Facebook Adds A.I. Labs in Seattle and Pittsburgh, Pressuring Local Universities | Hacker News Reader

Facebook Adds A.I. Labs in Seattle and Pittsburgh, Pressuring Local Universities

nytimes.com

50 pointsrbanffy39 comments

Previous discussion

https://news.ycombinator.com/item?id=17003244

Mobile website strikes again :-(

University that want to retain computer science professors, especially those working in hot fields such as artificial intelligence and machine learning, will have to pay them well. The main problem may be political -- can they pay professors in different departments very different amounts?

One reason it is difficult to attact high school teachers who can teach computer programming well is that teachers' union contracts say that you have to pay gym teachers and CS teachers the same amount.

> can they pay professors in different departments very different amounts

In the US this is already common. But salaries in AI (and to a lesser extent CS in general) have recently gotten high enough that you'd need much bigger differentials, plus you start getting to absolute salary numbers that become unaffordable to all but a handful of places. An early-career academic at a top university might make, say, $60-80k in the humanities versus $100-130k in CS, but companies will pay $150k+. A prominent tenured professor in CS can make $200k+ in some cases in recent years, but if you believe recent reports [1], senior AI researchers are being paid $1m+ outside of academia, which seems really unlikely to be matched.

I'm not sure it's necessarily a problem per se, though. Depends on what your goals are. As a fairly junior CS academic, it's good for me personally. CS faculty jobs are now much easier to come by than is usually the case in academia (look at the nightmare stories of physicists and biologists doing 10-year postdocs trying to land a tenure-track position). And pay has risen enough that it's now pretty good, even if still not as good as industry.

[1] https://www.nytimes.com/2018/04/19/technology/artificial-int...

Professors taking sabbaticals to work at private companies seems to be a common pattern as well—and probably a generally good one for everyone involved.

That shouldn't really be a problem, and is already true in a lot of cases. For instance, professors in business schools get paid significantly more than professors in the economics department of the same university.

This, most economics programs aren't attached to business schools because of that. At a more granular level, Finance also gets a nice bump due to private sector pressures.

"can they pay professors in different departments very different amounts?"

I've worked at a couple of universities, and as far as I understood it there were special provisions for professors of law and medicine, meaning that they got paid more, and that they could work outside of the University to a greater degree than senior academics in other departments.

can they pay professors in different departments very different amounts?

Absolutely. The baseline salary for a given rank varies across divisions (salaries in STEM fields start higher than in the social sciences, which in turn pay better than humanities) and varies across schools (law and medical schools pay better than arts and sciences, since the former compete directly with the private sector). More importantly, faculty compensation is negotiated on an individual basis, and deans have tremendous leeway in how much they can offer to retain talent in the face of competing offers. At a first- or second-tier research university it's typical for star faculty to have a salary that's 2x or 3x the department baseline (or more).

This wave of hate starts looking funny already. Facebook pays researchers decent salaries? Let's say it "pressures local universities".

Is $180,000/year not a decent salary in the Seattle area?

Not for top AI talent. That starts at $400k/yr

Hope they save because "AI" won't be hot past 2020.

Au contraire, my friend, it’s only in the beginning of an upward hockey stick.

You mean because it'll take all our jobs, including programming AI? ;)

What makes you think that? Media? What are you including under the umbrella of AI? ML, Deep Learning? Are you in the industry?

What industry?

I sell AI robots and its bullshit

Ah, different use case then as opposed to using AI or ML for drug discovery.

I agree, robots are hit and miss

I will make sure to revisit this comment in 2020.

AI won't disappear definitely. However, more graduates will enter market as we move forward. Thus there will be more supply than demand which leads to drop in salaries.

This has happened in many markets in recent years. Don't you recall big data hype around 2007? You would be considered as God if you can barely install Hadoop back then.

Competent big data folks still pull down $300-400k fairly easily. I’m talking about people who _really_ understand the field and have a track record, not “graduates with a degree”.

starts at 400k? where do you get that number? I can imagine a director 3 type position, which would be true for most fields. But interviewing, talking with recruiters, and looking at salaries on glass door, facebook pays a median 130k. Amazon pays closer to the 100k side with a maximum of 165k, an absolute ceiling that shall not be passed. There is stock, but even being equal to base, still no where near a starting 400k

Someone coming in as a research scientist (ie, the people that would have become professors) is going to be making more.

We pay this much to strong PhDs with a good publication record. So do Google and FB. Must be able to code really well, though.

OTOH, having many of our most brilliant minds all working on putting more ads in front of more eyes to get more clicks isn't such a great thing.

hmm, they will develop more intelligent systems which will have applications in all domains.

Scenario: a well-placed middle manager has a mid-life crisis and authorizes a skunkworks project to have an AI surreptitiously replace them while they travel the world. When other managers discover this, they demand their own AI surrogates rather than reporting the scheme to the C-suite. When this is eventually exposed, FB collapses in a firestorm of controversy, only for its investors to quietly start a bunch of new companies intentionally based on AI management. Eventually Accelerando, I guess.

A generation of brilliance wasted on surveillance technology.

To be brutally honest, I must confess that as a European in a peripheral university, also working in an AI subfield, I feel a bit of glee about this.

Those elite American universities have been sucking talent away from all over the world, and working with an amount of resources that provide them with a considerable advantage.

Now they receive some taste of their own medicine.

Note: I'm aware that there is nothing bad in what those universities have done (probably every university would do the same if they could or knew how) and I shouldn't feel like that! But I'm just human, and unfortunately far from being a perfect human... and being beaten to results by people that have 3x as many PhD students and 10x as much GPU power ends up creating bitter feelings.

This is just going to make it more competitive for everyone. Even more European professors will be hired away for even more money.

Note: I am not saying it’s better or worse, just (much) more competitive.

Yet I'm not sure that this Schadenfreude is at all useful for you, since these will instead also outGPU your research. I hope you can come up with a great AI business and compete with these companies, rather than be sad at the structural realities of global capitalism!

In my opinion, shame on any professor who joins an organization which has shown such disregard for their users without at least demanding reform.

I agree!

I just can’t tell whether you’re referring to universities or to tech.

He is clearly referring to the guys charging 50k / year to 19 years old.

*19s year old

Surely lecturers will just keep one foot in each world? Back before '08 when finance gurus were killing it, many of them kept ties with academia. It's a useful insurance policy, and it's great to be able to fall back to teaching or research once you've had enough of industry.

I love that even the new york times is apparently incapable of introducing a cookiebar that conforms to the actual law they're trying to conform to. It's not especially complicated to not give me cookies unless I say I'm OK with receiving them, instead they've gone with the "implied consent" option (again) that is pretty explicitly not consent anymore.

Reply on news.ycombinator.com