Slashdot Mirror
AT&T Labs' Brain Drain
Frisky070802 writes "The Newark Star-Ledger has an article on the brain drain at AT&T Labs, which laid off close to half its researchers two years ago this month, another good fraction last spring, and has lost many of the rest through voluntary departures. The article claims that only Microsoft might have the money to fund basic research as Bell Labs did years ago, though many (including me) would put IBM in the same camp. It cites problems at AT&T, ranging from researchers paying their own way to present at conferences to a loss of free espresso and bottled water. Many luminaries, such as Lorrie Faith Cranor, Avi Rubin, and Bjarne Stroustrup, are quoted --- with Stroustrup saying the lab was "mugged" by Wall Street. (Rumor has it that the losses haven't stemmed, with more top-notch researchers going to academia in the coming months.)" (Non-registration ZIP and age demographic collection.)
He's an endowed prof at my alma mater www.tamu.edu. Hope this improves the CS program there.
When travelling, it's ok if the airlines lose your emotional baggage.
Well, let's not forget the original point-contact germanium transistor (granted it was called Bell Laboratories back then.) Pretty much set off the entire solid-state revolution in electronics, which after nearly half a century has culminated in that paragon of technological debauchery known as Slashdot. But seriously, Bardeen, Brattain and Shockley would no doubt be hurt not to have their brainchild included in your list of great inventions.
The higher the technology, the sharper that two-edged sword.
Bell Labs was the brain power of AT&T and they went to Lucent when the company spun those business off several years ago. Did I miss AT&T picking them back up or something?
'Same speed C but faster'
But is open source really 'picking up the slack' that fully blown, say again **paid** research is?
... and at least it will help support Open-Source stay afloat.
/me goes back to watching Mozilla compile...Only 30 minutes to go!!!!
Most people who contribute to open source have other jobs and can't spend huge amounts of time with the project they start/help out at. And when they move on to other projects someone else has to play catchu-up and figure out the source before any more progress is made.
Open source projects need a constant revenue stream and the model needs to acknowledge that as donations don't seem to crack the honey-pot enough. If companies like AT&T, IBM, AOL and, heaven forbid, Microsoft supported the open-source initiative and gave projects a certain amount of funding then of course more people would support it. But the real question is 'how do you get that sort of funding?' (You can't... it's generosity)
Now what if those projects who have an open-source root, like Poseidon UML just off the top of my head, all donate a portion of THEIR earnings to a generic project each month? Again, another reliance on generosity.
I don't mind the idea of commercial versions of Open-source software as well as free versions, it makes sense - but the advantage is that she's not only a revenue stream but in addition it's one that gives you money thats an indication of how much effort you put into good ideas.
On the other hand, since generosity is hard to come by... what about commitment?
Firefox/Thunderbird, TightVNC, AngBand, Apache, MySQL+MySQL Front, PHP and most definetly Open Office - the top 7 things I love about free source - all staying afloat and continuing their 'good work'. All Free, All being updated regularly enough to keep people happy with.
But when you go to sourceforge and see 'Stage 1 - planning: December 2001' (and nothing put out since then) it makes me wonder whether some people are REALLY dedicated to what they do.
And oh bugger.. it didn't take even nearly long enough to write this up...
I have to disagree here. Especially in storage (look at IceCube as an example), there's all sorts of good things going on, and that have gone on in the past (like magnetic disks).
Mencken had it right. So glad that's old news.
I remember, back in 1987 or so, getting a good look at a computer industry study that showed gross revenues, margins, and so forth for pretty mich all of the companies in what one would consider "the computer industry" of the time. It also showed how much they spent on R&D.
Sperry spent a decent amount; so did Cray, and Hewlett Packard, and AT&T, and NCR, and so forth.
IBM spent more on R&D than the rest of them put together.
In fact, IBM spent more on R&D than the gross revenues of the second-largest company. Not the profits, mind you -- the gross revenues.
That was the single most gobsmacking business statistic that I heard until the one a couple of years ago about how Microsoft could purchase the airline industry out of its cash reserves -- twice .
If you noticed, I had said Microsoft Research - not Microsoft itself.
:)
You don't have to believe me - look at Microsoft Research's Publications page - that should give you an idea
The thing is that what business wants is not necessarily today's research - that is precisely my point. At this moment, sure, what MSR patents may not be quite important - but tomorrow, it could be the basis of a whole lot of things.
Microsoft has done some really really cutting edge work in Natural Language Processing, Graphics, Knowledge and Data Mining, HCI and Ubiquitous Computing - in fact, its hard to go without seeing atleast one or two publications by MSR in any respectable conference/journal that deals in these areas.
But - its not all out in the open - and thats what you should be worried about. Because Microsoft has NO need to bring it out in the open, until it has to. Hidden knowledge is more potent.
I was at conference at Bell Labs/Lucent not too long ago and I think part of what is happening is a natural shift in what matters in corporate research. I got the sense that Bell Labs was shifting slightly from its physics/hardware roots to math/algorithms/software future. They still do physics, but they also do proportionally more R&D in the idea/software space. (Disclaimer: I didn't see any budget figures or top secret stuff, so who knows what they really goes on in Murray Hill)
I'm not saying that we should stop R&D on hardware, solid state physics and materials, only that new software and software-related tools would help everyone get the most out of the current portfolio of hardware technologies. Given that we just discussed "Why Programming Still Stinks" and have not discussed "Why Hardware Still Stinks," I would suggest that the bigger research opportunites are in software.
I also suspect that software is more commercialization-friendly. If you look at research advances in hardware/materials it takes 20 years before it makes it out of the lab. By the time a fundamentally new invention is in mass-production its is off-patent. I know BellLabs invented the transistor and the laser, but I wonder what fraction of semiconductor and laser industry's profits actually went to BellLabs/AT&T. In contrast, software can be more self contained and follow a much faster adoption curve.
In summary, I would say that scientists and engineers already have a reasonably good handle on atoms and that the real R&D opportunities are in getting better with bits.
Two wrongs don't make a right, but three lefts do.
I rather create something than make something better.
But perhaps we need some sort of balance between the two.
As a former labs person, one who was included in last years outsourcing, it is not a surprise to see this happen. For right or wrong the new management has chosen this path and they are succeeding in an alarming rate. What they are succeeding at I have no idea beside the destruction of the Labs and the company as a whole. Am I a bitter ex-employee? sure... but that doesn't change the fact that that it is happening.
The president of the labs is to credit or blame as you see fit. He has a strategy and he is going about it quickly. Is it a good strategy? Time will tell, but it is not one I believe in, nor do I believe in their president, even when I was a loyal employee. He is downsizing research and development and trying to buy off the shelf products for a company that really has no peer in size. Let's face it, the reason why AT&T had to develop all of their own stuff internally was because there was no one on the outside developing towards that market and could achieve the quality that was desired. They have special needs that outside vendors, for the most part, can't fill, but they try and stick the square peg into the round hole.
When AT&T was "The Phone Company", it funded bell labs with an internal "tax". That means that every department in the company would take 10% off the top of what they brought in and send it to bell labs. It was very well funded and the R&D consistently paid off.
Now the stock market is a major player in moving money in and out of compaines and they don't like research. It even appears that most of the major funds consider it a bad gamble in most cases. The side effects of the short sighted profit is that the US economy is loosing 1.3 billion dollars a day and the pyramid scheme that used to prop up some of the economy is falling apart.
Congress needs to start intorducing tax cuts for real R&D.
Writing standards actually work is the job of engineers, just as basic research is the job of scientists.
It's not that research can't and doesn't happen at universities, it's that the success and mass of the big corporate labs meant that larger projects were within their grasp.
[Set Cain on fire and steal his lute.]
You give the Nobel Committee too much credit. Many in academia find the whole concept of the King of Sweden identifying a handful of outstanding scientists distasteful. For every Nobel prizewinner there are hundreds of people who did or are still doing work just as good. It's very demoralizing to the remaining hundreds, especially when the winners are usually late comers to the field, often hijacking others' ideas and then politicking hard to get a prize.
They also did not maket and made very little money off Unix. It was a failure ( in business, not technical or scientific) sense.
The problem I see is that big business want profits, not innovations. Its not worth innovating unless something is profitable. If something is profitable then you have to figure out if its worth the cost to develop and market it.
MS Windows is a perfect pure example of this paradigm. Bill Gates wanted every feature developed only if it could be profitable. It was not profitable to make a good OS because that would cost more money. For example it lacked many software out of the box that Unix had, and did not include protective memory or premptive multitasking until NT. Only when businesses demanded these things that they were added.
There were many multitasking multiuser operating systems and programming langauges like C/C++ in the early days. Not as good of course but from a business standpoint good enough.
In this day and age of maximizing productivity, Ken Thompson would probably be fired for writing Unix as a joke on Multics back in the early 70's by wasting company time. He was not ordered to write it at all actually. I am not agreeing here since what he did was great but I am just giving the business standpoint and different philosphy of research vs profitability.
Xerox could have been profitable but had incompetant management. They had waaayyy to many employees so they would need to charge a fortune for every gui driven desktop in order to make a profit. They did neither of this well, which caused %90 of it to close shop.
Today since businesses do not like hiring they hook up with Universities for small projects. That is where real innovation is that is not hampered down by need of profits or managment bueacracy.
Mosaic is a classic example which started the www revolution.
http://saveie6.com/
That is the problem stated by the author.
Its not profitable to innovate and write good software.
MS was very efficiant in terms of making the bare minimium because it was cheaper to develop.
AT&T on the other hand lost money by being too innovative. Unix was great but made very little money to AT&T oddly. There were many other OS's that were multiuser and multitasking. This brought demand down.
I think there is a conflict between R&D vs profits in todays world. CEO's are obsessed with having the maximium productivity done with the least amount of people or white collar workers ( cough India).
Ken Thompson and Ritchie would be fired in a second if they worked on Unix or C today because they were not ordered to do so and it would not be profitable. Sigh.
http://saveie6.com/
SCO, to their credit, is the only company who seems to be determined to cash in on some of that old Bell Labs Intellectual Property. [Understandably, SCO also harbors a deep, abiding grudge against IBM for the way they were betrayed in Project Monterey.]
But, in the world of oenophilia, Mondavi's Opus One is an utter and complete joke:
> The truth is, none of the world-class pure research labs (Xerox PARC, Bell Labs, TJ Watson, etc.) do a good job of helping their parent companies in the long run.
I used to work at Bell Labs. The amount of money made from patents and intellectual property licenses derived from Bell Labs research is staggering. You have no idea.
There one around (I don't remember which) that was still generating >100M a year more than 20 years after it was filed.
IP is where the money from research is made.
"I'm an old-fashioned type of guy. I worship the Sun and Moon as gods. And fear them."
First, most industrial conference fees are not in the thousands of dollars. About five hundred is more typical. These conferences do, after all, want good attendance. Second, it is rather unusual for employees to pay their own way to conferences, especially out of town ones (which most are). Third, research is by no means all or nothing. Most of it is incremental improvememnt on existing science, and gives a corresponding return on investment. Sometimes a radicial advancement is made, and this can make headlines, but that is the exception, not the rule. Fourth, you accuse companies of "stealing" ideas at conferences. Well that's the whole idea, ding dong. When one presents a paper at a conference, it's to disseminate ideas. People are supposed to "steal" them, and I take great pride when people steal mine.
I am a software developer and researcher in geophysics. In that community at least, the top researchers are about evenly divided between industry and academic, and no, the industrial researchers are NOT mostly ex-professors.
I have never detected any disdain for industry researchers from university researchers - indeed there are many consortiums between them. I suspect most academics are jealous of industrial researchers because they often have better financial backing and are involved in more "real world" problems. I also think industrial researchers are jealous of academics because they have more time and freedom to tackle basic, pure research. Together they make a powerful combination.
So far as your assessment of the quality of conference papers from industry is concerned, it's just complete garbage. Free enterprize is a highly stimulating environment that attracts talented people, and the papers reflect that. The weakest papers, I'm afraid, tend to be from graduate students, although I have seen many excellent ones. Sometimes, too, overtly commercial papers get presented, although conferences fight to minimize this.
Your rant is misinformed pretty well from top to bottom. I can't imagine why you would make such nonsense up, but it has no relation to reality.
While MS is certainly active in terms of amount of research, I can't say that I think they do the same caliber of research as something like Bell Labs, or even something like IBM.
I can't really give concrete numbers or anything to support my argument. In this regard, it's more of a subjective impression of what they're doing.
I can say, however, that I do research in statistics and information theory, and there is a certain degree to which research from Bell Labs and IBM pops up again and again. The same thing can definitely not be said about Microsoft.
Perhaps someday I will change my mind. Perhaps my experience is with a limited domain. But that would only strengthen arguments that MS research doesn't have the same breadth as other labs.
I have been a research universities in those limbo positions and I have to say that the level of talent I ran into from AT&T Labs was very impressive. Even among larger academic research groups the membership is transient and manytimes the scale and level of excitment could not match those at the AT&T laboratories. Nor the consistent quality of output.
Not that R&D isn't good - but academia specializes in that. I'm just saying that it's no surprise to me that companies heavy in R&D spending aren't doing so great. It's a tough time in technology and there's still room to capitalize on the R&D we already have.