Slashdot Mirror
AT&T Labs vs. Google Labs - R&D History
An anonymous reader writes "Ars Technica has a piece looking at the history of corporate R&D, in response to an article on the BusinessWeek site essentially calling the telecommunication giants aging fossils of communication. The Ars piece looks as several innovations to come out of the AT&T Labs over the years, as well as the era of innovation brought on by the Cold War." From the article: "The Cold War, with its 'Pentagon socialism', combined with large corporate monopolies that were expected to provide lifetime employment and pensions, made for something of a golden age for American technological innovation. This is the era that brought us the transistor and the predecessor to the Internet, an era where all the seeds of today's 'information economy' were sown and carefully cultivated at great private and public expense. The great labs of this era--Bell Labs, Xerox PARC, and IBM's labs--were places with massive budgets, where the world's top scientists were invited to pursue "blue sky" research into areas with no immediately apparent commercial applications. The facilities were state-of-the-art, and there was no pressure from management or shareholders to do anything but science for science's sake."
Meanwhile, back in America, a perfect storm of rent-seeking behaviors by entrenched players, a broken patent system, a lack of substantial corporate oversight, and old-fashioned executive greed threatens to drown the fabled "two entrepreneurs in a garage" just as surely as those two guys helped sink the blue sky research labs of the Cold War era.
I love America. God Bless the USA.
While Google is definitely doing some cool stuff, what they are creating, and the environment that they are creating it in can't really compare in scope to what happened back in the heyday of big r&d. Google Maps/Earth is cool, but how does it compare to shaping everyones lives like color tv and the transistor. The innovations of Google are significantly more evolutionary vs revolutionary.
We used to call it Bell Labs. Getting a job there was like the ultimate geek cred.
A feeling of having made the same mistake before: Deja Foobar
You know the world of today sucks when you're nostalgic for your parents good old days.
*** Sigs are a stupid waste of bandwidth.
Take those "go-getters" of the hey-day, compare the educational curriculum, pop culture, and political philosophies of their childhood to those of our children today.
Just a hunch, but I suspect that comparison will show darker times ahead for the U.S.
@HbFyo0$k8 tH!$
Google has some of the best scientists around. Yet what do google labs give us? autocomplete for search strings? The only thing that seems worthy of notice in Google labs is google sets, which has that 'next gen AI search' feeling to it.
:)
The same goes for Microsoft research: while there are some gems in there, you will see people presenting research on new ways for drag and drop and similar stuff. While that's useful, it's nowhere near what IBM, PARC and others were/are doing. Even Sun seems to have cooler research projects.
Either those next generation companies are not as scientifically inclined as the old 'dinasaurs', or maybe the truly amazing stuff MS/Google have is hidden from prying eyes till the market is ready for them
No we don't. Patent trolls BUY patents, or patent OBVIOUS things, and then use them as weapons for extortion. Bell Labs and PARC invented real technologies. I'm not saying that they didn't do their share of patenting stupid shit, but they did real research, and their parent companies/divisions actually deployed much of their technology in the real world, not just as a licensor.
I'm not sure but I'm pretty sure I've been getting emails telling me how I can make it bigger and/or longer.
And then comes a series of decade-long court battles over who invented what.
Take for example the Xerox PARC "Unistroke" patent. I happened to visit PARC before I saw the first PalmOS machines come out, and saw Unistroke in action. Some conference rooms had wall-mounted "sign up" devices on the wall by the door, which offered unistroke entry. PalmOS comes out with a very similar "Graffiti" concept. Great fit for the idea-- arguably better than the whole-word recognition that Apple Newton was trying. Several years pass where everyone who was anyone learns how to jot down stuff in Graffiti. And then the lawyers got involved. Over ten years later, the dust is starting to settle, and for what?
And those who didn't enter their thoughts in one-stroke alphabets entered their thoughts with teeny two-thumb keyboards. Hm, that sounds familiar... RIM Blackberry vs who was that?
No matter which side you choose to support, and I think everyone's put forward good arguments for and against every conceivable angle, when it ends up in court, everyone loses .
Pure research is great. Xerox got burned in the whole Apple Lisa / Macintosh thing, so they sorta swung the other way with Unistroke. There has to be a middle ground, though. Right?
[
No, I refuse to think that is a digg title, the Digg story title reads something like:
OMG_HOW_GOOGLE_KICKZ_AT_T_LABS_AZZ_!!!1
Ubuntu is an African word meaning 'I can't configure Debian'
I really miss school. Now, all anybody wants is results.
Research shows that 67% of those who use the term "research shows", are just making shit up.
The question back then was, "How can we outdo the rest of the world?"
The question today is, "How can we maximize our ROI?"
Once money becomes the driving goal above all else quality and innovation suffers.
I remember Alan Kay saying that Xerox wasn't easy on Xerox PARC. It was PARC's directors that shielded the researchers from the corporate pressure and gave them the time and space to do their work. Not Xerox'. So I don't think these historical companies had a grand vision of research. They had good research directors. Note also that some well known projects survived because they were kept below the management's radar and caught on outside the research lab. Both UNIX at ATT and HTML/HTTP at CERN took off partly because the management didn't care much about them.
Although you got modded down, I was thinking about what companies are like Bell Labs/PARC/etc. today. It's a pretty short list. I'd say that IBM is still on there; they still do some stuff that gets into pure research, although I think it's become more market-focused than it used to be; Google strikes me as someone who is trying to take up the helm that was dropped by Xerox PARC -- a combination of marketable stuff and real blue-sky tech ... but I think a lot of other research has moved from the corporate sphere to the realm of small startups. \
It seems like people who are coming out of grad schools now don't hope to get a position as a Fellow at IBM as much as they hope to get a big wad of funding from somebody (usually without thinking too hard about who "somebody" might be) and playing the startup game. Even though as a startup, you usually don't have much flexibility or opportunity to do research, it's all about productization.
I'm still not sure though that I would put Google into the same category as the old research companies of the Cold War era. Google's stuff is good, and it's definitely innovative, but in many cases it looks less like actual new knowledge development than just new and different ways of recombining existing stuff. That definitely has value -- don't get me wrong -- but it's different than the huge amount of capital investment and long time horizons that used to be the norm at Bell Labs, for instance.
Honestly I think it's the time horizon issue that's the worst part of today's market. I don't know if it's a product of instability -- nobody is sure what's going to be going on in 5 years, so they only plan for two -- or if it's just the desire to make short-term gains, but I think that we're starting to see the effects of lots of places not having a very coherent long-term strategy. Stagnation is bad, but a certain amount of predictability in the market can be good, if it lets people plan for longer, and thus take bigger calculated risks.
Nobody is willing to pay for research that might take 10 or more years to productize in today's market, and thus the burden falls on government and academia. They're basically some of the only institutions left that can afford to plan in multiple-decade ranges.
"Ladies and gentlemen, my killbot features Lotus Notes and a machine gun. It is the finest available."
In 2005, Microsoft spent about $7 billion on research and development (R&D). By 2008, the R&D budget will grow to $8 billion. If my memory serves, no American company spends more money on R&D than Microsoft.
The research division at Microsoft is the #1 industrial laboratory in the United States. To understand the magnitude of the largesse, note that Microsoft succeeded in convincing several tenured/tenure-tracked professors at top-notch private universities (e.g. Stanford University) to quit the university and to join Microsoft.
Like the pre-breakup AT&T, Microsoft is funneling its monopolistic profits into a massive R&D budget. Microsoft laboratory has become the "Bell Labs" of the 21st century.
Look at every generation, and its parent generation and... you'll see that not generations were equal, as scientific progress goes. It goes up and down like a yoyo, and it did so since the beginning of time.
E.g., ancient Egypt must have started with some really bright minds, since they discovered a lot of things. And I mean including a ton of medical and other stuff, not just how to pile stones in a pyramid. Yet right before the macedonian invasion it was already at a stage where nothing much was invented any more. Medicine for example had been solidified into something that was religion, law and malpractice insurance rolled into one, and everyone just followed the same official texts literally, and never tried anything new. For _millenia_.
E.g., in Europe the golden ages of Greece and Rome were followed by what we call the "Dark Ages". It's not just that they discovered fewer things, it's that actually a lot of information has been _lost_ in that time.
E.g., take China. It was at one point one of the most technologically advanced places. They have a long list of inventions, including stuff from paper to gunpowder to trebuchets to crossbows (including the repeating kind) to the compass to god knows what else that they invented more than a millenium before the Europeans. They were _that_ advanced. Even their less glamorous stuff, e.g., the composite bow, might get less hype, but you can see its efficiency against European equipment and tactics when it was brought over by the Huns.
Yet then came an age of decline and it ended up with the Manchu Qing dynasty, where literacy actually decreased and the government was literally more concerned with enforcing a uniform haircut (yes, I'm not joking) than with any kind of science or technology pursuit. The Chinese army actually regressed from having _some_ guns during the Ming dynasty, to all spears, swords and bows during the Qing dynasty. That sad.
Or take Japan. Yes, now they're doing damn good technologically and have been even more impressive as progress goes during the Meiji Restoration. But before that they had periods when it stagnated or even regressed. E.g., the Heian period, also remembered for the rise of the Samurai caste, is also considered by some a time of stagnation and even regress.
So, yes, times can change. Sometimes for the better, but sometimes for the worse. Some societies fail to give those "mundane Joes" incentive to go and learn or research something. Yes, each individual can decide what to do with his life, but if on the whole it's a smarter or more popular choice to aim low intellectually, people may well do just that. And then stagnation and even regress follow.
A polar bear is a cartesian bear after a coordinate transform.
What happened, as best I can tell, is that shortsighted corporate executives forgot that (applied) R&D rarely produces new fundamental knowledge about the universe while that is the main goal of pure research. A lot of great research is done when true scientists are given a budget that has already been written off by the bean counters, as IBM and (the old AT&T's) Bell Labs demonstrated many times.
The problem is that such research tends to be very expensive and non-geeks just aren't interested in results they can't understand. The only reason we have nuclear power today is that the United States was willing to spare no expense to develop a bigger and better bomb in order to win WWII quickly an decisively. Nazi Germany sponsored a lot of good science and then took some of the results with military potential and did a tremendous amount of R&D to create amazing new military technologies...tech that just happens to have had amazing commercial potential. Jet aircraft and booster rockets come to mind.
You will hear NASA fans gripe because now that the Cold War is over, NASA has to justify whatever it does to the drones in government who get paid to eliminate government waste. NASA is no longer a great source of new scientific and technical knowledge, but it probably could be again. So could a lot of private enterprises if NASA and other parts of the U.S. government didn't have a practical monopoly on many interesting areas of research.
For major research projects to get significant funding now, they either have to have tremendous (and fairly obvious) commercial potential, or be extremely trendy, in a politically correct sort of way. No expense (to the taxpayers) is spared protecting "endangered species" that (AFAIK) have no real significance except that they are about to succumb to Darwin's Law -- despite all the bleating of the ecowackos, wasting money on the Red-Cockaded Woodpecker is not going to produce new knowledge or improve the chances of Man surviving another century. Having plentiful, cheap sources of energy would.
But try to get money on the scale of the Manhattan Project for the purpose of finally developing nuclear fusion power plants... That is not by any means pure research, but the amount of pure research that can only be done with the kind of energy a large fusion plant could produce is staggering. But why stop with fusion? Total conversion seems about as likely to be a practical source of energy now as utilizing light pipes and orbital spacecraft as the backbone of a worldwide communications network did during WWII.
Do you think the U.S. might have fusion power plants online and/or total conversion reactors in the lab by now if such projects had received oh, say $100 BILLION dollars in additional research funding since WWII? That's a Big Pile O' Money! It also happens to be roughly what the U.S. has wasted on handouts to Israel since that nation was created by fiat in 1948. Why not just cut all foreign aid for non-humanitarian purposes (Israel gets only about 1/3 of the U.S.'s foreign aid largess, after all) and use the proceeds to fund a pure research lab or ten that are operated by private sector organizations that have track records of doing cutting edge research and producing useful knowledge?
Stop real government waste and use the savings to fund hard science research projects that short-sighted bean counters consider waste because they know no better, ignorant touchy-feely nitwits in search of warm fuzzies and/or vote generating pork-barrel projects that they are.
"You're young, you're drunk, you're in bed, you have knives; shit happens." -- Angelina Jolie
You do not realize the significance of these?
PageRank is a method by which billions of related and interlinked pages of information can be searched across, that returns relevant results.
They managed to (nearly) tame the beast that the World Wide Web had become. The fact that they managed to do this using an almost sociological approach is all that much more amazing.
Which also represents a new form of interaction with threaded information. Not the most revolutionary thing in the world, but hey, technically the LED is just another form of light.
Which represents just one application of research into machine learning.
I am on a subscription mailing list for intern employees. Two topics that come up often are car pools and drinking. Google's contextual advertising engine is so smart, it starting showing me ads for DUI lawyers next to emails from this distribution list! That freaked me out a bit, Google's computers had managed to learn that this distribution list consisted of people who drove around a lot of drank a lot of alcohol. Woh. The fact that Google is using that technology to show ads does not make it any less impressive. As it is also impressive that Gmail knows when my GF sends me a short message "go see superman next Saturday?" Gmail asks me if I want to add "Going to see Superman Movie" to my calendar.
Google's research is rather limited in that they primarily (solely?) deal with information theory, but within their research domain, their findings are quite amazing. Indeed, others have tried hard in the past to achieve the same results, and others still try today. Ask.com has managed to pull off some pretty amazing stuff (which is then replicated by Google is, oh, say, about 3 seconds.
Need help treating your acne? Come here!
and although I've never worked for, nor personally known anyone who has worked for Google Labs, they seem to be about the closest thing I've seen since.
Bell Labs served as the R&D arm of AT&T, Maw Bell, "The Telephone Company", a highly regulated utility. Because of it's monopolistic and legally protected position, AT&T back then through off copious amounts of cash, and was considered an exceptionally safe investment.
Bell Labs was funded by part of this cash flow and had an incredibly broad mandate towards basic research which showed up in the work people did, that often didn't have (immediate) commerical applications.
Unix, for instance - AT&T couldn't even sell it back then, due to their monopoly. But folks at The Labs kept on exploring, improving, conducting basic research into Operating Systems that we still benefit from decades later. My office mate at the time was working in fiber optics, and thought back then in 1984 that his work "might" have commericial telephony applications sometime past the year 2000 after the development of several enabling technologies.
Everyone was encouraged to present papers internally; every day there were loads of seminars and working groups during office hours and, of course, the informal meetings and brainstorming sessions that took place at pubs and strip clubs along the New Jersey coast.
Your manager typically was also doing his or her own research, and would help you to explore specific areas of interest that might not be precisely part of the department manadate.
Highlight of my time there: taking several lunchtime seminars in a new programming language called C++, presented by Bjarne Stroustrup himself.
I think Google with their model of lettign engineers do whatever they'd like one day a week is the best sustainable compromise between fully commericial companies such as Microsoft or Apple and pure research organisatins such as Bell Labs.
Short of government funded, open ended research military reseaerch - I did that as well, and while it may not seek to commericalise the research but they sponsors will have "other" uses in mind - it's probably the best thing we've got going for us right now.
I'll leave you with a toast that I picked up from some of the older engineers during my time at Bell. We used it during many an evening at the local strip clubs:
"Stronger Whiskey, Younger Woman, Faster Computers"
Ahh, the good old days.
A message from our sponsor
Having been to Bell Labs (Murry Hill NJ) and worked with some of the people when it was in its prime, I think the article fails to appreciate some history. First, AT&T is gone. And when it was recently brought by SBC it was a fraction of what it was.
Back in the day, there was AT&T which owned Bell Labs in Murry Hill NJ. This facility was the envy of every major company in the world. They did research in both hardware (physics, chemistry, integrated circuits, etc) and software (UNIX, C etc.) Of course they had their "feet on the desk noble prize winners" but the majority of the researches had goals that served the corporate interest. They did understand that fundamental science can pay off in the longer term, but today's short-sighted next quarter stock price mentality prevents this type of strategic thinking. For instance, AT&T developed in-house hardware and software because they needed a way to track (and bill) phone calls. They needed to understand fundamental physics and chemistry because deep sea cables and communication satellites are things that are not easily repaired.
Now what many people forgot, or don't know is that AT&T broke in two parts many years ago: AT&T Communications (took software R&D) and Lucent (took hardware R&D). Lucent took over Murry Hill as its HQ and AT&T Research moved to Florham Park, NJ. Lucent has since also spun off Agrere. AT&T sold their wireless business to Cingular, and what was left at that point went to SBC. So saying AT&T of today (a renamed SBC) is has a powerful research arm is like saying Micky Heart is the Grateful Dead. They do good stuff, but the magic is gone.
As for Verizon. Their only claim to fame is the biggest tax bamboozle ever pulled off by a company.
HPC for Primates. Read Cluster Monkey
That $7B includes all product development -Vista, next version of SQL server, etc, all in there.
The amount spent on "corporate research" is a lot less, probably no more than $100M, though that is a rough guess.
The other thing is yes, they've hired some great people. Lamport, for example. But hiring people because they did great work in the past does not mean they will do great stuff in your company. I've seen that in my own.
Microsoft R&D is, on most cases, not true R&D, but product development. The same can be said about Google. So far, big science funding by large corporations is solely represented by IBM, who funds research on fields from nanotechnology to biological research. Look at how many Nobel Prize winners they currently employ. Now tell me how many are working for MSFT. Do you really believe you compare some of the finest IBM research with, let's say, winFS? And what is good in IBM research is that some of this research is actually translated into profitable products, what let the shareholders happy enough to make them let the money flow to R&D without complaints.
Your ad could be here!