'We can't compete': why universities are losing their best AI scientists
theguardian.com
theguardian.com
HTTP/1.1 402 F_YOU. PAY_ME.
Very few things grind my gears more than academic fat cats whining about losing their slaves to those who make slaves lives better.
Schools want to retain talent? Pay it more money. It is that simple. Can't afford it? Maybe you should spend less money on salaries of professors ( who need those slaves to actually do the work ), administrators, fancy buildings, marketing, sports programs, etc.
None of the professors would take 4/5s pay cut and stay, why should those who do the work? Everything else is secondary.
To be fair, if you're not then there's little reason to start a PhD and it makes perfect sense to drop out of it and go do something else...
Btw, it's also perfectly possible to start a PhD after a few years of work in the industry. Not everyone goes on to have more studies right after graduating.
Pardon my french, but it is a stupid reason for someone who can barely has any money to support himself or herself. We have empirical evidence these grad students are not stupid because when Google or Apple or Facebook offers them $150k/year they move.
I don't think there's any better reason to do research than to be interested in your research subject.
Actually, my guess is that this is the real difference between academia and the industry: if you're motivated by money, you'll optimise for making lots of money; if you're interested in generating new knowledge, you'll optimise for that.
It's also not obvious at all it's a bubble. AI field is already much more complex than the regular CS field, which already suffers from talent shortages. AI is the future, and the demand is here to stay.
This is basicly the boom times when most companies want a website but don't have one. Once the obvious applications are built these bubbles tend to burst as matence pays far less.
More importantly, the performance of many AI implementations could be improved with the increased sophistication, so the demand for such complexity is there for businesses in competitive sectors.
Any examples of topics in other fields of CS in practical use that are more complex than recent AI advances with potential practical applications?
It's breathtakingly obvious:
1. Large numbers of people getting poached from academic programs by industry.
2. "This time it's different" is what we've said during every bubble ever.
3. Lots of money chasing a few ideas.
The only thing that's not clear about this bubble is when it's going to pop. People who didn't see the potential of the internet in 1996 missed out. People who got into it in 2000 were left holding the bag.
The industry is paying five times what universities are paying. Not Fifty percent. Not double. Five times.
Professors/administrators whining about it are pure scum unless they themselves are making one fifth what they would be making in a different university that offered them a job purely because they love this specific college doing the specific research or teaching.
What is really pissing off these professors is that with their worker bees gone they themselves would need to do the work - something they are not quite used to.
> The industry is paying five times what universities are paying. Not Fifty percent. Not double. Five times.
Nope, not missing that. You don't think this happened in 1999-2000?
I agree with you that grad students are ruthlessly exploited. I've seen that firsthand. I routinely discourage people from doing a Ph.D. for that reason. And if you're not too far along, taking a terminal master's and leaving for industry is a solid option, especially because this AI bubble won't last forever. But if, like the one student in TFA, you have just one year left, quitting grad school is not a great idea. Now you have, at best, an awkwardly long stretch of grad school on your CV with a master's degree on the end of it, and at worst, no degree at all to show for your pains. At that point, just suck it up and finish the degree.
I do. I'm yet to see people who jumped away from academic CS tracks at the time being upset about their decisions. I do know of dozens degreed people in their late thirties reporting to C students while making quarter of the salary of the same C students.
You can't do that with AI. The amount of material you need to get on top of is really something else, especially if you want to be able to understand the foundations of what you're doing- which stretch back several decades before statistical machine learning, popular as it is today.
Do you know that it was possible to get a PhD in genetic sequencing not a very long time ago from very serious schools? That's right, what today can be done by a random student or a random outsourcing shop in a third world country used to be a PhD worthy specialty.
In 5 years today's ML/AI PhDs would be in the same peculiar position as PhDs of genetic sequencing.
Are you speaking from experience? Because my own experience from my degree and Master's AI courses and lectures is completely different to what you describe. Particularly during my Master's, code was only incidental; the meat and potatoes of each and every lecture was theory.
It depends of course on the quality of the teaching in a given institution, but, for instance, you can find here the lecture notes from Oxford's Deep NLP course:
https://github.com/oxford-cs-deepnlp-2017/lectures
As you will see, it's anything but just code, or lousy such. Of course, that's Oxford, their teaching is world-class- but most AI courses I've found on the internet are similar in scope.
Really, it's pointless trying to teach AI with "just code" - because in AI you can't really code much, unless you understand the theory.
>> In 5 years today's ML/AI PhDs would be in the same peculiar position as PhDs of genetic sequencing.
I think what you're saying is that in 5 years, people knowing Tensor Flow or Torch will be a dime a dozen.
Perhaps- but someone will still need to figure out some way to acquire new knowledge.
Also, it bears repeating that AI is not just statistical machine learning and statistical machine learning itself is not just deep learning. AI is a broad field with communities focused on many different subjects. Just as an example you can find information on, easily, try probabilistic programming.
These days it's really easy to get all the information you need to teach yourself whatever you need to. If you have problems, ask the Internet for help.
And these days, if you want to work on the actual leading edge, it's likely at Google, not at a University.
The information may be online but it's not organised in any way and if you're learning on your own, first you have to figure out what information is relevant to what you want to learn. University offers focus, support and guidance that you don't have when you're on your own.
I think a lot of developers treat machine learning as just another technology they can get good at if they apply themselves to it, like blockchain, non-relational databases, VR etc. In truth, AI is nothing like that. You really need to grok some background knowledge before you can do anything useful with it.
You can sure learn to use the free-access tools available today, like TensorFlow or Keras etc, but, for instance, you're not going to just figure out how to roll your own deep learning framework just by reading a few books you found online- like you could do for, say, a web framework. Nor does it mean that, just because you learned how to use a machine learnign framework, you can discover new knowledge and invent new machine learning algorithms, or optimise existing ones.
Which means- someone else always has to be there to do the hard work for you.
I'd say, that's the reason why it's AI PhD students who are being poached, and not just random devs who declare themselves "passionate about machine learning" on their facebook pages. Because the AI PhDs already have the background the big AI companies are looking for.
I would argue: what's the tangible difference between someone who has gone through all five years and someone who has gone through "just" four of them?
Plus, i don't think applied research is less important than base research.
More so than junior developers? In my experience, "junior programmer" means you'll do ten times the work for one tenth the salary - and everyone will treat you like an idiot on top of that (because, hey, if you were smart you wouldn't be a junior, amirite?).
I do think the exploitation of grad students is a US phenomenon, btw. And the article is about universities in the UK, where I'm pretty sure the conditions are different.
Still pays more than a grad student.
Because if we're going for absolute money amounts, we might as well all give up on any sort of career in technology and shift our focus to getting some managerial position in a big corporation, somewhere- that's where the money's at.
Now it is possible that you are independently wealthy and it is irrelevant to you, but I can assure you anyone who is not independently wealthy would always trade shit job (A) which pays $K for a shit job (A) that pays $K + several thousand dollars, which is the complain of this article
So certainly not "anyone" would pick the best paying job, without any other consideration. Generally, people in academia are not in it for the money and their main motivator is not money- if it was, they would not be in academia.
I don't know where you get your ideas about academics from but I can tell they're not coming from experience with academics.
Also I'm not sure where you get your ideas about motivation. In my experience, the majority of people will choose work they find fullfiling over work that pays more but is not as interesting. Example: the typical developer prefers to be hands-on with code than become a manager, even if managers make more money than developers (in most companies).
Economists have predicted 15 of the last 10 recessions.
That's the real problem. There are only so many fundamental things you can do with deep convnets as function inducers.
There's a lot of money to be made by applying those to a thousand different things, but even that has to taper off at some point.
For instance what about learning value functions for reinforcement learning (e.g. AlphaGo)? Or natural language processing? These are definitely not vision problems, or if you believe that they are then 'functions for vision problems' is actually a pretty huge class!
The universal approximation theorem backs up my claim [0] - we can approximate arbitrary functions with neural networks. I think this theorem is overemphasised in practice: we don't generally want to approximate arbitrary functions, we _want_ to encode specific prior information into the function we approximate, as you rightly say. But that doesn't mean that we have to do so, or that we only have one idea about what functions to encode, or even how to encode them.
[0] https://en.wikipedia.org/wiki/Universal_approximation_theore...
I wouldn't say we have only one idea about functions, but I would say I haven't seen much of an active pipeline, outside maybe DeepMind, on coming up with new kinds of priors over functions that we can apply to larger-scale or more structured tasks. At some point, applied deep learning will run out of steam, and someone's going to have to go back to doing basic research.
That someone may find, as many have, that in terms of sample efficiency and transfer learning, deep neural nets are not always so great.
You say this like someone who thinks academia deserves those people and industry is "stealing" them.
When you include robots for industry (e.g. Sewbots, Warehouse robots, ...) and household (e.g. Aibo, Amazon Echo, ...) it is clear that the markets are potentially much larger than their current sizes and will grow with improvements in AI.
Recent papers in AI and Deep Learning have shown that it is possible to train robots to adapt to unpredictable environment and human interactions. This will realize many new potential applications in the physical space over the next decade.
Although the fundamentals are the same, the variations are legion and could become highly complex as the environment dictates. It takes a few years at least to become an expert in the field with experience in practical applications. Not everyone possesses a cognitive toolkit necessary to become one either. There will not be an oversupply of experts in AI and ML within the next five or even ten years.
The rest of us just kept working and earning.
Get a PHD.
That's how universities can compete -- let you work on the things that interest you. There doesn't have to be a commercial application.
What's the starting point? 0.01%? Because 100x improvement in that case is just 1%.
https://en.m.wikipedia.org/wiki/Opportunity_cost
It’s true that a PhD is valuable for your career. Also, grad school is a special, and hopefully enjoyable period of your life (and therefore valulable).
On the other hand, you are not only forfeighting the earnings that you could be making —- you are also sacrificing the difference in growth of those earnings. For example:
Option Grad School: $0, $0, $0, $0, $200, $210, $220, $230 = $860
Option Worker Bee: $100, $110, $120, $130, $140, $150, $160, $170 = $1080.
Tweak the numbers however you want. In particular, many good non-PhDs soon prove themselves and earn just as much as PhDs each year.
Not wasting your 20s being poor, eating microwave noodles and doing someone else's research is worth something.
For many people, grad school is where you form some of your best life-long relationships, in an environment where you’re surrounded by like-minded people. Many people would tolerate ramen noodles and free pizza for that trade.
The situation that professors and schools would love to continue indefinitely.
Schools and especially professors need the cheap labor and cheap ideas provided by the grad students. The age old adage "Those that can do, those that can't teach" applies. The moment grad students move from making peanuts, dealing with politics, kissing ass of their professors and doing work of their professors to making bank, dealing with politics, kissing ass of their managers and doing work for their managers academia as we know it would collapse.
Also, do you have a proposed solution to open their eyes? Or do you think the situation is hopeless?
What? I don't understand.
What stops you from going back to grad school when the bubble pops?
You can happily work for free then and still come out way ahead.
I don't understand your argument. Wouldn't someone with 2-5 years of work experience, networking, and a much higher salary leave them better off than staying in academia? Either way you're still in the field you think is "an obvious bubble."
Besides, if a bubble exists and it will pop in two or five years, staying in school is worse than getting into industry now. When the bubble pops, no one will hire any new grads any more, but existing workers may survive the chopping block, and are far more likely to reenter the field when it recovers, simply because they have work experience.
I have no sympathy for academic labs that underpay. For 99% of academic fields there is a gross oversupply of workers.
I also worked full time, where just the projects that couldn't have been done without me being there would have paid for my salary for 30 years.
who knows where that money goes, but they can definitely pay a lot more for the people who actually do things. the university i mentioned has a 12 billion $ endowment. they would actually make more if they stopped having such high turn-over.
salary caps on tech workers like the article suggests is a horrendous idea. pay caps will only create more shortages. The universities just need to pay their workers market rates instead of paying near slave wages and complaining about turn over.
edit: mathattack comments about job security is incorrect. many of the positions are grant funded, you can loose your job at any time, sometimes whole labs go in a blink of an eye.
It is not as simple as looking at a large total endowment and jumping to the conclusion that they should spend it on staff.
But I think you're overestimating how much these professors are paid. Universities feeling a crunch between their income and the value of their employees on the outside may have chosen to screw over their expendable grad students first, but professors aren't that much farther up the totem pole.
Part of the problem, though, is that Apple can use the scientist's skills to generate >>$1M/year in revenue. Private industry will happily pay them $250k/year, hire secretaries and managers to be crap-catchers and crap-umbrellas under and above them respectively so the scientist can be most effective, and otherwise take care of the goose that's laying golden eggs.
But the university isn't cashing in on this revenue-generating potential. They're (ostensibly) using the professors and grad students to teach undergrads, and also using them to do research that doesn't generate revenue. The existence and function of the university does provide a valuable service to society, but they're not being appropriately compensated for that service if it's worth so much on the outside - and they're squandering a lot of that societal contribution by letting it be siphoned off to the firms providing the student loans and the administration of the university.
The latter obviously leads to the former, but actually receives a great deal less funding. The former, after all, gets the deep pockets of multi-billion-dollar companies, while the latter relies on various national science agencies.
It's only gotten worse since then.
If universities are institutions of higher learning, then the people doing the teaching should have a larger say. Right now (as with many companies), the accountants are in charge.
IIRC, that was the complaint at Digital. The company was doing well when run by engineers. And then the MBAs took over... soon the good engineers left, and the company started building crap.
On the plus side, in machine learning, most of the big companies are doing a good job of publishing their cutting edge research.
The real culprit is administrative staff bloat, and facilities spending. University administrative org charts have exploded in the past few decades, and so has spending on lavish new dormitories, administrative and academic buildings, monuments, etc. While some of the buildings are funded at least in part by donations, the rest has all been fueled by the easy money of student loans, and the premium tuition paid by international students.
"Nearly every university loses money on sports. Even after private donations and ticket sales, they fill the gap by tapping students paying tuition or state taxpayers." https://www.usatoday.com/story/news/nation/2013/09/15/athlet...
http://www.cleveland.com/datacentral/index.ssf/2016/03/how_m...
This is a problem at universities of all sizes and reputation.
Someone gives you a few million, but it always comes with a laundry list of conditions about how it can be spent, and usually an audit schedule and reporting requirements. This means you need a huge staff of people who largely shuffle money around between various groups to make sure group A who have millions in capital expenditure money can pay their janitors, that money comes from group C, who launders this through an agreement with group B to provide services in exchange for use on group A's equipment with group C's support staff.
The UK government values neither research nor teaching very highly, and has additionally wrecked the economy of universities through Brexit.
Not even computer science would have gotten very far- it's another niche subject that took a long time to get any traction to the mainstream.
Anyway in the UK the public distrusts "boffins" and is likely to put serious impediments in their research, if, again, such matters were put to the vote.
AI is a remarkably poor choice for making this argument by the way, as very little research was being done into AI by academia in the last 20 years. The renaissance of neural net research was primarily driven by Google's Jeff Dean taking an interest in Geoff Hinton's research, and then showing that it could be combined with large datasets to achieve impressive new results. The EU in particular has essentially nothing to do with this: the EU's presence in the AI field is essentially via DeepMind (which will soon be ex-EU).
But I suspect if grant money were reined in it'd be primarily in the non-STEM fields. There are a lot of EU grants awarded for things that I doubt anyone would miss much:
http://www.telegraph.co.uk/news/worldnews/europe/eu/7909787/...
Researchers have unearthed a series of grants issued to schemes deemed “confidential” that have not been subject to outside scrutiny. Taxpayers’ money spent on the projects, many of which have been described as “crazy”, has increased since the onset of recession.
The schemes include £145,000 to print 736 postcards that “reflect the current problems in Europe that generate social exclusion” and £166,000 on a street circus project whose aim is to “strengthen international understanding”.
Producing the postcards in six EU countries cost nearly £200 per card.
Whilst these aren't academic grants specifically, they are indicative of the level of quality control that occurs in systems where tax money is spent without accountability to taxpayers.
Why? What other spending matter is put to the vote, like that? I don't know about the US, but in the UK at least, voters vote for a government, then the government publishes a budget. Voters don't get any finer control than that.
>> Producing the postcards in six EU countries cost nearly £200 per card.
Like you say yourself that has nothing to do with research grants, or generally funds going to science.
Teacher dissatisfaction is extremely high, and a lot of money and effort has been wasted on the "academy" scheme. It was inherently vulnerable to being scammed by the private sector and a lot of the "flagship" heads are now either sacked or jailed as a result: https://www.tes.com/news/school-news/breaking-news/charting-...
In re Brexit, the three big concerns are ability of foreign (especially non-EU) students to study without being deported or discouraged by the Home Office (they pay full price and effectively subsidise HE); Funding from European research bodies; and the free movement of academics and researchers, including to attend conferences.
Most PhD students in Computer Science are not paying the University, the University is paying the students.
The pay is low but the opportunities for growth are high if you work with good people at the right time. Think of it like a startup where instead of paying workers in equity they are paid in terms of professional development. Like a startup, this form of compensation is sometimes not always worth anything, but can also be worth far more than a salary. It is a high risk venture and should be assessed as such.
I do not want to pinpoint any specific names, but salaries for tenure-track assistant professors that started after the AI boom seem to be on par with salaries in the private sector for positions right after the PhD (100k-150k range). Also, professor salary usually includes an extra summer salary that comes from grants, I don't know whether this summer salary is included in the database [1].
Also, AI academic positions are spread throughout the US, as opposed to tech AI positions that are concentrated in the bay area. This, as well as the almost complete freedom that academics enjoy, may appeal to some talented PhDs.
There's an argument that you should be paid a salary to receive an education. However I think it would also apply to undergraduate degrees (and indeed in the UK students used to receive grants).
PhDs in the UK are often funded, the student receives a stipend to cover their living costs. But it's mostly assumed that the work you do is a form of education/training in how to do original research and that your education is the compensation.
They don't have a problem getting the money to retain talent, they just don't think paying faculty well is worth it
It's kind of strange to hear academics described as "fat cats", especially if they're being compared to, say, managers in Google or Facebook.
Also, I'm not sure where your experience comes from but in the UK PhD students are anything but slaves.
Maybe you're thinking of grad students in the US? I hear the situation is a bit different there. From my understanding, US grad students are forced to do a lot of unpaid work? This is not the case in the UK, as far as I can tell (I'm just starting a PhD).
* Lots of data
* Lots of CPU/GPU power
That's two places where Universities just can't compete with the money which companies can throw at this problem. Also, all AI companies have no interest in working with academia, just taking the staff/students, so there is no useful 2-way communication.
EDIT: I just wanted to expand one small point. The data and CPU/GPU power seem to let you research and solve fundamentally different problems. Consider a classic AI technique, like SAT. If you have lots of computers, you can solve bigger SAT problems, but there didn't seem to be an interesting research which required massive clusters of machines, or massive datasets. The big research improvements were all possible on a single reasonable desktop machine, and a selection of commonly available benchmarks.
[1] TPU: tensor processing unit. [1] https://deepmind.com/blog/alphago-zero-learning-scratch/
More importantly, AlphaGoZero is the final evolution. To reach this stage one needs to experiment a lot, and most of the experiments are failures. To do any meaningful research requires a budget of several times or dozens of times the amount of computing power, and this is only for one project.
The computation resources at universities are definitely not sufficient.
> “He was offered such a huge amount of money that he simply stopped everything and left,” said Maja Pantic, professor of affective and behavioural computing at Imperial. “It’s five times the salary I can offer. It’s unbelievable. We cannot compete.”
Glassdoor suggests that a salary for a machine learning or natural language processing engineer at Apple averages about $125k. It's not like "6 figures" means $999,999. ...And this was five times the salary that the professor could offer, implying the grad student was valued at $25,000 a year. Here are Imperial's salary guidelines [1]. Given that the student was one year prior to their PhD, it appears they were capped at £35,850, which is closer to $47k, so maybe Apple paid $250k, but $47k is still not as much as the person is obviously worth.
And then she advocates pay caps.
The reason that private industry is willing to pay these people so much is not some nefarious desire to monopolize the skills and outputs. It's because their work is able to generate value in excess of that salary.
Universities need to offer professors and grad students a job with a value - salaries, benefits, and investment in their future - that makes it worthwhile for them to work there. They've been able to mistreat altruistically minded students who don't know what they're worth, and allow shortfalls in their education so they lack the skills to function effectively in the workplace, but while that may work in some sectors AI scientists are currently so valuable that it's not using. Let's fix the universities, not place blame on the companies who are hiring the scientists out of those universities.
[1] http://www.imperial.ac.uk/media/imperial-college/administrat...
999,999+ would be “seven figure”, the university in this case can only off low- to mid- five figure salaries.
PhD students in the UK typically receive a stipend from their research council. For 2017/18 it's £14,553 [1]. You don't pay national insurance or income tax on a stipend.
I lived in London on such a stipend for four years. It's tough.
Yep. What a surprise. An article in the Guardian attacking tech companies and arguing they need to be regulated to stop them harming society by, er, paying great wages to skilled individuals.
She also throws in "getting them to pay their taxes" as if these companies don't pay taxes.
What a poor piece of journalism. The entire story can be summed up as "industry pays better wages than academia for skills in demand". Is that really news or is it agenda pushing?
I've heard Prof Pantic advocate for salary caps for PhD students leaving academia before and just can't understand it. It could mean that _all_ jobs must be capped - this seems like an overreaction to market dynamics in one corner of computer science. If not, is she advocating for salary caps for PhD students exclusively? This would be almost trivially easy to game and also downright exploitative.
In addition, we're told almost constantly how few academic jobs there are, and how we'll all have to leave academia. As soon as this starts to actually work well, suddenly it becomes a problem. I've tried to construct a strong/charitable version of this argument before and failed - does anyone have one?
On that note, are there financial institutions and tools that let you essentially take out a reverse insurance policy against your equity? They get the equity which may be worth a mint in 10 years, or maybe nothing, you get $XX/month cash?
A/B testing and ways to improve ad conversion or stickiness revolve around getting people to do things. This is different than say new math proofs that discover something.
When there is overlap and nature reveals a secret to use in society, there is high leverage and often high pay. It is not the norm though. Some may argue that it shouldn't be because it biases the purity or whatnot.
I think though zeroing in on a goal like money loses the curious branching paths that discoveries are made of. Exploration and implementation may be fundamentally oppossed. Children are fundamentally money and time sinks because they are pure learning discovery. Soldiers fundamentally can not afford to be curious. Etc...
More recently, the working environments of most universities have become much much worse; even exploitative. Tenure is much harder to get, more and more classes are taught by part time adjuncts, the administration bureaucracy becomes ever larger and more powerful. Horror stories are everywhere: The frantic scramble for jobs, the poverty wages, the professors sleeping in their cars or turning to prostitution to make ends meet, the politicised witch hunts, the grinding bureaucracy.
In Pantic's story, their student was probably hired way for a salary in the $125k-$250k range. Yes, that's high, yes it's more than a university is probably able or willing to offer. But a lot of people are not motivated only (or even primarily) by money. If you pick a top phd candidate, and make it clear to him that there is a real, viable path towards him obtaining tenure, making maybe $60-80k, having a well equipped lab, and having grad students of his own, and that he'll be a respected, high status individual in the campus hierarchy...many, many people would take that deal in a heartbeat. And that is something which was on offer 40 years ago, and is not an offer now, but could be. (Universities are still well funded; what's changed is the priorities.)
There's no shortage of people who want to be academics; the problem is that the deal being offered to them today is terrible; it's no surprise that the few who have compelling options outside academia are tempted. Boiling it down to being all about money is tempting, but I think it missed the point. If the Guardian had bothered to track the student in the story down, I suspect money would be a part of it, but only a part. And maybe not the largest part either.
Good point. Sometimes "both sides" journalism is ridiculous, but in this case it's very relevant and they didn't even have an "X couldn't be reached for comment".
As you say, the academic environment is only suitable for people who really want to be there and can put up with the terrible pay, uncertainty and lack of choice. If you're not interested in that then there are few good reasons to do a PhD, as it's essentially an extended work-sample interview for the process.
This isn't even an entirely new problem. The finance industry has been offering people much better pay for certain kinds of maths for decades.
I've been asking professors for more than 10 years at my Alma Mater (Technical University Delft, Netherlands) to start an AI bachelor + master program. There is only some machine learning and robotics hidden away at mechanical engineering. Graduates don't know most of the basics (apart from backprop or general RL, e.g. say, VC dimension). It's not a full-fledged education as it should be.
If there are not enough people trained in these matters, there will be really tough competition.
The most important factor is the environment. All researchers love an environment which encourages idea exchange. But the university (& the department) doesn't seem to see the point. The environment is not comfortable to work in to begin with, not to be mention about productivity and exchange, though, to be fair, it is a common problems across all academic institutions. So, if you want to keep them, give them a good environment please.
The problem rather is:
- Why require the scientist to invest a lot of time (thus opportunity cost) before they can get (very unlikely) a tenured position when they have the alternative to earn lots of money in industry now?
- Even if they are somewhat idealistic and would actually like to give lectures - why not offer a format such as block courses (this format is not uncommon in Germany) that fits better the industrial obligations than one lecture per week over a semester?
I believe both issues can easily be solved by the university without having to invest lots of money.
E.g. unbundling would be a great idea. Also make teaching a class less of a lifestyle choice and more of a "we made sure this person knows his/her shit". Taking classes from people who have a lot of industry experience would be dope.
Also setup research institutes like Inria of fraunhofer. a lot of good researchers become professors to do research but don't give a fuck about teaching. Which sucks for everyone involved.
Personally, I think that Professor Shanahan did the best thing. He gets a major pay raise while still retaining his scholastic position which he can fall back on in case his private sector job goes kaput.
Edit parent was calling out pay caps and then was editted.
The problem isn't that these companies are offering huge amounts of money, it's that the universities are ignoring the market signs and not raising pay. I find this odd though because at my school, all the professors who taught lucrative fields (engineering, cs, business) were paid much higher than the other ones because they wouldn't be able to retain them otherwise.
This becomes especially an issue in continental Europe, where private universities are basically unheard of.
Do you want your tax money to pay many many times more for a single professor at a university, than for even the head of state?
But they do need to pay enough to stop the poor buggers feeling like paupers in an expensive city. Much more importantly, though, is that there need to be genuine and clear opportunities for career advancement.
Somehow, we've got ourselves in a situation where some of our brightest people are paid crap, in crap conditions with crap prospects. And then are surprised when they are attracted elsewhere.
[1] "The average public college president earned just over $428,000 in 2014" http://money.cnn.com/2015/06/07/pf/college/highest-paid-publ...
The highest paid university president in Germany earns about 150'000 EUR per year, including all boni. Data from 2014.
Now remember that Professors will certainly earn less, and try to compete with Google, which will pay even an entry level developer more.
As for non private universities, well that's just what happens when socialists and capitalists compete for scarce resources.
I'm going to try to interpret this as favorably as possible.
The goal of the education system in a country is to provide everyone with a good education, so that a person's success in life is only dependent on their own actions, and not on what university or school their parents could afford. To do that, you have to be able to acquire and retain talent.
This wasn't an issue until research started to move more and more into the private sector. When research is done in companies, it becomes hard to have professors that teach and research at the same time, as they do at universities. Even harder when wages become competitive.
There's no reason education has to be provided entirely by the state. In voucher schemes the state takes a slice of people's taxes and gives it back to them in the form of use-limited credits that can only be spent on e.g. education. Mandatory insurance schemes are similar in concept. You can have entirely private provision in such a scheme.
Price controls result in shortages rather than everyone benefiting equally - this always happens, and in this case the price controls on academic pay are resulting in a shortage of academics.
Before that it was mostly independent researchers, or ones employed at places such as the Prussian Academy of Sciences.
And in that model, research was open and public.
In today's world, research is jept secret, never published, but a software patent is created based on the use case, so that the company can sue whoever copies it, but never has to publish either (software patents do not require publishing implementation detail, in contrast to real patents).
If that is good is an entirely different question (and I think it's not).
> There's no reason education has to be provided entirely by the state. In voucher schemes the state takes a slice of people's taxes and gives it back to them in the form of use-limited credits that can only be spent on e.g. education.
That's basically what the US model already does, and it leads to ever growing tuition fees, and people far over their head in debt before they even leave university.
My recollection is that the entire industrial revolution was based on innovations that came out of the private sector. Or, is the distinction you're making one of teams vs individuals? The size of teams needed to make new inventions has gone up over time I'm sure, but companies led by an inventor with other people working for them on developing the invention certainly predates WW2!
That's basically what the US model already does, and it leads to ever growing tuition fees, and people far over their head in debt before they even leave university.
That's because western society convinced itself that people had to have degrees or else they'd be useless and unsuccessful for life. In the UK universities are basically all taxpayer funded - the so-called student 'loans' are all from the government and aren't structured in the same way bank loans would be, so it's essentially public funded, and people end up in lots of debt there too.
Of course, those who couldn't "bear" (i.e. afford) it didn't take the course.
And then buy top notch to-be-professors' time by an hour.
Make them compete one against another if there's a choice.
All kinds of paper-pushing regulations have to go now, along with people who carved nice positions for them to do said paper pushing, because it's public money completely down the toilet.
An AI winter because the current state of the art is valuable and applicable enough
It's a tragedy of the commons, yes, but traditionally we've used government to handle these. Would that be desirable here? What form would it take, even?
(If someone can provide broader context on https://tuftsdaily.com/news/2017/08/29/professor-computer-sc... and http://tech.mit.edu/V130/N28/tenure.html, I would be grateful.)
1) their are no conflicts of interest in relation to IP. 2) it's cheaper. Companies that give money for grants are bombarded with overhead costs where a large percentage of funds go nowhere near the intended project the grant was given for.(I forget the name and it varies by school) 3) Kids are payed better( non slave wages) 4) Professors get to work with large amounts of data
What usually happens is that do their time,work with the data they always dreamed of, realize what they are doing, are happy or not happy get bored /pissed off,come back and teach.
The entire academia system in the United States is so fucked behind belief, I really have no sympathy having been their myself.
Those do not make a major negative society impact because they tend to be narrow -- e.g., the current one sweeps in just AI/ML, not all of CS.
This means that while there are a few minor short term disruptions (e.g., a new student might find a scarcity of professors in his #1 choice area for a year or two), it opens up a bunch of opportunities, too (tenure spots, grants, etc.) and in the super-competitive world of modern research universities they quickly get a bunch of qualified applications.
My 2c -- best of luck to those who move to industry, but society impact will be minimal if any 10 years down the road.
I don't think innovation is being stifled. Every week there is a new amazing publication on AI/ML, but instead of coming from University it sometimes comes from DeepMind , Google, FB, etc...
If you want to research self driving cars, wouldn't you want to work at Google which has tons of data?
No-one outside Google knows how deepmind works. No-one can reproduce it.
For anything, actually.
They train all their networks on data taken from users, and publish neither the data nor the models. And the scientific world doesn't have the resources to replicate anything close to it.
In my experience, often all you ever get are the papers.
Now the tables are flipped and academics are the ones not getting access.
A culture of real open access would be nice.
As to why it got published? I don't think it should have been.
I'm not going to compete for a lottery ticket to take a huge pay cut to go work as a professor, with fewer resources and a worse schedule than even mid-level SDEs have. That's clearly not a workable solution for having the best and brightest train the next generations.
But I am concerned by the prospect of a brain drain in the training pipeline and am interested in helping with that. What are viable strategies to try and share some expertise back from industry?
Setting aside for a moment if my employer would agree or not, what should I even be pushing for as a collaboration/training program?
I guess there's a huge amount of hype about statistical machine learning right now and some people will think they 'd better capitalise on it while it lasts.
Still, that's short-termist. If they hype ends soon, the majority of those students leaving their degrees will find themselves out of work _and_ without a degree. If the hype doesn't end soon then there's no need to rush.
So in any case, there's no need to rush. You'll get a better long-term deal if you stay where you are.
This seems to be a case of them getting angry at people for having listened to them. If you tell people that the reason to go to university is to find a higher paying job, it's not their fault if they don't value your university once they find a higher paying job.
So why the concern here?
What will these PhD students gain by staying, when it is so saturated? It appears this is more about tenured professors not being able to pump out papers as easily as they could in the past. For them, the students are merely tools to do so.
As someone else commented: Where is the student's perspective in this piece? Many, many PhD students, if given a much better prospect at a TT position at a university, would reject large salary offers.
https://www.cloudpassage.com/company/press-releases/cloudpas...
I just wish the Apple, Facebook, Amazon, MSFT, Google could get together somehow and provide an alternative to the system we have now. The disruption would be welcomed among the millions of students and tax payers with open arms.
Today, you are losing top talent to another company in your country. Tomorrow with that kind of attitude and "solutions", you will lose them to another country. For better or worse, money is one of the top motivators and the society just has to deal with it.
No more no less.