investment, on par with a place like Basel? I don't think so. By a pure coincidence of geography, companies like Roche have a presence in the general area (Mannheim I believe), but I don't think you can argue that any deliberate actions by EMBL or DKFZ have led to deep private sector talent pools.

I actually think you could turn this argument around pretty easily and show that even enormous investments in institutions like EMBL and DKFZ don't actually create the private sector spinoffs that central planners would imagine.

Like the GP, I suspect that there is a combination of environmental factors at work, that we can't fully identify at present.

Maybe you are right and I misunderstood his question. I was under the impression that he asked whether a European governmental facility could aggregate talent in some place and came up with an example where this worked and why I believe an EAIL could work as well. I'm not sure that either of us understood his emphasis on industry.

Regarding your point: Heidelberg and Cambridge have both profited to some extent, but I'm convinced that a place like EMBL has profited Europe as a whole much more, keeping excellent talent on the continent and attracting excellent talent to the continent. Bork, Ellenberg and co. would have otherwise been long gone to the US and with them a lot of excellent scientists.

Regarding cluster formation: I agree, EMBL and DKFZ alone are not sufficient for creating the critical mass that you see in Boston and SF. You need a couple of big companies as well as the academic infrastructure to get these clusters up and running. I guess only Munich has the critical mass in certain sectors in Germany.

I find the comparison with CERN (made by the Guardian article; CERN is not mentioned in the open letter) misleading and somewhat troubling.

As of 2016 (last year for which there are figures; new ones should be out in a couple of weeks) CERN had 82 research physicists on staff, out of a total headcount of 2531 [1].

The vast majority of researchers working at CERN are associates (982), users and visitors from other institutions, mainly universities in member countries. CERN's primary role is to provide shared facilities (accelerators and detectors) which are too expensive for individual members, not to employ researchers.

The equivalent in machine learning would be shared computing, data collection & labeling resources capable of competing with what Google, Facebook etc have to offer. A state-led initiative on shared computing resources might be more cost-effective than buying time on Google Cloud, but the main thing could well turn out to be pooling data from national databases. Yay?

[1] https://cds.cern.ch/collection/CERN%20Annual%20Personnel%20S...

I don't even think CERN, Higgs boson included, is worth the man hours or cash invested...

HBP doesn't seem worth it. ITER is a huge failure that is scheduled to receive billions more euros for more than 25 more years. I don't know... Europe seems to fall victim to these sort of things very readily.

While everyone in these comments seems to try to defend these great European projects on their merit alone, I'll try a different approach.

In my opinion, these projects are one of the last great human efforts to inspire generations with things that are not just about making more profit. But it's not about this project or that. It's about far more.

You say this money should go to hungry people instead. Let's say it's not just starving children in Africa, but maybe even a project to feed the homeless or help under-privileged children get an education...even in that case, do you actually think that if all that money was freed up, this would actually happen?

Of course it wouldn't. It would go to line even more pockets of some faceless international organisation / corporation / hedge fund / military investment buddies of some regime.

There is no way in hell that any poor or under-privileged person would benefit from the abandonment of these projects. No chance whatsoever. What little money there is, is all concentrated in the hands of a few people who have absolutely no other mission in life other than accumulating more money and power.

They have made most science happen for private sector profit-reaping already. This is some of the last stuff done for all mankind. We'd be making a huge mistake abandoning it.

Again, the people in charge will not help the poor or homeless. If you wanted to do that, you should tax multinational corporations and get rid of tax havens.

There is a kind of money hidden away in tax loopholes all over the globe that would make the LHC, HBP, heck, the ISS look like chump change.

Look there first.

Are you defining "worth it" as whether there is a profit to be made at some point? How much would you have paid for the higgs? Are some pieces of science best left in the dark because they cost too much?

YES. There is no limit to the cost of modern science, but resources are still scarce in the real world and we cannot afford to spend unlimited funds on projects that will not pay off in human flourishing in the near future. To answer your first question that is the profit to be made and it absolutely is necessary to consider when directing tax dollars which could otherwise be spent feeding starving people or curing dying people. I would have paid $80 for the Higgs.

While you are correct to say that scarce resources means we can’t do everything, pure/fundamental research like the LHC tends to have extremely high payoffs several decades after it is completed. Nobody knows what exactly, if we did we could focus on that stuff in particular, but its one of the best things we can do to grow the economy.

(Also, if everyone in the EU paid $80 for the Higgs, that’s about three LHCs depending on where in its history you take the exchange rate).

Actually, good basic research has a much quicker payoff scale, like 10-20 years. Transistors, fiber optics, lasers, paid off almost immediately- and were invented ON PURPOSE by the way, in order to facilitate better communication for AT&T. And I would pay $80 because I'm personally interested, but I'm the 1%, and I don't think you should force people to pay for it, therefore you should revise your estimate to .03 LHCs.

You’re taking a bunch of things deliberately invented to solve an existing problem as an example of fundamental research paying off quickly?

Counterexamples: one of the early prime number researchers was proud that his work had no use at all, and now it’s the foundation of a major class of encryption. That was a century or two.

Or, Maxwell published “A Treatise on Electricity and Magnetism” in 1873, and it took another 30 years to become voice radio.

Of course, as the LHC only discovered the Higgs in 2012, even your 10 year lowball would be four years in the future.

I think you have no idea how (fundamental) research works...

"Groundbreaking science is too expensive, we are only capable of investing in one thing at a time!!!"

Like it on not, science and research is what powers humanity flourishing, if you want to get your pitchfork out about misused resources, go complain about corporate tax cuts or money invested in pointless ad-tech, not science.

> Like it on not, science and research is what powers humanity flourishing

Not all science and research though. That's the whole point of my argument - we should direct funds towards projects that actually have tangible benefit to humanity in the future. I will eat my shoe if any application of neutrinos comes about in the next 50 years. Transistors? Great - those weren't invented by accident, and John Bardeen certainly deserves his two Nobel Prizes.

Remote monitoring nuclear reactors for signs of weapons development: http://iopscience.iop.org/article/10.1088/1742-6596/593/1/01...

Communication through any obstacles, including (in principle) the entire diameter of the planet: https://arxiv.org/abs/1203.2847

Considering we have already have cheap things to detect nuclear proliferation, such as seismic detectors that have proven effective w/ North Korea and how expensive a neutrino detection system would be to build, I don't see any marginal benefit to humanity there. Considering the fact that you would need to build a beamline for any transmission of a neutrino message and a 5 ton detector to receive it (unlikely to be reduced since it's so hard to stop neutrinos) I'll stick with my > 50 year timeline for neutrino message passing being used in any practical sense. Maybe they could put it in a nuclear sub, but the design process would take 50 years, thanks to similar incentives built into the military industrial complex.

Seismic can only detect an explosion. Neutrino can detect the construction. If the threat comes from someone who has reactors and already knows how to make bombs, seismic is too late.

A 5 ton detector isn’t particularly implausible in a whole range of scenarios.

Eat up. :)

Still waiting...

Someone literally just provided you with an explanation of a possible application, you can't just dismiss their comment because you don't like it.

Isn't there already an application of neutrinos: to detect nuclear material (so weapons inspection and so on)?

And how will we know what science is going to be useful and "profitable" until we do it?

I'm not even going to dive into the huge discussion that is "why profit driven science is an objectively terrible idea".

ITER started as an equal partnership between EU, US, USSR (yep, not Russia) and Japan. It is insanely expensive, but I wouldn't consider it a Europe-specific madness.

Not worth the WWW? (Amongst many other things)

If you had the power to view into parallel universes and told me TBL would not have invented the integration of hypertext and the TCP system without the support of CERN I would begrudgingly give CERN funding until 1995.

But that's the thing: before 1995, you couldn't know that you had to keep funding the CERN to get the WWW. So you'd have cut it off before, and now you'd be arguing that you made the right call, because we wouldn't know that the WWW would have been a thing.

But it is very unlikely that another parallel universe doesn't have TBL or someone else inventing WWW in another setting. I would argue that the invention of the web was dependent on TBL, not CERN.

I'm not sure we would have computers at all if we followed your approach.

I think you're trying to paint me as someone who is anti-science, I'm the opposite. I have extreme confidence in what our scientists can do, especially when you stop giving them easy money. Note that CERN also employs thousands of very smart particle physics PHDs who's median contribution to humanity is roughly equivalent to a giant land sloth's. I want to demand much more from them, and I'm also extremely confident that they can deliver.

"I have extreme confidence in what our scientists can do, especially when you stop giving them easy money"

Yeah, as a scientist there's nothing like underfunding when it comes to doing great things! Having to reduce the team letting people go, work with obsolete/scarce equipment, and spend the majority of the time writing grant proposals to beg for some money to move forward, are great boosts indeed. And surely the great advances of science have not been associated to large labs with huge resources, often related to the deep pockets of the military...

No, I'm painting you as someone who has way too much faith in their own abilities to predict what can be achieved by pure research.

You can only demand more if you're certain that more is achievable. I don't see what evidence you have that allows you to conclude that.

>are there many examples of planned efforts like this actually working, where some group is able to attract top talent in some industry and keep it for the long term?

Many, I'm not sure, but how about Silicon Valley? Which was borne mostly out of government backing of military projects and the university system. https://en.wikipedia.org/wiki/Silicon_Valley#History_(1971_a...

OpenAI, Allen Institute for AI, MIRI, etc. are quite successful in the AI industry. Those are all in the US and pay competitive salaries though. It seems to me that Salary is a primary driver here, why stay at Cambridge for 40k gbp/yr when you could do the same research at Google Brain for $200k+ but if those two numbers aren't so out of wack I think it becomes a much more difficult choice.

If you had any idea what scientists make in the EU, you'd see

1) why this is happening

2) why this effort won't change a thing

They could, you know, pay competitive wages. Ridiculous suggestion, I know, at this point that'd be a 5x and more (500% and more) raise.

I fully agree.

German media constantly complains about "IT expert shortage" yet salaries grow like 3% per year. So IMO it's either not a shortage or the German industry doesn't know how to attract good people.

To put some data points there (in EUR)

1. Germany Uni Doctorate Position: 32k, 22k after deductions

2. Germany AI / Data Scientist Position: 80k, 50k after deductions

3. Luxembourg Uni Doctorate Position: 32k, 26k after deductions

4. Luxembourg AI / Data Scientist Position: 80k, 60k after deductions

Source: Friends and family and me

The above two countries belong arguably to the economical top tier, yet they can hardly compete with a US Senior Software developer (>= 100k$?)

Now imagine how the non-top tier countries fare (i.e., 80% of the EU population)

I was an AI researcher and now work as ML/AI/Data Scientist consultant (mostly for non-European companies) and I yet have to see that European job offer of 120k€ per year, hell, make that 100k€.

I’m in the UK, intending to move out for obvious reasons (would’ve gone a year ago but relative got an Alzheimer’s diagnosis early 2017). Considering Berlin, Silicon Valley, or a PhD just about anywhere if the cost is in my range.

Once healthcare (at NHS level) and rent are accounted for, $100k in Silicon Valley isn’t as good as £37k in the UK. Germany has more expensive healthcare than the UK’s NHS, but significantly lower rent and overall is only slightly more expensive than the UK… unless you intend to eventually buy a house, in which case it’s suddenly much cheaper.

Europeans always seem to forget that well-paying white collar jobs in the US (e.g. tech) usually have good to amazing health insurance. My health insurance is fortunately in what I would consider the "amazing" category and I highly doubt I would get better care via the NHS - the cost would be the same (I pay nothing, not even copays) but I have access to some of the world's best hospitals and even onsite clinics... and this is a benefit that I receive before tax, so it's not coming out of that hypothetical $100k figure.

Furthermore as it seems most tech jobs in the UK are in London, I don't think the rents are different enough to make up for the discrepancy in pay.

From my research into working abroad in the past, it seems that only Australia and Switzerland could compare to the US developer market.

> $100k in Silicon Valley isn’t as good as £37k in the UK

I don't know about the UK but I know a number of excellent engineers in other european countries that refused offers to move to Seattle or California over healthcare, housing and personal safety.

Maybe it's just me but that's just bullshit. Sorry to say, but IF you've got a great private healthcare provider US healthcare is superior. Obamacare didn't change that at all.

How much does it cost to get zero excess (copay? Not sure the terminology in the US. Where you pay nothing other than the insurance premium) for all hospital and doctor visits, including home visits and ambulance use, and up to £8.80 excess (less if you need lots) for any prescription medicine?

And cover for your family, partner and kids.

And absolute cover for all pre-existing conditions.

And it will cover you in retirement or if your employer closes down or anything else like that.

The ones you get as a tech worker will do all 3. Cover for entire family, including pre-existing, and you can continue the cover after you quit/are fired.

> I’m in the UK, intending to move out for obvious reasons

I thought you meant pay by this, but then you include Berlin in your list of potential places... Is there other reasons I'm missing out on? Genuinely curious

Brexit. I could go on for ages about specifics, but experience says that nobody outside the UK really cares (it’s so 2016), and nobody inside the UK is changing their minds.

Haha! Oh right. I'm an immigrant in the UK myself. Curious as to why the natives would leave :)

Ah! Partly the total incompetence of the people who are in charge of making Brexit happen, partly because I don’t trust the UK government with human rights unless bound by international convention (a bind which looks like it will be a lot weaker after Brexit).

It’s unfortunate, really, there’s a lot of nice stuff in the UK too. And I know I’ll miss it.

They just put someone in the pen for 8 months for having a laser jammer on their car and flipping cameras the bird.

I'd be making my way toward the door, as well.

I think he's talking about the upcoming Brexit :)

Another datapoint for perspective, from the Czech Republic (also an EU country):

* PhD stipend: €3.3k / year (untaxed)

* median CZ salary: €13k / year, €11k after taxes

* data scientist: €30k / year, €22k after taxes (royalty! although still below a doctorate student in Germany, apparently)

€30k buys you a hell of a lot more in the Czech Republic than in Hamburg or Munich.

Or in some cases, a hell of a lot less :-)

Little known fact: people from CZ routinely travel to Germany for shopping. It's not just for the cheaper prices (electronics, drugstore), but even where the prices are comparable, for the much higher quality of products (dairy, food).

The most expensive country I've lived in was Romania, where the salaries are even lower than CZ. This trend of higher salaries => cheaper and better goods is rather counter-intuitive (though services are cheaper, to be fair).

It might be counter-intuitive but, on the contrary, high salaries and cheap goods are effects of economic power.

Contrast it with powerless people that have low income and pay relatively high prices for basic goods or rent: the stereotypical coal miner from 1800, people in prison, poor people in some western countries...

Consumer goods are about the same between CZ and DE, I think what OP meant is things like housing or healthcare, that is much cheaper in CZ.

This is because of cheap German supermarkets (Aldi, Lidl) with ruthless efficiency dictating prices for the German market.

There is also the issue that food manufacturers have been selling products with different ingredients but the same packaging in Eastern Europe compared to the ones they sell in Western Europe.

While I don't disagree with your argument, the disparity is not quite that high (talking about Germany).

A Uni Doctorate Position in Germany is 44k in the first year, ~49k from the second year (EUR before deductions) as per official TVL13 tariff. You are probably referring to the fact that it's common for PhD students to have 'half' or '3/4' positions, but afaik in CS, and especially AI, full positions are the norm.

The Industry figures seem on point.

There might also be a difference between the Länder.

Case in point: A friend and colleague of mine had a full position in Dresden (AI) and he got a little more than 30k (AFAIR before deductions). That was around 2009 I guess.

Maybe in high rent cities (Munich, ...) it's different?

There is a difference between east and west Germany, but not specifically between the Länder.

In 2009 you would have gotten (before deductions) ~34873 in the east and ~37852 in the west. In 2018 it is ~45863 in the east and ~45900 in the west. This is all assuming a full-time E13 position.

Source (German): http://oeffentlicher-dienst.info/tv-l/west/

I think this framing is a bit misleading. Salaries aren't exogenous, they track productivity. If a data scientist in Germany is making half of what he would be making in SV, it's because he's much less productive in Germany. Changing this isn't simply a matter of paying more to attract better talent, it's a much deeper issue.

Actually, as someone who's actually been an academic in Western Europe, I'd say the reverse is true. Certainly mathematical ability is better in Western Europe (on average). The top, no doubt, SV's top performers far exceed those Europe has (perhaps because a decent part of them are from Europe, and those weren't selected randomly. I mean to some extent that's fun because I actually know almost 10% of the top ones. From working with them before they went to the US).

But average among the population, even at universities, no doubt whatsoever, Europeans are better at math than USians. But for the top, the brain drain makes the playing field very unequal.

Besides, it's not just ability of people at universities itself. Money for programs is also very different. The US military supports US research ... and in Europe ... it's also the US military that supports research (but obviously not equally, and I must say, I find it hard to blame them for that). And in the US you regularly see academics pay for their own organizations outside of the direct academic context. Obviously that works better with better pay. There are government grants, but it doesn't compare to US university endowments (or however they sponsor that). Many of the people "brain drained" didn't leave until they were 10 years into their academic career, out of frustration of just not fighting on a level playing field just because their governments just don't support them.

If German universities paid the same, and stopped the brain drain, I have very little doubt they would exceed the output of SV. But with the current levels of people leaving, never going to happen. And to be fair, Germany's universities are nowhere near as bad as Southern Europe's universities, and even those are far ahead of most of Eastern Europe. That doesn't mean there isn't the occasional great person working there, there is, but most of the good ones leave.

Note, research and scientific staff is equally underpaid, if not more, in the US.

Since fundamental physics and bioinformatics have been mentioned several times in this thread, if you're curious, look into salaries at places like the National Laboratories, or even independent labs like the Allen Institute.

The market distortion in the field of AI/ML isn't as simple as poor wages in academic research. There are many strong European industries that could benefit from more AI/ML and that can definitely afford attractive salaries. Those jobs already exist around the German car makers for instance. Unfortunately, there's always more money to be made selling ads.

What are some major reasons?

Is it because of the culture of not valuing software as much as 'hard' engineering or just lower average GDP per capita? (I read someone commenting that pay for software engineers in Japan and Korea are similarly uncompetitive for the former reason.) Swiss pay is supposed to be quite competitive with American, and other Western European countries are right around their borders.

European countries and companies are less likely to go big or go home in my experience. The USA has the expression, "bet the farm" and American leaders like to use it frequently. I can't think of any European equivalent saying off the top of my head. That results in bigger bets, bigger payouts when they succeed, more capital to invest in even bigger bets, etc.

The UK case of DeepMind is a good example. A big bet by British standards - lots of money raised and spent on pure research with no business model to drive it. By British standards that's incredibly brave and/or reckless. Yet a successful gamble, sold for half a billion dollars. But, sold to an American company literally called Alpha bet.

> Is it because of the culture of not valuing software as much as 'hard' engineering

No - they are paid about equally.

Here is a table (German text, does not matter, the table is clear enough): https://www.ingenieur.de/karriere/gehalt/wie-viel-verdienen-...

"Engineers with >2 years experience by sector", from top to bottom: chemistry/pharmaceuticals, automobile, energy, IT, mech. engineering, electronics, engineering and planning offices, construction

As you can see IT actually is right in the middle.

>I read someone commenting that pay for software engineers in Japan and Korea are similarly uncompetitive for the former reason.

Pay for software engineers in China is actually more competitive than in Japan/Korea, in spite of China's GDP per capita being much lower. I suspect a key difference is the presence of many startups and large tech companies, something that Japan and Korea seem to lack.

Places like EMBL or ESO are comparatively well paid, much better than at universities. Still less than in industry, but that's the case everywhere.

Doesn't this happen to US universities as well? I bet Mercedes et al could do the same as Google if they didn't require you to move to company towns in the middle of nowhere.

I think you're thinking of Audi (Ingolstadt, Neckarsulm) or Volkswagen (Wolfsburg) middle of nowhere towns with a severe lack of estrogen. Mercedes/Daimler and Porsche are in Stuttgart, a major city (for Germany at least). BMW is in Munich, doesn't get much more major city than that.

The bigger problem I see is German company's unwillingness to pay engineers more for performance. You basically get the same salary whether you're pulling the whole team or slacking off.

Stuttgart is not a world-class city and Munich is filled with Bavarians (I.e. conservative, burgeois). Cologne is too small, Hamburg has no tech scene. Germany only really has Berlin for the kind of tech and city that I imagine the HN audience is interested in.

>Munich, doesn't get much more major city than that

San Francisco, Shenzhen, Paris, London, Berlin? Because the competition is international. Regional/National thinking is part of the problem.

Stuttgart to me is more world-class than the hordes of homeless and poor infrastructure in San Francisco. I have a 15 minute commute to work. I have a little forest 100m from my apartment. I can get to downtown in 15min as well with public transport. There are over 20k expats in the biggest expat group on Facebook, and there are 3 others.

It's lacking a bit in the international food department. But otherwise it's much more liveable than SF, London and Paris. Shenzhen I can't comment on, because I haven't been. I could imagine it's amazing, because Chinese food. Berlin I don't like personally, but not for rational reasons like the other cities.

In Munich you get the international food. However you pay in rent and commute. Still better than SF, because public transport and almost no homeless. Munich is conservative, bourgeois, clean and friendly, as if that's a bad thing.

The cities you listed are 3 times older than even the idea of an American continent. They are culturally rich and unique in their own way. They have survived several countries around them and are historically significant regardless of their size today. Cologne for example is the 4th biggest city in Germany and 2000 years old. They have traditions the rest of Germany finds odd and a strange beverage they seem to think of as beer. If that isn't enough for you culturally, than Germany really cannot compete. I don't see any reason to create some artificially big city for the sake of competing world-wide.

Also you forgot that those cities are often very close to each other. The Rhein-Neckar (Mannheim, Heidelberg, Ludwigshafen), Rhein-Main (Frankfurt, Darmstadt, Wiesbaden, Mainz etc.) or Ruhr (too many to count really) area do form hubs, which can compete worldwide and are economically MUCH stronger than Berlin ever will be.

Berlin is the only true international city of Germany, and with its active startup and IT culture - ~20 hackerspaces, countless IT MeetUps, startup events, well-known STEM research labs spread across 3 full and several applied universities and research institutes like MPI or Fraunhofer; industry research (Amazon) - the only place in Germany and one of the few places in Europe that is interesting for the HN audience.

The areas you mention are only places to go if your startup wants to tap into the (indeed very established) mechanical engineering industry, or very specific areas of research. But definitely no place for people who have a more international view on things.

> Meanwhile, some universities had been hit so hard that they > had lost an entire generation of talented young researchers.

The schadenfreude is strong with this one. Maybe next time try treating your researchers less like cattle and start paying them market value salaries.

Trouble is, it's not really a "market". EU universities are (mostly) subsidized from other people's taxes, so calling for "market salaries" is essentially calling for increased taxation. A tough sell, especially given the distant, lukewarm connection between the people paying and the people being paid.

I want to thank everyone who contributed to this thread. The variety of experiences and viewpoints provided an incredible insight to someone who's been mostly in the USA for the last 2 decades. My conclusions:

a) There's not a snowball's chance in Hell that a European AI hub compete with the US,

b) Europe remains far more fragmented than I thought it was; even more fragmented than it appeared when I was in high school. "Brexit" may actually increase Europe's (and Britain's) organization.

c) Britain and Europe have tons of good people but a lack of organization suitable for pure research, be it by governments, corporations, "hubs", or non-governmental organizations.

d) China is similarly SNAFU'd but worse in some respects (unsettled markets, lack of transparency & property rights). Luckily, there doesn't appear to be a ghost of a chance of having one of China's many languages adopted as a scientific "lingua franca". [Note to self: cross the (formerly) relevant item off my "bucket list".]

e) Perhaps restoring near-universal use of the character encoding UTF-16 or even US-ASCII is not a completely dead idea but merely one to be delayed until the adoption of English as the "lingua franca" of the Internet and Europe, salvaging much-needed bandwidth. Time will tell.

SV has always been heavily subsidized. A lot of the fundamental tech comes from government research projects. Startups have access to easy capital and there is no transparency of the 10/1 or 100/1 failures, how this is sustained and where all the money is coming from.

There is definitely govt money including from the CIA, NSA at work here. So there is no reason the EU cannot do the same, infact they are late.

The open source movement has also played a key role in most of the SV success stories in the recent past. Zuckerberg, Page and others are the celebrated 'free market' 'wealth creators' but the work of Stallman, Torvalds, Rasmus and thousands of others have undeniably added billions of dollars of value.

Our currently ideological narrative of 'wealth creators' and 'free markets' completely fails to account and value these and other critical social investments. We make the narratives we want.

All the SV money further subsidizes companies like Uber and other startups hungry for market share, who can under price and run out competition in other markets which is dumping. Its just that no one is going after them in Europe and elsewhere yet but the rules are there and can be enforced when needed.

This reminds me of the European attempt at competing with the American BRAIN Initiative [1]:

"The key difference between Europe's HBP and the U.S.'s BRAIN Initiative is that the latter does not depend on a single scientific vision. Instead many teams will compete for grants and lead innovation into different, unplanned directions. Competition is happening via the nimh's traditional peer-review process, which prevents the conflicts of interest that plagued decision making at the HBP. Peer review is not perfect—it tends to favor known scientific paradigms—and American science funding has plenty of problems of its own. But the BRAIN Initiative's more competitive and transparent decision making is far removed from the political black box in Brussels that produced the HBP."

[1] https://www.scientificamerican.com/article/why-the-human-bra...

The BRAIN initiative started after the proposal of the Human Brain Project. The Human Brain Project also followed the Blue Brain Project, which was started in 2005. Leaving criticism of the HBP aside - it would be wrong to put it as if the HBP is the "European answer" to the BRAIN initiative.

The BRAIN initiative came after the Human Brain Project so it would be more accurate to call the BRAIN initiative America's attempt to compete with the European HBP.

There was a vision - it was Henry Markham's vision for simulating a cortical column. However, it was a bad idea, and it all blew up, due to a bad problem to focus on.


The question is really if Europe can compete with the whole US ecosystem without protectionist policies (and there justified). Nowhere is the US domination more visible than in the tech sector, where Europe essentially is a tech vassal state. The European countries need to build not only an AI hubs, but also an competitive business ecosystem that attracts top talent through salaries, opportunities, challenges and benefits. I don't think there is much appetite in the European population for competing on lavish salaries and raw capitalistic incentiments (both categories in which US leads).

European leaders can just look at China and see successful protectionist policies. It makes sense to have some policies not only in US-EU relationship, but also inside EU that spreads the talent on a regional level so regions don't just get drained. This is also a big issue for East- and Central-european countries.

There's actually something fascinating about the dominance / vassal premise: it applies within the US, and it applies within Europe as well.

Silicon Valley, Boston, NY, Seattle, and a few others dominate the US on tech, with the rest of the US mostly fitting into your tech vassal state premise.

The exact same thing is true within Europe for industry and tech. Germany overwhelmingly dominates Europe on auto manufacturing for example. What kind of cars are Russia, Spain, Portugal, UK, Poland, Ukraine, Greece, etc producing? The Germans are out selling and out producing the rest of Europe to an epic degree.

How many Airbus planes - and or what share of components - are manufactured in Romania, Czech, Greece or Poland? In Europe these are industrial vassal states to France or Germany.

The headquarters are in Germany, but tiny Slovakia, for example, has huge car production facilities.

US hardly has a monopoly on AI. Canada and China are also leading countries. It's great that the EU wants to invest too, the more the better! Science doesn't hold allegiance to any country and never should.

The open letter does talk about North America in general (not only the US) and China. The Guardian summarized all that to "the US" in their headline.

This is great news. We need more tech hubs and more competition in the tech industry. I'm still confused as to why europe decided to cede the entire tech industry to the US. Do europeans love us this much? That they want us to have all money/technology?

But this doesn't go far enough. Europe has 600 million people. They should have 2 or 3 tech hubs.

Look at what china is doing. They are developing their own tech industry. That is what europe should do. Foster their own tech industry.

Where this is shown up starkly is in Self-driving startups. This list of startups [0] only has 1 or 2 in Europe out of 33.

Mercedes does most of it's self-driving research in the US.

[0]: https://angel.co/job-collections/top-self-driving-cars-start...

is it just me or does the guardian symbol in the browser tab (white G on a black disc) always look like the pirate party symbol (black flag on white background in a black circle)?

Ahem ....to compete with China.

Unless it is supported by capitalist intent, it cannot compete with us.

You have already lost. Because 'us' almost certainly does not include you.

Not sure i get you. Us=the us, and it does include me.

I don;t know about AI, but when it comes to using formal methods to nail down software development and make it reliable, the EU is already outpacing the US.

Considering the vast majority of all CS and modern software languages and practices have originated in the US in the first place and the US still, by any rational or unbiased measure, has the largest and most innovative tech companies in the world, I am not totally convinced of what you're claiming

This is an assertion without data to back it up. Care to elaborate?

INRIA, where Coq development takes place.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact