The Rise and Fall of Industrial Research Labs

The Rise and Fall of Industrial Research Labs(cacm.acm.org)

71 points by yorp 11 years ago | 24 comments

greggarious 11 years ago |

This article seems a bit hyperbolic.

Yes, the Silicon Valley branch has shut down, but MSR's main office in Washington is still open and doing great research. It's probably the best industrial lab that's still running.

I also have interned at PARC (no longer "Xerox PARC", but still kicking). I was given remarkable freedom both by my direct supervisor and the division manager.

Down the road, HP labs was still going (though a few of their staff had migrated to PARC). Likewise, IBM and Bell Labs still are going.

And Google seems to be really spinning up their research. I won't comment on specifics in a public forum, but I've noticed signals that they're taking basic research seriously. (They used to be very short term, product oriented)

anonetal 11 years ago | |

I agree that MSR is probably the best CS industrial lab out there right now, but I think Moshe's point is that even the rest of the MSR is probably a bit wary right now given what (and how it) happened to MSR SV. I don't think HP, IBM, and Bell Labs (with a few specific and localized exceptions) are doing basic research as they were doing a few years ago (or in Bell Lab's case, a few decades ago). Maybe you are right about Google, but generally speaking, there are very few papers from Google Research (relative to their strength).

DrDimension 11 years ago | | |

Agree. It matters because nothing undermines scientific integrity and productivity like financial insecurity.

Qworg 11 years ago | |

There's also many other MSR labs in many other places: China, Cambridge, Israel, Egypt, Munich, others. Redmond is the just the biggest.

I believe that the SV cut was the right choice vs. "spreading the pain around", but we'll see how it shapes up in the end.

walshemj 11 years ago | |

And not all RnD labs are CS the same retreat from RnD can be seen in a branches of STEM unfortunately.

jacquesm 11 years ago |

I think it's more that the low hanging fruit has been plucked than that there is a 'fall' of industrial research labs. Just like with technology in the early 20th century there was a golden age where advances were relatively quick and dramatic because the fields were brand new. Once the field matures it takes more effort to get up to speed on what has already been learned and evolutionary progress rather than revolutionary takes over.

And so far we seem to be doing pretty good with that level of progress. Any faster than this and I'm not sure we'd be able to properly integrate technology into our culture and society before it had become obsolete already.

Just look at for instance the succession of audio recording and distribution methods to get an idea of that: the gramophone record lasted for many decades, CDs succeeded them and lasted for a couple of decades, digital formats are dying out about as fast as they are being created (with the exception of mp3).

At some point you're going to have problems of interoperability simply because of the speed of progress (we're seeing something quite close to that on the browser front right now).

aragot 11 years ago | |

To explain this acceleration of technology, let's remind the exponential growth of population: There are more people on Earth than dead population since the birth of times

So there's as much potential for science advance in 1 lifespan as since the wake of Humanity.

danieltillett 11 years ago | | |

>To explain this acceleration of technology, let's remind the exponential growth of population: There are more people on Earth than dead population since the birth of times

Actually there have been around 100 billion people born over this history of human kind. Given a bit over 7 billion people are still alive that suggests most people ever born are dead.

As an aside I have heard this suggested as an argument that the probability of death is not one. If only ~93% of all the people born have ever died then you can't be certain that everyone born will die.

alecco 11 years ago | |

I beg to differ, based on experience. The low-hanging fruit is gone only for the paths mainstream research take. There were some great discoveries recently but since they don't resonate with the "in" crowd it gets mostly ignored.

jonmrodriguez 11 years ago | |

There is a lot of low-hanging fruit in genetic engineering and robotics!

fest 11 years ago | | |

Re: robotics- you mean low-hanging fruit research-wise or application-wise?

I'm not closely following latest research in this area but I get the feeling that most of the focus is geared towards autonomous cooperation and nano-scale. What are other directions?

Zigurd 11 years ago |

I'm usually on the side of pointing out the squishyness in some numbers, like US productivity and GDP. But we seem not to be in a crisis or even decline in corporate research. Microsoft and Google and, to a limited extent Facebook have taken over from IBM and Xerox and Bell Labs.

The Bell Labs business model was highly artificial, and that artificiality was used as an argument not to break up AT&T. Current structures are more sustainable.

Some ways of structuring corporate research are new: At Google, and at startups funded by Founders Fund, you will find R&D that's been selected for impact.

angersock 11 years ago | |

Bell Labs got us microwave communications, lasers, transistors, Unix, C, awk, Plan 9, fiber optics, and a bunch of other things.

The new "sustainable" structures, frankly, aren't even a substitute.

sgnelson 11 years ago |

http://cacm.acm.org/magazines/2015/1/181626-the-rise-and-fal...

For those of us who don't want to have to read the mobile version.

lukastsai 11 years ago | |

https://getscroll.com/r/h1h6t

DrDimension 11 years ago |

Why fritter away all that money on fundamental scientific research when you can spend 2.5 billion for fucking Minecraft?

michaelochurch 11 years ago |

Being in technology, you have a front-row seat to the concept of elective decline, or the idea that civilization's decline is often a collective but intentional choice, rather than the result of external stress. It seems sad, to a historian, that a great empire would "collapse" (an affair that usually happen imperceptibly over hundreds of years) but the reality is that it's the unintended product of millions of individuals making self-interested and possibly self-beneficial decisions... that end up not keeping up or advancing the civilization. At some point, they just decide that it's not worth it to be civilized anymore, and investment ceases.

Since the 1970s, we've seen elective decline in the U.S., in science, and in technology. Abstractly people want scientific progress, but no one wants to pay for it, and people within the masses would rather be guided by their resentment of academics than fight back when academic or research jobs get cut. When state legislatures cut funding for public universities, the hoi polloi don't care because their resentment for professors is stronger than their sense of a need to keep up the society.

One might hope for Silicon Valley to be better, and look to it for leadership, and it may have kept its integrity for longer, but the current "M&A has replaced R&D" era is just fucking disgusting. It's easy to focus the hatred on a few unlikeable celebrity founders (and I've done my share of that) but the truth is that the problem is really deep and probably unalterable. We have a front-row seat, if we work in science and technology in the U.S., for elective decline-- why it happens, the individual actors who push it forward (not wanting decline, but valuing self-interest more), and the often one-way erosion of trust that tends to make it irreversible-- but we have no power to change it. And, just as one might read about a civilization that collapsed 3,000 years ago and think, "They would have been fine if they just <X>", we can easily come up with solutions that will work but never see implementation, because (as seems to be a constant of human organizations) the wrong people are in charge.