Microsoft SVC

By now, the news that Microsoft abruptly closed its Silicon Valley research lab—leaving dozens of stellar computer scientists jobless—has already been all over the theoretical computer science blogosphere: see, e.g., Lance, Luca, Omer Reingold, Michael Mitzenmacher.  I never made a real visit to Microsoft SVC (only went there once IIRC, for a workshop, while a grad student at Berkeley); now of course I won’t have the chance.

The theoretical computer science community, in the Bay Area and elsewhere, is now mobilizing to offer visiting positions to the “refugees” from Microsoft SVC, until they’re able to find more permanent employment.  I was happy to learn, this week, that MIT’s theory group will likely play a small part in that effort.

Like many others, I confess to bafflement about Microsoft’s reasons for doing this.  Won’t the severe damage to MSR’s painstakingly-built reputation, to its hiring and retention of the best people, outweigh the comparatively small amount of money Microsoft will save?  Did they at least ask Mr. Gates, to see whether he’d chip in the proverbial change under his couch cushions to keep the lab open?  Most of all, why the suddenness?  Why not wind the lab down over a year, giving the scientists time to apply for new jobs in the academic hiring cycle?  It’s not like Microsoft is in a financial crisis, lacking the cash to keep the lights on.

Yet one could also view this announcement as a lesson in why academia exists and is necessary.  Yes, one should applaud those companies that choose to invest a portion of their revenue in basic research—like IBM, the old AT&T, or Microsoft itself (which continues to operate great research outfits in Redmond, Santa Barbara, both Cambridges, Beijing, Bangalore, Munich, Cairo, and Herzliya).  And yes, one should acknowledge the countless times when academia falls short of its ideals, when it too places the short term above the long.  All the same, it seems essential that our civilization maintain institutions for which the pursuit and dissemination of knowledge are not just accoutrements for when financial times are good and the Board of Directors is sympathetic, but are the institution’s entire reasons for being: those activities that the institution has explicitly committed to support for as long as it exists.

85 Responses to “Microsoft SVC”

  1. HDB Says:

    Great post. However, this reemphasizes that academia comprises the sole set of institutions whose “entire reasons for being” is the “pursuit and dissemination of knowledge.” It’s naive to think that businesses, including those with industrial research labs, would prioritize basic research over profit in the long run.

  2. John Says:

    I disagree that’s this scenario implies that academia is necessary.

    Imagine a world where everybody (all us CS theorists) worked for Microsoft or IBM or Google or another company, but we were all still allowed to do research like Microsoft Research allows. But none of us have tenure, and small groups of us could be fired when a lab shut down for no particular reason.

    Would that be so bad? It depends on the firing rate. But it wouldn’t be all that different from the current world. We’d still do our research. And in the current world, a huge fraction of research is done by temporary workers (postdocs, graduate students), anyway. They are mostly in academia, but do not have any job security.

  3. Scott Says:

    John #2: I’d tend to agree with you, if it were clear that the system was in “steady state”—i.e., that for every corporate lab that suddenly fired its researchers, others would be hiring a similar number. I’m more worried about a general economic downturn, or a change in attitudes, that causes many corporate labs to shed their basic research simultaneously. Such things have happened in the past, and in my view, they underscore the need for basic research institutions that are sheltered from the vicissitudes of the market. (Notice the irony: the existence of those institutions could make it easier, rather than harder, for researchers to accept corporate jobs, by mitigating the risk.)

  4. aram Says:

    Another reason academia helps is that the industrial research labs have to compete with academia which helps encourage them to give their researchers some amount of freedom to pursue their own ideas.

  5. Chris W. Says:

    From Scott:

    Such things have happened in the past, and in my view, they underscore the need for basic research institutions that are sheltered from the vicissitudes of the market.

    Trouble is, much the world is being pressured if not actually run by influential entities led by people who believe that nobody should be sheltered from the vicissitudes of the market. The abruptness of Microsoft’s decision seems almost typical of the way such adjustments are made these days.

  6. Scott Says:

    Chris #5: Yes.

  7. kris Says:

    I do think that the shutdown of this lab is a big loss for the research community, and I hope that everyone who was employed there will find suitable positions (given its reputation, I am sure they will land on their feet). However, the nature of the layoffs are not unusual. This is how jobs are cut, and these are the risks of employment for anyone who is not a university professor. This is not to say that it is not painful, or the uncertainties of employment are a desirable state of affairs.

    Unfortunately, academia is also no different (in fact it is worse) for many ( >> tenured/tenure track faculty) employed there. There is the well known problem of adjuncting for bare minimum wage, and the very many postdocs who have no guarantee or even reasonable certainty of finding employment if their advisers run out of funding or decide to get rid of them.

    Sadly, the world of work is an increasingly insecure and harsh place, with the general attitude among the plutocrats being that it is immoral for people to have a reasonable amount of employment and economic security.

  8. Bram Cohen Says:

    Private company research institutions are a bit of an odd bird. They’re mostly set up by giant evil corporations as a PR thing to make them look a bit less evil. Now that Microsoft doesn’t have the kind of unassailable position in the business world that it once did, it has less reason for such things.

  9. Tony Says:

    Chris #5: nobody, that is, excluding themselves 🙂

  10. Scott Says:

    kris #7: Yes, of course you’re right. Much of the “shock” comes precisely from the mismatch between how things work in our little academic bubble, and how they (now) work in the rest of the world. My brother works in the financial industry, so I know from him that the manner of Microsoft SVC’s shutdown would be totally “normal” there.

    It’s true that most people in academia have no job security either. But at least grad students and postdocs know they’re only supposed to be around for a limited time—that “success” means finishing and leaving. At least faculty denied tenure are given a year to look for a new job. And at least whole departments can’t just be arbitrarily closed, without provoking howls of protest.

    As a result, maybe the culture is a little different. In my own experience of academia (which admittedly might be atypical), I’ve witnessed administrative staff who had serious performance issues kept around for years, because firing someone who you knew as a person just didn’t seem like the kind of thing that you do.

  11. Sid K Says:

    Industry research vs. academia research seems to be analogous to benevolent dictatorships vs. democracies. Yes, the former can sometimes get things done faster and more effectively, but is quite susceptible to instabilities and can become indifferent or non-benevolent quickly. And yes, the latter is slow and cumbersome but it is unlikely that the main values would be sacrificed at the whim of a single person or a small group of people.

  12. luca turin Says:

    The comments and reminiscences at Windows on Theory brought tears to my eyes. That seems to have been one hell of a place. What a clumsy way to go about things.

  13. fred Says:

    “Like many others, I confess to bafflement about Microsoft’s reasons for doing this.”

    1) Microsoft is in the process of cutting 18,000 jobs.

    2) Microsoft just bought the video game Minecraft for 2.5B$.

    Unlike Google, Microsoft doesn’t have some magical scheme to summon unlimited amounts of cash and must rely on selling traditional software and hardware to stay in business.
    Someone at the top must have figured that it’s time to trim the fat and focus back on simple and proven products/services that can earn them money.

  14. vzn Says:

    yep a real tragedy but anyone who has studied history knows silicon valley has a lot of ghosts in its halls and even (at times) “skeletons in the closets”. there are many precedents for the closing of the lab & industrial R&D seems to be much more scarce and less sustainable in current economic conditions. there are some fundamental incompatibilities between capitalism & R&D and yes, academia tempers it somewhat, but increasingly academia is subject to heavy capitalistic pressures also. seems its a deep problem in social systems. rising tuitions, large campuses run like corporations, many temporary workers, harder to get tenure, etcetera. its whats known as a “red queen race” in a book by matt ridley. few have the awareness to see it for what it is. one might even indict the decades long political preoccupation/slant with supply-side economics….

  15. Michael P Says:

    IMHO, “publish or perish” in pre-tenure career is not that much different from “develop something useful or get fired” in industry.

    Here’s, BTW, an opinion about tenure shared by quite a few researchers outside academia:
    http://www.theonion.com/articles/newly-tenured-professor-now-inspired-to-work-harde,35169/

  16. John Sidles Says:

    Dystopia  Dystopian STEAM narratives abound nowadays … the comments on today’s Slashdot topic “Ask Slashdot: Finding a Job After Completing Computer Science Ph.D?” provide multiple examples.

    Dystopia redux  This week’s news from the Middle East is even worse: the Duffel Blog’s dystopian historical survey Middle Eastern Card-Stacking Championship Placed On Hold Again is commended to all Shtetl Optimized readers.

    Utopia envisioned  Utopian STEAM narratives consequently have premium value … that’s why Tim Gowers deserves our appreciation and respect (as it seems to me) for the following passage from his recent trip report “ICM2014 — Barak, Guralnick, Brown”

    I always like it when little themes recur at ICMs in different contexts. I’ve already mentioned the theme of looking at big spaces of objects in order to understand typical objects.

    Another one I mentioned when describing Candès’s lecture: that one should not necessarily be afraid of NP-complete problems, a theme which was present in Barak’s talk as well.

    I’m particularly fond of it because I’ve spent a lot of time in the last few years thinking about the well-known NP-complete problem where the input is a mathematical statement and the task (in the decision version) is to say whether there is a proof of that statement of length at most n — in some appropriate formal system.

    The fact that this problem is NP-complete does not deter mathematicians from spending their lives solving instances of it.

    What explains this apparent success?

    I dream that there might be a very nice answer to this question, rather than just a hand-wavy one that says that the instances studied by mathematicians are far from general.

    Needless to say, STEAM history provides plenty of inspiring examples in which Gowers-style optimism was justified.

    An utopian case study: Hugo Black and Bell Labs  Particularly relevant to the Microsoft-labs present-day invitality is Hugo Black’s IEEE Spectrum article “Inventing the Negative Feedback Amplifier”. Black’s obvious-in-retrospect invention boosted phone-line capacity by three orders of magnitute, and revenue associated to Black’s US Patent #2102671 (and its numerous descendants) funded Bell Labs throughout its glory years.

    An utopian question  Do STEAM opportunities of comparable vitality to the Black/Bell narrative exist today? Surely it’s reassuring that Tim Gowers imagines that this might be so! And perhaps Ken Regan and Dick Lipton (who is this year’s Knuth Prize winner) are right too in saying

    “Often—not always—but often when we apply theory to areas from other parts of science we run into negative results.

    Telling people in another area of research that they cannot do X is not usually helpful.

    In many cases they need to do the work anyway, and so telling them a negative is really less than useful.”

    Resolved for Purposes of Debate  BosonSampling researchers can reasonably embrace the optimistic tradition of Black/Gowers/Regan/Lipton by attempting to demonstrate the following postulate (which is opposite to arXiv:1011.3245 [quant-ph] and its descendants):

    The P-BosonSampling Postulate  The data-sets associated to n-photon Scattershot BosonSampling experiments can be indistinguishably simulated with PTIME (classical) computational resources

    Question  What is the probability that the P-BosonSampling Postulate is true under realistic assumptions — both informatic and physical — associated to the definition of “indistinguishably”?

    Provisional answer  The central teaching of Black/Gowers/Regan/Lipton-style STEAM optimism is that we should all hope to live in a world that cherishes diverse opinions in regard to the P-BosonSampling Postulate, sufficient to induce some researchers to conceive of 21st century Black-style innovations, sufficient to support 21st century Bell-scale enterprises.

    Conclusion  Scott Aaronson and Alex Arkhipov richly deserve all of our appreciation and thanks for providing — crucially! — precisely half of the creativity and commitment that are necessary to foster the Black/Gowers/Regan/Lipton variety of 21st century STEAM dynamics in the context of BosonSampling.

    ——–

    @article{Black:1977fk, Author = {Harold S. Black}, Journal = {IEEE Spectrum (USA)}, Month ={December}, Number = {12}, Pages = {54--60}, Title= {Inventing the Negative Feedback Amplifier},Volume = {14}, Year = {1977}}

  17. Scott Says:

    Michael P #15:

      IMHO, “publish or perish” in pre-tenure career is not that much different from “develop something useful or get fired” in industry.

    One of the great ironies of this closure is that a single Microsoft SVC researcher, Sergey Yekhanin, apparently developed something for Microsoft’s business (related to locally decodable codes) that alone would justify the entire cost of the lab. I didn’t discuss this in the OP only because I don’t know the details—but anyone who does is welcome to share them here.

  18. bjuba Says:

    Re: Scott’s comment above.

    In slightly more detail, Sergey’s work was on systematic locally reconstructable codes. You can think of this as an improvement of the parity-check code used in RAID — when there are no errors, you can just read the data off directly from the code, and when disks fail, you can reconstruct the lost disk efficiently by only looking at a few of the other disks. It apparently went into production with Windows Azure:

    C. Huang, H. Simitci, Y. Xu, A. Ogus, B. Calder, P. Gopalan, J. Li, S. Yekhanin “Erasure coding in Windows Azure Storage” Proc. USENIX ATC, 2012.

    As has been noted by many other bloggers, you can draw your own conclusions about the message that the leadership at Microsoft is sending.

  19. Chris W. Says:

    Tony #9: Yes, indeed. The solution proffered by the ownership class is “become one of us”.

    In that connection, see this 2013 interview of Charles Murray by Kai Ryssdal of NPR’s Marketplace. Murray’s views, combined with his position at AEI, say a lot about how we got here. Here is one commenter’s response:

    Imagine my reaction to hearing Dr Murray bloviating from his pedestal at the American Enterprise Institute about lazy Americans as I sped home from my day job to change my clothes for my night job. As a 60 year old who was laid off for being too old at a Fortune 500 company, I now have a day job, a part time night job and freelance. The kids I work with at my day job bust their asses, and I dare say that if I were take a dozen of the navel-gazers at the AEI and put them to work at my day job, they couldn’t get through a shift. Keep pontificating Dr Murray – it shows how out of touch you are about real Americans….

    Sorry Scott, I’m drifting off-topic…

  20. Rahul Says:

    I don’t know. Sure its sad for the people involved. I’d have been pissed off if it was me being laid off too. And I agree about the undesired abruptness. The shutdown could’ve been done more gradually.

    But I don’t see the underlying criticism about closing the lab. Can’t a private company ever close a lab or a division? What if management felt one lab or one location or one area of research was over-represented, under-performing, mis-aligned with their long term goals etc.? Maybe the division’s employees weren’t doing enough to “turn ideas into reality”?

    Would this be a generic criticism of closing any research division by any corporate? Further, it’s puzzling that Scott’s post doesn’t include the larger context of approximately *18,000* employees world-wide that Microsoft announced laying off starting July this year. Should MS-Research guys be singled out for special treatment? Insulated from harsh realities of money & budgets & priorities?

    If Mr. Gates did chip in with an exclusive money airdrop, how’d it be fair to the other 16,000 MS employees also being laid off worldwide? Or should he chip in to rescue all of them?

    I don’t buy that this will bring “severe damage to MSR’s painstakingly-built reputation”. In any case, MS has more to think about than just MSR. Maybe the shareholders will like some deadwood being pruned?

    Besides the press note says a part of the laid of researchers are being absorbed at other MS locations. Is that true? How many?

    Yes, people rarely get sacked from academia. Although at the margin I’m not sure if that’s a bug or a feature. And maybe yes it is a feature, but that’s no reason why the whole world has to work on this “no-layoffs-ever” model.

    As a lobbyst for TCS, this would be an excellent writeup, but if one wants broader, non-biased reporting I’m not sure this conveys a good or fair viewpoint.

  21. Scott Says:

    Rahul #20: That’s an absurd straw-man. Not one of the commenters here, or on any of the other blog posts about this, ever suggested that “no private company can ever close a lab or division.” All of the criticisms have been specific to this particular shutdown: to

    (1) its abruptness (showing a callous attitude toward the academic job cycle),
    (2) the fact that the other MSRs are now irreparably damaged in their ability to hire and retain great researchers (this is not speculation—I know of people at the other MSRs who are now seeking to leave),
    (3) the fact that SVC wasn’t in any sense “underperforming” or “deadwood,” but had been one of the best CS research labs in the country, and
    (4) the fact—discussed in comments #17 and #18—that a theorist at MSR SVC, Sergey Yekhanin, made a major contribution to Windows Azure, which alone could have justified the entire cost of the lab.

    As for some of the MSR SVC researchers being offered jobs elsewhere in Microsoft, the latest intel I’ve heard is that it’s just a few of them, maybe 10% (the Turing Award winners, Cynthia Dwork, Sergey Yekhanin, maybe one or two others). The other intel I’ve heard is that the choice to end SVC had more to do with internal Microsoft politics (which I don’t know the details of) than with the actual good of the company.

  22. Sid K Says:

    Scott #17:

    Sergey Yekhanin, apparently developed something for Microsoft’s business (related to locally decodable codes) that alone would justify the entire cost of the lab.

    In this vein, I recently came across a nice example: DART, a logistics planning AI tool developed for the US military. It supposedly saved the US military more money in the 1990-91 Iraq war than all the money that DARPA had invested in AI research in the previous 30 years.

  23. Chris W. Says:

    On Scott’s last comment (#21) it should also be noted that Windows Azure is one of the strongest businesses Microsoft has right now, which is widely considered to be an important reason why the current CEO got his job.

    Of course it wouldn’t be the first time that company politics resulted in an obtuse and shortsighted decision.

  24. Rahul Says:

    Scott #21

    Ok, I think I misunderstood your original post then. Sorry. Your #1 I agree with (like I said in my earlier comment too)

    To clarify, would you be ok had MSR closed another one of their locations? Which one?

    Or are you against MSR doing any layoffs? Or are you saying layoffs should have been dispersed & not location specific?

    Like you described, Sergey Yekhanin did great work & indeed they are retaining him. Or trying to. But that may not by itself be a great reason to retain everyone else at the same location.

    I don’t know. All I’m saying is presumable MSR top bosses had the information to judge. Are we qualified to second guess them? If we want to, I’d like to hear names that specifically were asked to go (i.e. not relocate) and we think were fantastic performers.

    Indeed company politics can lead to obtuse decisions. I’m trying to understand the details here of why this particular set of layoffs is obtuse. You don’t have to wait for a crisis before you start laying off.

  25. Darrell Burgan Says:

    It’s sad whenever anyone loses their job without cause, but that is basic reality in business. People lose their jobs all the time for all sorts of reasons having nothing to do with their capabilities or performance. In technology, for example, such risk is part of the game, and nearly all of us in it have experienced it at one time or another.

    I’m unclear why researchers should somehow be exempted from this unpleasant reality merely because they are researchers. Isn’t that the purpose of academic institutions and tenure, or am I being naive?

  26. krishna Says:

    As a society everyone benefited by the researchers. Though researchers do research for fun but somehow it always benefits the society in many ways. Even operating system was once a theory now a money making product. If intellectuals are not respected and nurtured and safe guarded the society will be deprived of many beautiful things. To pay 75 researchers doesn’t cost that much money and it’s actually healthy for them and for US as a whole to keep researchers engaged in research. I am sure there are many things that microsoft could have done to trim the cost like reducing the pay for top level management and I always wonder whether top level management deserve that much money. The management always think scientists output are not tangible. Why bill gates did not intervene in this decision? As gates knows a little bit about research as he published a paper with legend chris papadimitriou. As long as company like microsoft makes billions of dollars in profit there is no necessity for them to trim of the scientist they can answer the stack holders in many ways. Rahul said “You don’t have to wait for a crisis before you start laying off.” Why you need to lay off when there is no crisis?. Moreover rahul can check the papers published by those 75 scientists online. Professor lke madhu sudan left MIT to join MSR new england.

  27. Scott Says:

    Rahul #24: The people sent packing include Omer Reingold, Frank McSherry, Raghu Meka, Guy Rothblum, Alex Andoni, Andrew Goldberg, and at least a dozen others who I’m forgetting right now just within theory (leaving aside the rest of CS). This is not a list of “underperformers.” It’s an all-star list.

    So the feeling I wanted to evoke is not one of pity for these poor unemployed people. They’ll all land on their feet. Indeed, one can be sure that other institutions are even now eyeing them hungrily—though so many great people on the market at once will create a glut unlike anything I’ve seen (others who’ve been around longer have seen such things). And yes, they knew the risk going in (indeed, they may have turned down the stability of academia for the larger salaries of industry), and yes, their “plight” is shared by millions of other people in the modern world.

    The feeling I wanted to evoke, rather, was one of sadness that such a fine research lab, which took so long to build, was destroyed for so little reason (with large “collateral damage” to the rest of MSR, and thus, indirectly, to computer science).

    To me, keeping the one guy in your basic research lab who did something that alone justified the lab’s cost, while firing everyone else, reflects a profound misunderstanding of how science works (it’s hard to imagine a profounder one). Science is a probabilistic process. It would be like a beach town keeping the one lifeguard who actually saved someone’s life, while firing all the others who were off-duty or elsewhere at the time. No doubt your one remaining lifeguard displayed “heroic competence” (to borrow a phrase from The Simpsons), but she can’t guard the whole beach 24/7.

    You suggest that we’re unqualified to second-guess the Microsoft managers who made the decision. But what about people at the remaining MSRs, who know Microsoft’s internal politics, and are vocal (in private) about the decision’s stupidity and shortsightedness? I don’t feel qualified to second-guess them.

    There’s no question of “rights” here. Obviously, Microsoft execs have the right to close a lab for whatever reason they want, the CS community has the right to complain about it on CS blogs, the execs have the right to ignore the complaints, etc. etc. On the other hand, to whatever extent large companies invest in basic research for reasons of image (not only for the public, but for their pool of potential recruits), one can hope that publicly shaming them, when they pull something like this, will slightly increase the cost to them of doing it again.

  28. Vadim Says:

    The main difference between researchers and everyone else, is that everyone else is expected to improve their employers’ bottom lines, either by bringing in revenue, controlling costs, or providing support services that make those functions possible. With researchers doing basic research, the connection to the bottom line is much more tenuous, which is why very few companies fund basic research and most such research is done at universities. For those companies that consider the uncertain benefit to be worthwhile and choose to fund it anyway, it makes sense to at least somewhat follow the norms of the “research industry”, otherwise they’ll only be able to attract researchers that have no other options for employment. So Microsoft’s abrupt closing of the lab wasn’t wrong from some lofty moral principle point of view; it hurts them from the free market perspective, the free market having different norms for different classes of workers.

  29. Dave R Says:

    I read that many of the researchers would be offered positions in other MSR labs… For the rest, I’m sure they already have offers somewhere else – how hard can it be for brilliant CS PhD’s to get job offers in Silicon Valley?

  30. John Sidles Says:

    Rahul excuses “MS has more to think about than just MSR. Maybe the shareholders will like some deadwood being pruned? […] I’m saying is presumably MSR top bosses had the information to judge.”

    Darrell Burgan sighs  “It’s sad whenever anyone loses their  job  family farm without cause, but that is basic reality in business.”

    Shtetl Optimized readers who prefer simple economic axioms that imply simplistic social conclusions — libertarian market fundamentalists for example — will sympathize with Rahul’s and Burgan’s remarks.

    Still it is well for STEAM students to reflect too upon diverse alternative viewpoints. These include:

    (1)  Essayist Wendell Berry’s 2012 Jefferson Lecture It all turns on affection, which can be appreciated in light of the following natural isomorphisms:

    \[\begin{array}{rcl}\langle\text{STEAM careers}\rangle&\simeq&\langle\text{family farms}\rangle\\ \langle\text{Microsoft}\rangle&\simeq&\langle\text{American Tobacco Company}\rangle\end{array}\]

    (2)  Artist Barry Deutsch’s celebrated cartoon “The 24 Types of Libertarian” (which a Google search will find), and

    (3)  Mathematician Alexander Grothendieck’s appreciation (as surveyed in Allyn Jackson’s two-part Notices of the AMS ReviewCommé appele du néant — as if summoned from the void: the life of Alexander Grothendieck”, 2004) that:

    “Our thirst for knowledge and discovery indulges itself more and more in a logical delirium far removed from life, while life itself is going to Hell in a thousand ways — and is under the threat of final extermination. High time to change our course!”

    Conclusion  If Grothendieck is well-justified (as seems likely to me) in reminding the STEAM community that it is “high time to change our course,” then it is plainly evident too that the governing boards of the world’s great corporation-universities, and their libertarian market-fundamentalist allies, can be relied upon neither to chart those course-changes nor initiate them.

  31. Scott Says:

    Dave R #29: See the previous comments. It seems that strikingly few of the MSR SVC researchers—even superb people you’d think Microsoft would be dying to keep—were offered other jobs at Microsoft. But in any case, I don’t think the “plight” of the fired researchers is the main issue here (they’ll find other jobs); rather, the issue is the loss to MSR and to CS research as a whole.

  32. Dave R Says:

    Well, how could it be a loss to CS research if someone else is going to hire them? Unless no one will hire them to research what they wanted to (which would be unlikely unless what they were doing was completely invaluable).

  33. Scott Says:

    Dave R #32: Econ 101. Yes, they’ll probably get hired somewhere else (the ones who don’t decide to leave the field), but that will make fewer slots available for other people, etc.

  34. Dave R Says:

    Scott –

    “that will make fewer slots available for other people, etc”

    not necessarily. There isn’t a fixed lump of labor. They might just have to start doing things the economy finds more valuable.

    Re your lifeguard analogy, maybe its more like a few of them were off building sandcastles (great ones, albeit). Point is that if they were doing something clearly valuable for MS, then MS would have kept them. If they weren’t, MS was right to let them go.

  35. Michael P Says:

    Dave #32:
    Sometimes research gets seriously hindered by a move from one position to another. I’m not sure whether this is applicable to theory, but in other subjects one often needs to spend months after the move developing tools that has been left in the previous company before any further work can actually be done. Also, sometimes one cannot just continue from the point one left off in another company, even if all the tools were available, because of IP issues.

  36. Rahul Says:

    Even though a researcher may be strikingly good your organization may not always want to be doing what he is good at.

    Further even though a researcher is good and you want to be doing what he is good at you may already have so many people working on it that you may not want any more.

    Vladim #28 mentions “researchers doing basic research” but it’s intriguing that MSR has the tagline “Turning Ideas into Reality”. That sounds like a more applied goal. What gives?

  37. Peter Nelson Says:

    Dave #34:

    …if they were doing something clearly valuable for MS, then MS would have kept them. If they weren’t, MS was right to let them go

    By that logic, no firing could possibly be subject to criticism. Arguments have been made throughout this thread that they were valuable to MS and that MS made a mistake firing them.

  38. jonas Says:

    Rahul: in “Do theoretical computer scientists despise practitioners?” http://www.scottaaronson.com/blog/?p=1981 , Scott points out that people in pure research have to pretend that their work is applied (and vice versa) to get grants from the government. The tagline is usually for marketing, so it could easily be part of such a pretense.

  39. Dave R Says:

    Peter Nelson: I guess it is just not clear to me what the hypothesis is then on why they were fired. Does Microsoft not know how to do probabilistic NPV calculations? Or they just got the valuations or the probabilities wrong?

    It seems to me people are somehow implying that Microsoft fired these people knowing that they were valuable but they just didn’t care. That is the argument that I am rejecting.

  40. Scott Says:

    Dave #39: Why did Apple force out Steve Jobs? Why did IBM agree to Microsoft’s licensing terms? Why did Coke introduce New Coke? Did these corporations not know how to perform probabilistic expected-value calculations? Did they—gasp—get the probabilities or the valuations wrong? Did they “know” they were doing something stupid and simply not care? Sure, those are all hypotheses worth considering. Or … maybe corporations shouldn’t even be thought of as value-maximizing rational actors at all. Maybe they’re huge organizations with turf wars and rivaling factions and unthinkingly-implemented bureaucratic policies and quotas and people handing off blame and responsibility to others—thereby allowing for the emergence of the well-known phenomenon of collective stupidity.

  41. Douglas Knight Says:

    Dave R, MS isn’t a person. It doesn’t know anything; it doesn’t care about the value of divisions. When people say that this is a “political” decision, they mean that some individuals decided that it increased the individuals’ career value at the expense of the company. Maybe you think that MS is well-governed and that this is implausible, but if you think MS is so well-governed that you can reason about it as an agent, you’re wrong.

  42. Rahul Says:

    Scott #40:

    Yeah, but had you been around back then would you have taken the right decisions on all those issues?

    Yes corporations cannot always decide wisely nor optimally but it isn’t clear to me that there exists someone else outside who might always decide better.

    I think a lot of people underestimate how hard it is to decide for a large corporation. If one just wants to lobby for a field, say TCS, that’s OK, because one could take the stand my field right or wrong. i.e. all layoffs affecting good TCS researchers = bad and worthy of criticism and protest.

    OTOH, if one is to be taking MS-positive decisions I don’t think that’s easy at all. Because then we have to take those messy hard decisions: How many should we lay off? Where all should those layoffs come from?

    Now those questions I don’t see answered in this thread and indeed those are the hard questions. If not to layoff TCS guys at this MSR location then who? That is the hard question. “Don’t layoff any TCS guys” is the easy answer but I’m not sure that was very practical.

  43. Dave R Says:

    Scott, Douglas: I think you are not giving MS enough credit. Just because the decision was announced hastily doesn’t mean it was a hasty decision. Working at a company of similar size, I can tell you that every decision is modeled out on 17 Excel tabs these days (what else do MBA’s do with their time?). In this case, I am sure it was done on an individual by individual basis, and MS decided that it was only profitable to employ a few of these (albeit very brilliant) people. (Evidenced by the fact that they seemed to keep a few who had contributed substantially to the bottom line).

    If you feel that their model assumptions were somehow flawed – that they were monetizing the future value of differential security or whatever incorrectly – then I think those are points that should be debated (here or elsewhere). But just claiming that MS is a bunch of would-be state senators or whatever who make irrational, selfish decisions is degrading a lot of other bright people at MS whom we don’t happen to know.

  44. lylebot Says:

    it seems essential that our civilization maintain institutions for which the pursuit and dissemination of knowledge are not just accoutrements for when financial times are good and the Board of Directors is sympathetic, but are the institution’s entire reasons for being: those activities that the institution has explicitly committed to support for as long as it exists.

    I dunno, I sort of think the reason universities have been able to support the pursuit and dissemination of knowledge over the last 50+ years is because of uncommonly good financial times and a sympathetic “Board of Directors”–i.e., Congress, particularly during the Cold War.

    (Yes, I know universities did it before the post-war boom too, but it was very different. Not nearly as many people had the opportunity to contribute to that knowledge, for one thing.)

    Call me cynical, but it appears to me that institutions like Harvard are becoming investment banks with a university as a PR front, while “lesser” institutions are seeing federal funding dry up and faculties shrink. Congress is no longer sympathetic. I think we’re at the tail end of the golden age of universities, and it won’t be long before we’re seeing programs, departments, colleges, and universities shutting down on a regular basis.

  45. Scott Says:

    Dave #43: Thanks—that really makes me see things differently! I don’t know why it didn’t occur to me that, by defending the lab, I was hurting the feelings of the MBAs in Redmond who must’ve worked very hard to prepare Excel sheets justifying why the lab should be closed, and almost all its researchers fired without warning. If any of those MBAs are reading this, I hope they can find it in their hearts to forgive me. Like they say, until you’ve walked a mile in someone’s black loafers…

  46. Florin Moldoveanu Says:

    I don’t get the fuss over the layoffs in a commercial company. This is how the commercial world works. Brutal and effective. Unpleasant, painful? Certainly. But this is noting out of the ordinary.

  47. Scott Says:

    lylebot #44:

      I think we’re at the tail end of the golden age of universities, and it won’t be long before we’re seeing programs, departments, colleges, and universities shutting down on a regular basis.

    I don’t know that you’re right, but if you are, then all the more reason to speak out for the necessity of institutions of higher learning that are sheltered from short-term market forces.

  48. Scott Says:

    Florin #46: The thing that makes it non-ordinary (with just a few historical precedents, like the destruction of the old Bell Labs) is that for 20 years, MSR had behaved like a gigantic academic CS department that happened to be run by Microsoft. That’s the reason why they were able to recruit so many first-rate academics to work for them. So the peremptory closing of SVR is a wake-up call showing that MSR can no longer be thought of that way. If you don’t care about CS research, it’s indeed not a big deal, but if you do, it is.

  49. John Sidles Says:

    To combine several recent Shtetl Optimized themes:

    Steven Pinker asserts [“brilliantly” per Scott’s evaluation] that  “Test scores, as far up the upper tail as you can go, predict a vast range of intellectual, practical, and artistic accomplishments.”

    Fans of Frans de Waal’s works (beginning with Chimpanzee Politics: Power and Sex Among Apes, 1982) will reflect that Pinker’s assertion seemingly predicts that high-scoring organizations — elite universities, elite corporations, and elite research laboratories for example — will exhibit a vastly diminished range of unreflectively ape-like and/or selfishly short-sighted and/or willfully ignorant political machinations.

    Yet there’s not much evidence of that, is there Shtetl Optimized readers?

    Young STEAM students (especially) commonly assume that ascending the ladder of professional and/or corporate and/or political status eventually leads to a utopian realm in which wise foresight and dispassionate judgment is normative behavior. And surely the hypothesis that marketplaces are rational … if it has any fundamental validity … will predict that this expectation will be fulfilled?

    Common sense  It is most people’s life-experience (including mine) and the plain lesson too of history both natural and recorded, that the apish ranges of humanity’s status-ladders extend a lot higher than humanity’s friends would wish.

    Conclusion  Whatever capability it is that test-scores measure and markets optimize, we hominids presently use this capability as much to justify and facilitate our apish impulses, as to inhibit them.

    Corollary  It is comparably plausible that Microsoft SVC was terminated because its computer-science scholars were asking suboptimal questions, as that Microsoft executives were suboptimally valuing their answers.

    Summary  Could we all be asking better STEAM questions, with a conscious view toward arriving at better answers?

    The world wonders. In particular Neil Stephensen has been wondering. Good!

  50. Florin Moldoveanu Says:

    Scott, I have experience in both worlds: academy and industry. Both have good and bad sides. Industry has much less politics than academia, greater mobility and opportunity, better pay, but it is brutal and cares only about the bottom line. Trust me when I say that in an executive’s mind, the consideration from your post were not even entertained for a split second. This does not make them callous or careless, this is how industry simply is. Now if a competitor would hire those people and in the process gain a competitive advantage, then this would be a big deal and cause to fire on the spot the person who decided the cut. But they must have done their due diligence and concluded that this is not a real scenario.

  51. Tony Says:

    #46

    Brutal and effective?
    No one who ever worked in big commercial company would ever say that.

  52. Darrell Burgan Says:

    John #30, you said:

    Shtetl Optimized readers who prefer simple economic axioms that imply simplistic social conclusions — libertarian market fundamentalists for example — will sympathize with Rahul’s and Burgan’s remarks.

    John, I’m not asking anyone to sympathize with me nor am I defending Microsoft. As someone who has personally experienced losing one’s job for no other reason than a change in executive direction, I think I have a right to question why a researcher should be exempted from the same unfair reality that I am not exempted from.

    If your point is that it is unfair, I will agree with you (although I’ll question what can be done about it). I disagree with you, however, that researchers should occupy a privileged position in the employment equation. We all face the same risks, together.

  53. Rahul Says:

    Scott #48 says:

    “If you don’t care about CS research, it’s [ closing of SVR MSR] indeed not a big deal, but if you do, it is.”

    I think Scott’s quote is at the crux of the divergent sets of comments. I’d rephrase a bit: “If you only care about TCS research the current set of layoffs are a big deal, definitely sad and worthy of criticism”

    OTOH, if you are willing to think broader, and include the interests of Microsoft as a company, the returns to its shareholders, or the interests of the non-TCS guys or even the Microsoft-products consuming public then it is not such a black-n-white picture. Then the MSR-SVR layoffs may yet be bad, or they may be good, or they may be essentially irrelevant but it’s not easy to pass a quick, critical judgement.

    One might also want to read market analysts writing about the layoffs. They sure don’t sound as pessimistic as Scott. Another thing to watch in the longer run will be MSFT stock prices. In the short run shareholders seem to have responded favorably to the layoffs but these signals can be fickle.

    Finally, when a researcher joins an industrial lab rather than academia he’s making a conscious tradeoff: higher salaries, closer to application, more resources, chance to jump ship etc. but with the downsides of lack of security & independence. To wish that MSR should have worked within the conventions of the traditional university model is not useful. If that happened in the end what difference would there be between the two models? I think there’s a place for both systems to coexist.

  54. Scott Says:

    Rahul #53:

      OTOH, if you are willing to think broader, and include the interests of Microsoft as a company, the returns to its shareholders, …

    I find it amusing that, in your terminology, thinking about the future of CS research (or, one could say: Microsoft’s future capacity to innovate, the future of Microsoft’s relationship with the CS community, the future of industry-supported basic research as a whole…) means thinking “narrowly,” whereas considering the returns to Microsoft’s shareholders right now means “thinking broader.” 😉

  55. Rahul Says:

    Scott #54:

    What’s the total number of people using Microsoft Products or holding their shares?

    OTOH what’s the total number of people invested into Theoretical Computer Science Research by virtue of their jobs, primary interests etc.?

    I’d like to think the former cohort was indeed “broader” than the latter. I could be wrong.

    PS. I don’t think Microsoft’s future capacity to innovate is very strongly correlated with the TCS Research at MSR or elsewhere. Again, I could be mistaken.

  56. John Sidles Says:

    In reference to adamantly market-fundamentalist narratives regarding the Microsoft SVC closure (e.g.. #50-#53 above), and with further reference to Wendell Berry’s Jefferson Lecture discussion of Howards End (per comment #22 above)

    Howards End
    Chapter 22

    There was one quality in Henry for which she [Margaret] was never prepared, however much she reminded herself of it: his obtuseness.

    He simply did not notice things, and there was no more to be said.

    Henry  “My motto is ‘Concentrate’. I’ve no intention of frittering away my strength.”

    Margaret  “It isn’t frittering away the strength; it’s enlarging the space in which you may be strong.”

    Henry  “You’re a clever little woman, but my motto’s ‘Concentrate’.”

    Summary It’s reasonable to regard Microsoft executives and SCV researchers as jointly complicit in obtuse concentration, in regard to the 21st century’s rainbow bridge opportunities in STEAM enterprises.

  57. John Sidles Says:

    Rahul discerns valuation in market-price “In the short run shareholders seem to have responded favorably to the [Microsoft] layoffs but these signals can be fickle.”

    The text of Wendell Berry’s 2012 Jefferson Lecture severely criticizes market-valuation:

    In this age so abstracted and bewildered by technological magnifications of power, people who stray beyond the limits of their mental competence typically find no guide except for the supposed authority of market price.

    ‘The market’ thus assumes the standing of ultimate reality.

    But market value is an illusion, as is proven by its frequent changes; it is determined solely by the buyer’s ability and willingness to pay.

    By now our immense destructiveness has made clear that the actual value of some things exceeds human ability to calculate or measure, and therefore must be considered absolute.

    For the destruction of these things there is never, under any circumstances, any justification.

    Summary  The 21st century’s creative destruction of hope, by the agency of unrestrained globalized market-valuation, has begun to “exceed human ability to calculate or measure.”

    Recommendation  “The end already is present in the means” is a friendly maxim that properly directs our concentration away from market-fundamentalist ideologies.

  58. fred Says:

    Among all the patents filed by Microsoft, how many actually originated from that SVR?

  59. rrtucci Says:

    Is MS going to close more? Come on MS, I bet the cost of living in the Boston area is almost as high as in San Fran.

  60. Scott Says:

    John Sidles #56: I assume I’m not going to get a comprehensible answer, but why were the SVC researchers “complicit in obtuse concentration,” whatever that means? (Does it mean: complicit in their own firing?) Do you know anything about them or their research, that would justify such a statement?

  61. Scott Says:

    Rahul #55: Not for the first time I ask, why do you come here? There’s a whole giant Internet out there for people who don’t place an overriding value on narrrow, parochial concerns like theoretical computer science, the advancement of human knowledge, or the eradication of ignorance. Go explore it! 🙂

  62. Vadim Says:

    Scott #61, I don’t think your suggestion to Rahul is very fair. I’m guessing that most people who read this blog (except maybe Lubos) have some interest in TCS and would like to see the field advance. On the other hand, to what extent people balance this vs. other factors – say, the financial outlook of the company that’s sponsoring the research, which impact things like retirement accounts invested in the company, thousands of jobs, etc – is going to vary, and unsurprisingly, TCS researchers such as yourself may find themselves leaning more heavily to the TCS-above-all-else side of things than others might. For what it’s worth, I agree more with your view, but I don’t think you want this blog turning into an echo chamber, and Rahul (while having strong opinions) always expresses himself reasonably and politely. Are you perhaps taking this more personally because you know some of the people who were impacted?

  63. Rahul Says:

    Scott #61:

    Because narrow concerns can be great fun too! 🙂 Hey I post on sailing forums and that’s probably more parochial than this.

    But jokes apart, I don’t think Shtetl is that narrow, is it. I’ve seen you post tons outside of TCS. You do a good job really. And I come here because you are smart & write well and attract tons of good comments. Doesn’t mean I always agree with you (fine, sometimes I do!)

    In other words narrow & parochial is perfectly fine so long as one realizes it is narrow & parochial! And the fact that “the advancement of human knowledge, or the eradication of ignorance” is a massively large superset of TCS.

  64. Scott Says:

    Vadim #62: OK, point taken. I didn’t mean to suggest Rahul was unwelcome here; I was just expressing mild bafflement and amusement. He reminds me of a guy who goes to Red Lobster, then spends the whole evening complaining about an overemphasis on shellfish.

  65. vzn Says:

    some reflections & collection of a lot of blog writing on the topic partly spurred by this dialog. microsoft silicon valley TCS research lab shuts down— easy come, easy go

  66. anon Says:

    A couple of points:

    1) The decision to fire MSR SVC was not taken by a bunch of MBA’s. It was taken by Harry Shum (who is the head of MSR and used to be a graphics researcher) and his team.

    2) It is not a secret that recent MS products have not done as well as they hoped: windows 8, windows phone, surface tablets, etc. MSR was given a target to cut about 5% of their workforce.

    3) While several options come to mind (such as spreading the layoffs evenly over all labs), all of them are painful and would have hurt the reputation of MSR. The disadvantage with spreading layoff across all labs/groups is that many employees which would be retained would lose their close collaborators.

    4) Indeed layoffs are sudden and out of sync with academic cycle. However people are getting severance package to support themselves financially. The hard reality is that at the end of the day, MSR is a part of MS which has a certain goal of workforce reduction in the next few months.

    So any thoughts of what *should* have been done instead? (besides the obvious answer that MSR should have been unaffected by layoffs)

  67. anon Says:

    With decline of the dominance of MS, we in the CS research community can only hope that other companies like Google and Apple (which have now gained the required size and stability) will start supporting basic computer science research. BTW MSR management has clarified that there will be NO further layoffs.

  68. Tony Says:

    #50

    Politics is everywhere in the industry.
    In fact, it is mostly politics.
    The executives are callous and careless, probably never more than nowadays.
    The industry is in Dark Ages, meritocracy is dead for all practical purposes.

  69. anon 2 Says:

    anon #67: At this point, everyone understands that “MSR management has clarified that there will be NO further layoffs” means “no further layoffs within the next 6-12 months”.

    Folks who have TCS Ph.D.s and are at the level of full-time MSR researcher are grown-ups (mostly). This means that people who have other options will go elsewhere, and MSR will get the people who don’t have any other option. That is not the kind of organization you want to build, whether it’s a research lab or a factory floor.

  70. Rahul Says:

    Scott #64:

    For the record, I love Red Lobster; and shellfish are awesome. 🙂

  71. John Sidles Says:

    Scott asks “John Sidles #56: I assume I’m not going to get a comprehensible answer, but

    [Q1]  why were the SVC researchers “complicit in obtuse concentration,” whatever that means?

    [Q2]  Does it mean: complicit in their own firing?

    [Q3]  Do you know anything about them or their research, that would justify such a statement?”

    Scott, please let me offer weak answers to your questions … weak in that the answers themselves are flawed by “obtuse concentration.”

    A1  “Obtuse concentration” is the cognitive trait of concentrating obtusely upon narrow questions that have convenient answers, thereby failing to “enlarge the space in which you may be strong”.

    This answer borrows from Margaret Schlegel-Wilcox’s memorable discourse in Forster’s Howards End (per comment #22 above);

    A2  No.

    A3  Yes, but only in that everyone knows it (or should).

    In particular, Shtetl Optimized readers are invited to verify for themselves that the Microsoft Corporation’s market capitalization (in constant dollars) peaked in 1998. The subsequent sixteen years of capital stagnation has been bad news for Microsoft Research Labs, whose employees are young, in which retirements are few, whose vitality depends in part upon ongoing recruitment of talented young researchers.

    To put it plainly, stagnating businesses cannot sustain vigorous research laboratories, any more than stagnating national and global economies can sustain vigorous research universities.

    In striking contrast, numerous computation-centric enterprises have exhibited exponential growth in this same time-period (see for example the market capitalization of NASDAQ: ANSS [Ansys Corporation])

    Obtusely concentrated summary  Microsoft SVC researchers and Microsoft executives were jointly complicit — and equally derelict in duty-to-shareholders — in not asking “Why isn’t Microsoft’s research strategy-and-development strategy more viably like Ansys’?”

    Broadly illuminating summary  Better/broader answers than market-valuation can supply are illuminated (as it seems to me) by the friendly comments of Tim Gowers/Dick Lipton/Ken Regan/Alexander Grothendieck/Neil Stephenson/Wendell Berry, as quoted in comments #16, #30, #49, and #57 above.

    There’s more  Much more can be said upon the general topic of STEAM-hope, and needless to say, it is neither necessary, nor feasible, nor desirable, that everyone agree in these matters. Perhaps we can all agree with John von Neumann (1955) though:

    “The most hopeful answer is that the human species has been subjected to similar tests before, and seems to have a congenital ability to come through, after varying amounts of trouble.”

    Here von Neumann is borrowing from then-SecD Charles Erwin “Engine Charlie” Wilson: “The price of progress is trouble, and I must be making lots of progress.”

    Postscript  Perhaps I had better say that my wife (Constance Sidles) has begun writing a book upon the topic of “hope”.

    Especially on our hot crowded 21st century planet, Connie’s essays crucially grapple with STEAM-hope, regarding which there is *far* more to say than “obtusely concentrated” NASDAQ market-measures (and oracle-grounded complexity zoos) can can ever encompass.

  72. fred Says:

    Seems pretty relevant to the discussion:

    http://arstechnica.com/science/2014/09/are-we-producing-too-few-or-too-many-science-and-technology-grads/

    We seem to be producing researchers fitted for academic setting but ill-prepared for commercial context.

  73. Rahul Says:

    bjuba #18:

    “Sergey’s work was on systematic locally reconstructable codes. You can think of this as an improvement of the parity-check code used in RAID — when there are no errors, you can just read the data off directly from the code, and when disks fail, you can reconstruct the lost disk efficiently by only looking at a few of the other disks.”

    Moving away from the layoffs-sage, does the Sergey-strategy circumvent the common worry about using RAID5 or RAID6 with large drive sizes? Say above 4 TB or so, the probability of dual / triple errors especially during reconstruction in the aftermath of a failed drive?

    Or the days or week long reconstruction time with large volumes?

    I’d love a blog post or more comments on Sergey’s work!

  74. David Says:

    The following is not my explanation, but I find it convincing:

    Corporates perform research which they think will be useful for them for the period they see as “their future”. Moreover, the closer the corporate is to be a monopoly, the better it is positioned to have a monopoly on the applications of the research; thus the more beneficial it is for the corporate to be open about this research.

    In the 50s and 60s Bell and IBM saw their future as infinite, and where monopolies, so having huge and well funded labs made sense (they where right by the way – the labs output only increased the barriers to entry for new players – eventually the US DOJ had to deal with both of them). When MS research was founded, Gates thought the same; he was wrong (MS was not a monopoly long enough to take advantage of the research arm’s output), and they are no longer a monopoly.

  75. gentzen Says:

    What can be really scary about such lay-offs is how hard it can be to check that they actually happened. I work in the semiconductor manufacturing industry, and heard in June 2014 the following rumors regarding “real” lay-offs:

    In the multi-beam direct-write segment, multiple sources indicate that KLA-Tencor is exiting this market. Officials from KLA-Tencor declined to comment.

    […]

    KLA-Tencor, meanwhile, has been developing Reflective Electron Beam Lithography (REBL), which was originally funded by DARPA. In fact, DARPA poured over $100 million in funding into the REBL program.

    KLA-Tencor was able to develop an alpha tool, based on the REBL concept. At the recent SPIE conference, the company reported some impressive results. The key to the technology is a CMOS-based digital pattern generator module, which enables more than 1 million beams at full current. Reports of its exit from this market leave its development efforts in question.

    I actually have business cards from people that work(ed) there, but would you want to send them an email asking: “I heard your division will be closed and all of you are laid-off. Can you confirm this?” Maybe next time I go to a conference, I will try to find out what really happened…

  76. gentzen Says:

    krishna #26

    Do you have any idea how big a staff of 75 is? I worked with IMS Nanofabrication, one of the remaining players in the multi-beam market, and they had significantly fewer than 75 employees total (including all support staff), and will probably succeed in bringing their new technology to market. How will they do this? They know their own strengths, and collaborate with experienced partners for the parts where they are lacking experience and workforce.

    Known big players like ZEISS Semiconductor Manufacturing Technology might have more employees, but they also have a much bigger technology and product portfolio to manage. Any single product and technology of them will probably have significantly fewer than 75 employees working on it. This works, because they too collaborate with other groups, both internal and external (ASML…)

  77. bjuba Says:

    Rahul: depending on how liberally you interpret “RAID,” it may (1) reduce the number of disks involved during reconstruction, (2) tolerate more failures and/or (if you’re already considering previously existing state-of-the-art error-correcting codes to fall under the umbrella of RAID) (3) achieve 1 & 2 while allowing the normal-case data access to be just a single read from a disk.

    If you’re interested in the concrete quantitative improvements in overhead, speed, and/or error tolerance, I’d advise you to have a look at the USENIX ATC paper. Section 5 is a reasonably complete comparison with the Reed-Solomon codes that were previously the code of choice in distributed file systems.

  78. John Sidles Says:

    gentzen remarks “What can be really scary about such lay-offs [as Microsoft’s] is how hard it can be to check that they actually happened.”

    Cringely’s latest book The Decline and Fall of IBM: End of an American Icon? is a thought-provoking analysis of the mechanisms involved … whose conclusions will surprise precisely zero IBM-research insiders.

    A countervailing trend is that industries all around the globe are pressing harder-and-harder against fundamental thermodynamic, quantum dynamic, and informatic limits … which of course are three different names for the same limits.

    This pressing-against-limits is characteristic of obvious technologies like sensors and photovoltaic devices, and of less-obvious technologies like semiconductor device fabrication, and (nowadays) of non-obvious technologies like hundred-ton wing-spar fabrication machines.

    Yes, 21st century composite-layup technologies are quantum-limited sensing-and-control technologies. For every one giant research lab that is closing, many smaller companies that supply these technologies are growing and hiring.

    Summary  Research domains as seemingly disparate as theoretical research in BosonSampling, observational research in advLIGO, and device fabrication at scales both macro- and nano-, nowadays compose a unitary STEAM-enterprise domain that is becoming mathematically naturalized, physically universalized, and economically globalized.

    Observation  Dystopian STEAM-trends are more plainly evident than utopian STEAM-trends; and nowadays both trends are accelerating.

  79. quax Says:

    Not quite sure why so much virtual ink has been spilled on this question from Rahul #24:

    I don’t know. All I’m saying is presumable MSR top bosses had the information to judge. Are we qualified to second guess them?

    Yes.

    Questions of such simplicity really deserve a simple answer.

  80. John Sidles Says:

    Another notable entry in this week’s “what is happening to the great industrial labs?” dystopian/utopian reading list is David Whalen’s The Rise and Fall of COMSAT: Technology, Business, and Government in Satellite Communications (2014).

    Background  COMSAT was founded as a publicly traded/privately held corporation in 1963.

    Utopian opportunities  During the subsequent 51 years, satellite communication revenues have grown to $100+ billion dollars per year.

    Dystopian performance  After a promising start, COMSAT lost market-share decade-by-decade, and guttered to final corporate extinction in 2007.

    Whalen’s diagnosis  (p. 223)

    “Both the DOMSAT [DOMestic SATellite] and DTH [Direct-To-Home] businesses required an entrepreneurial frame of mind — something COMSAT lacked. The ‘best and the brightest’ engineers were not good entrepreneurs — or were they?”

    Will the grim fate of Bell Labs and COMSAT befall IBM and Microsoft too?

    Dystopian and utopian epigraphs  For no clear reason (that I can see) Whalen’s concluding chapter begins with three grimly dystopian epigraphs:

    History is a pack of lies about events that never happened told by people who weren’t there.
       — George Santayana

    History is little more than the register of the crimes, follies, and misfortunes of mankind.
       — Edward Gibbon

    The very ink with which all history is written is merely fluid prejudice.
       — Mark Twain

    For STEAM-researchers and STEAM-enterprises, a more apt epigraph might have been

    Never doubt that a small group of thoughtful, committed citizens can change the world. Indeed, it is the only thing that ever has.
       — Margaret Meade (attributed)

    Open questions  Margaret Meade’s “thoughtful committed” individuals undoubtedly exist today, within organizations like IBM and Microsoft just as these individuals existed at Bell Labs and COMSAT … whether their voices and actions can impart transformational ideas and initiate transformational enterprises remains to be seen. The world wonders!

  81. aviti Says:

    is this related to the MSR SVR?

    This week, word got out that, as part of its current restructuring, Microsoft decided to shut down its robotics group. [1]

    [1]. IEEE Spectrum

  82. rrtucci Says:

    aviti#81, what do you mean? You mean the people at MSR SVR were robots? IMG, IMG, complexity theorists are a more advanced life form than I thought.

  83. aviti Says:

    rrtucci#82, I mean it seem like similar politicos shut down the MS robotics. Maybe current management is tired of those researchers enjoying their passion!

  84. microsoft silicon valley TCS research lab shuts down— easy come, easy go | Turing Machine Says:

    […] 8. Shtetl-Optimized » Blog Archive » Microsoft SVC […]

  85. aviti Says:

    Sentiments to the closing [1, 2] and how Microsourus response [3].

    [1] http://thmatters.wordpress.com/2014/10/14/letter-re-closing-of-microsoft-research-silicon-valley/#comment-2116

    [2] http://spectrum.ieee.org/view-from-the-valley/at-work/innovation/scientific-community-blasts-microsoft-for-closing-of-silicon-valley-lab/?utm_source=techalert&utm_medium=email&utm_campaign=102314

    [3] http://blogs.msdn.com/b/msr_er/archive/2014/10/21/harry-shum-open-letter-to-academic-research-community.aspx#comments