Sunday, June 17, 2012

Startup University

The academic research process is incredibly inefficient when it comes to producing real products that shape the world. It can take decades for a good research idea to turn into a product - and of course most research never reaches this phase. However, I don't think it has to be that way: We could greatly accelerate the research-to-product pipeline if we could fix the academic value system and funding model.

Here's the problem: Some of the smartest people in the world have spent their entire careers building throwaway prototypes. I sure never built anything real until I moved to Google, after nearly ten years of college and grad school, and seven years as a faculty member. And by "real," I don't just mean a prototype that we developed for a couple of years and then threw away as soon as the papers got published. In effect, I "wasted" millions of dollars in funding, and countless man-years of development effort by my students and lab staff -- apart from a bunch of papers, nothing of practical value came out of my entire academic research career. (Maybe I'm being a little hard on myself, but let's take this as a given for sake of argument.) And I don't think my lack of real-world impact is at all unusual in a university setting.

What would the world be like if all of this hard work had actually translated into real, shipping products that people could use? How could we change the structure of academic research to close the gap between playing in the sandbox and making things real?

The plight of the academic is that there is often no direct way to translate ideas into reality -- you don't have the resources to do it at the university, and the academic process forces you to bounce between ideas every few years, rather than sticking it out to turn something into a product. In theory, academics are supposed to be patenting their ideas, and companies are supposed to come along and license the patents and turn them into real products. However, I am not aware of a single project from a computer science department that ever been commercialized through this route. This approach is more commonplace in fields like biotech, but in computer science it is rarely done.

A far more common (and successful) approach is for academics to spin out their own startups. However, this involves a high degree of risk (potentially career-ending for pre-tenure faculty), and many universities do not structure their sabbatical and leave policies to make this easy to do. Most universities also make starting a company painfully difficult when it comes to questions of IP ownership, licensing, and forcing the academic's research to be dissociated with their commercial activities. As a result, you get a bunch of super smart academics who play it safe and stay within their tenured faculty jobs, subsisting on grants and rarely commercializing their work. This means that a lot of great ideas never get beyond the prototype phase.

What I'd like to see is a university with a startup incubator attached to it, taking all of the best ideas and turning them into companies, with a large chunk of the money from successful companies feeding back into the university to fund the next round of great ideas. This could be a perpetual motion machine to drive research. Some universities have experimented with an incubator model, but I'm not aware of any cases where this resulted in a string of successful startups that funded the next round of research projects at that university.

Typically, when a startup spins off, the university gets a tiny slice of the pie, and the venture capitalists -- who fill the much-needed funding gap -- reap most of the benefits. But why not close the air gap between the research lab and the startup? Allow the faculty to stay involved in their offspring companies while keeping their research day job? Leverage the tremendous resources of a university to streamline the commercialization process -- e.g., use of space, equipment, IT infrastructure, etc.? Allow students to work at the startups for course credit or work-study without having to quit school? Maintain a regular staff of "serial entrepreneurs" who help get new startups off the ground? Connect the course curriculum to the fledgling startups, rather than teaching based on artificial problems? One might joke that some universities, like Stanford, effectively already operate in this way, but this is the exception rather than the rule.

It seems to me that bringing together the university model with the startup incubator would be a great benefit both for spinning out products and doing better research.

51 comments:

  1. Congrats on your first post. I totally agree with your point of view. UC Berkeley and UC Davis have started their own incubators as well. We'll have to wait and see how successful those new companies will be and if they can support future companies.

    ReplyDelete
  2. I am not sure I agree. I think ultimately the focus on commercialization does more harm than good by limiting openness and the flow of ideas, data, and code. I have sat through too many seminar where the most interesting slides were mostly blacked out due to patent reasons. I have even been on a PhD committee where I couldn't ask questions about certain chapters unless the rest of the audience left the room (PhD defenses are public in Denmark). And why ask questions? I was forbidden to talk about it with anyone else anyway. Very anti-science!

    However, what directly hurts my research is all the code that I'll never see or be able to reuse because it ended up in a commercial package. If I want to use it I (i.e. some poor PhD student) have to re-implement 1000s of lines of code from scratch. Very wasteful!

    I collaborate with several companies and the single biggest hindrance to this is my Tech-Transfer office and their protracted negotiations over rights. It slows the project down and makes it harder for the industry scientists to get permission to engage in new collaborations because their research units are "billed" by their legal department.

    Of course it should be possible to start a spin-off company, but I don't think actively pushing for more this advances science.

    Wrt "This could be a perpetual motion machine to drive research" I recommend this excellent post by Anthony Nicholl's on the financial aspects.hink ultimately the focus on commercialization does more harm than good by limiting openness and the flow of ideas

    ReplyDelete
    Replies
    1. forgot the link: http://www.eyesopen.com/en/blog/bayh-dole-act

      Delete
    2. Hi Jan - that is a great point. I'm not arguing that we should dispense with "Real Science" or that a startup-focused university would replace the need for more traditional universities, but I agree that there is danger in the idea of allowing commercial interests to dominate scientific enquiry. (On a more cynical note, though, one could easily argue that this is already the case.)

      Delete
  3. Wonderful idea that needs to flourish. I hope more universities get on board and try out this model.

    ReplyDelete
  4. >What I'd like to see is a university with a startup incubator attached to it

    Sounds like you're describing the vision for Cornell NYC.

    ReplyDelete
    Replies
    1. Gun, perhaps - I was not aware they were going to a startup-focused funding model though.

      Delete
  5. > In effect, I "wasted" millions of dollars in funding, and countless man-years of development effort by
    > my students and lab staff -- apart from a bunch of papers, nothing of practical value came out of my
    > entire academic research career.

    I would say this is so true of a lot of research in general and a lot of researchers know this in heart that they are just publishing papers & thats it. Where is the real stuff that actually people can use or it benefit their lives? Thats one reason which made me feel research career is not for me.

    ReplyDelete
  6. > The academic research process is incredibly inefficient when it comes to producing real products that
    > shape the world. It can take decades for a good research idea to turn into a product - and of course
    > most research never reaches this phase.

    What you see as a problem I think is the way it is supposed to be. Academics are supposed to be generating ideas. There is a certain value in incentivizing just creation of ideas without attaching profit-value to them which is the role universities have. Rather than create the one product that might be immediately useful, it is valuable to do a scientific exploration of a series of ideas, some of which don't lead themselves to returns (at least not instantly). IMHO, wanting to turn ideas into products, and wanting to explore new ideas are competing goals. Optimizing for one takes away from the other and the universities are clearly well-tuned for the latter.

    As for the second part, it seems to be the trend that most engineering schools have an incubation center of some sort attached to the engineering school. The Wyss institute at Harvard and the Stevens Institute at USC are two examples I've seen. While they are not as streamlined as you'd like, I suspect thats just a matter of time. Many of these institutes are very nascent and are figuring out the best working model, the individual strengths, and best commercialization strategies for their respective schools.

    ReplyDelete
    Replies
    1. Karthik, I agree that idea generation is _one_ function of academia but it feels like we've created a bunch of artificial barriers for academics who want to do more practical stuff. But I agree in general that academics should not only be doing products.

      Delete
  7. It is intriguing that some research fields have the potential to result directly in products. But I find odd the assumption that the *purpose* of academic scientific research is to create products for profit. It is not; the purpose is to discover new knowledge. Let the corporations worry about investing money in discovering knowledge that makes money for its discoverer. Some knowledge is worth knowing whether or not it leads directly to profitable products. Hence the societal need for government-funded research.

    What product did Watson and Crick have in mind when they discovered the structure of DNA? Countless products are possible because of this discovery, but the direct purpose of that research was the knowledge itself. There is nothing wrong with academic scientists making money through patents or consulting or startup companies. But they should be judged on how important and influential is the knowledge they discover, not judged based on how much money it makes.

    ReplyDelete
    Replies
    1. Dave - Yes, I agree that academics can and should be doing long-range research and not just product-oriented stuff. But I'd like to reduce the barriers for academics who want to get their ideas out into the real world, and would be able to do so if the resources and structure were in place to make it happen. I had a lot of stuff I worked on at Harvard that could have been spun out into companies, if I had the time and were willing to risk my academic career to make it happen. Since I personally didn't care about "getting rich", the safe bet was to stay in academia and forego the potential for commercial impact.

      Delete
  8. As others have pointed out, this is not so simple for various reasons. Two points I'll add: almost all of this research is publicly funded, so the idea that private interests should benefit from this is suspect. Second, want to have immediate impact beyond a paper? Open-source your code!

    ReplyDelete
    Replies
    1. There is no limitation on commercializing the results of publicly-funded research. This is one of the key functions of government funding for science: To create new industries and drive economic growth. If people were not allowed to commercialize publicly-funded science, there would be no Internet, no Google, no WiFi, in fact, no computing technology whatsoever.

      I agree that academic code should be open sourced, but don't overestimate the impact that open sourcing has. Most academic code is junk. I'm including everything I ever wrote in that category :-)

      Delete
    2. Sounds like you're arguing both sides here: "If people were not allowed to commercialize publicly-funded science, there would be no...computing technology whatsoever." Seems like the system is working just fine then. Of course, it's far, far more nuanced than than, but I'll just make that one point for now.

      Delete
    3. My argument is that we would have a lot more innovation translated into real-world products if we could lower the barriers for getting good work out of universities.

      Delete
  9. The core of your idea is ostensible and resonates well with individuals (researchers) in the academic community, who spend great deal of their time publishing papers with no direct application/commercialisation plan. However, there is a great incentive problem when organisations like universities taking direct role in identifying the valuable/commercialisable projects. The core of the problem is that in a university those who will ultimately make the decision on project fundings, don't have much at risk! This introduces inefficiency and can lead to the downfall of the mechanism.

    Another general point is that I personally think commercialisable products/applications do not necessarily lead to highest welfare. I think even binding academic projects and open source community can have substantial benefit - maybe not directly profitable though.

    ReplyDelete
  10. tremendous post. brilliant minds in universities all throughout the
    country and they publish papers. it makes no sense and the model
    needs to change. I always knew of the "problem" but your solution
    is great.

    mark

    ReplyDelete
  11. You did not waste millions of dollars in funding. We the taxpayers did. And as far as making everything open source, those programmers deserve to get paid. Giving away your work is not how to do that.

    ReplyDelete
    Replies
    1. Those programmers did get paid, by our tax dollars that funded the research. In the same way that the research results should be freely available (which Matt has supported in previous posts), the software developed should be released also.

      As Matt describes, the code produced is not product quality. Some enterprising entrepreneur can take that code, run with it, and reap the benefits.

      Delete
    2. I'm all for open research - but open research that never results in a tangible improvement to society is useless. The function of a university should be to benefit society. If academics are unable to have impact on society by publishing obscure papers that a few hundred people read, it seems they should at least be able to generate products.

      By the way, most research done at universities is never open, specifically because the university and the researchers intend to exploit it for commercial gain - this is especially true in biotech. The basic ideas might be published in the scientific literature, but the core IP is typically locked up in patents.

      Delete
    3. Sorry if this is off-topic but software is really not like other products. When you buy a car, you get to modify it all you want, but you don't get the keys to the factory. Open sourcing your code is more like giving away the keys to the factory.

      Delete
  12. Maybe you should look at what Lille (in Frane) is doing with its Euratechnologies center : we have on site a lab (INRIA), a few universities (HEI in particular) and about 100 startups. And there's a plan to a "innovation campus" that would bring together small companies, big companies, labs, researchers and students in a building own by the Lille University. Some pointers in French : http://index.lavoixeco.com/campus-innovation-lille.html

    ReplyDelete
  13. Have you heard about the DMZ (Digital Media Zone) at Ryerson University in Downtown Toronto.

    ReplyDelete
  14. Intuitively, this sounds to me like something which would yield much short-term benefit, some medium-term benefit, and would in the long-term lead to the failure of the universities which embrace it. Education and research simply don't make money fast enough to be of interest to the markets as they are constituted today. If they did, then companies would get involved in providing education and doing research. Instead, only insanely profitable companies do research because they have the cash to burn and can afford to gamble on very long longshots (Google, Microsoft, IBM) and only deeply sketchy organizations provide for-profit education.

    The free market has spoken: research is not a good way to make money. If we try to businessify and entrepreneurify our universities, why would we expect the outcome to be any different?

    ReplyDelete
    Replies
    1. Bell Labs, which seems like a counterpoint to my thesis, was actually mandated into existence by the agreement ("consent decree") between AT&T and the US government which allowed AT&T to be a monopoly for 50+ years. AT&T was required to spend x% of their profits on research, and had little interest in bettering the phone system which was making them money hand over fist, and therefore invested that money into research into whatever. When AT&T was broken up into the baby bells, the consent decree became null and void, and the writing was on the wall for Bell Labs.

      Delete
    2. Peter, this is a very interesting perspective and not one that I have heard before. Let's see. So you're saying that the research that led to the development of new technologies, such as the Internet, RISC, and mobile broadband did not result in multibillion dollar industries and drive tremendous economic growth? I have a hard time buying it. It's true that most research projects are long shots that may not result in that kind of impact, but I would argue that long-term R&D is absolutely essential both for economic growth and maintaining technological leadership. Much has been written about this elsewhere.

      Delete
    3. Hi Matt

      No, I think what's Peter's pointing out is not that research has no economic value, but that often it can be hard to monetize that value in the sense of earning profits in the market. Even in Bell's case, only a tiny fraction of the value of the Bell inventions you mention would have gone to Bell in the form of profits (this seems clear since the long term effects of their inventions now underpin a large chunk of world GDP, but I don't think any truly huge sums went to Bell).

      Delete
  15. So How those theoretical guys could start up their own companies? How should profs/PhDs balance their research with real product? Will all academic publication favor product related topic or theoretical analysis?

    It seems to me the ultimate answer leads to one conclusion: The academic is not suitable for real product invention especially in IT industry.

    ReplyDelete
  16. completely agree that the more integration the better, but the devil is in the details: conflict of interest, barriers from the university non-profit status, attractiveness of early stage and therefore risky technologies if the university gets too aggressive in its negotiation. interesting topic, but not an clear cut pathway

    ReplyDelete
  17. Matt: Let's start with your premise, the first sentence:

    "The academic research process is incredibly inefficient when it comes to producing real products that shape the world."

    One possibility is to accept this statement, and point out that it's like saying: "The Google industrial process is incredibly inefficient when it comes to generating original works of Spanish literature." It's perhaps a valid point, but it's not particularly sensible. I didn't realize that "producing real products that shape the world" was necessarily the point of research.

    But you seem to be saying this in a larger sense, such as what are the contributions of the research process to the world. In which case, I am skeptical of your statement. The company Google, as I recall, began from the academic research process. There were even papers written (which I have students read and teach about in graduate class). So your current employer provides a pretty clear example that your starting premise is at least questionable. I could certainly name a few other companies as well. (Locally, we've got Akamai. Or Sleepycat. I'll let others go on.)

    But if we're looking at "contributions to society", then I think you need to be looking at the big picture. One commenter noted the "discovery" of DNA, for example. I think you can look for similar groundbreaking work within computer science. Les Valiant's work on learning theory (for which he recently won the Turing award), for example.

    I therefore find myself disagreeing with your conclusion:
    "We could greatly accelerate the research-to-product pipeline if we could fix the academic value system and funding model."
    I'm not saying we couldn't accelerate the research-to-product pipeline; perhaps we could. But I don't think you need to fix the academic value system and funding model. I think your plans to fix it are just as likely to break the value of what universities provide.

    "In effect, I "wasted" millions of dollars in funding, and countless man-years of development effort by my students and lab staff -- apart from a bunch of papers, nothing of practical value came out of my entire academic research career."

    Perhaps, Matt, this says something more local, about your research choices, or at a higher level perhaps about how systems research currently functions. But I feel you've extrapolated from your own individual experience and feelings to condemn the entire system. I'd question if that's really justified.

    ReplyDelete
    Replies
    1. Hi Mike,

      So, good points, but I think you are actually agreeing with me, and helping me to make my case!

      I never said that the sole function of universities or research should be to *only* produce products. I'm just saying that universities are generally bad at spinning out companies, that there are ways that the impedance mismatch could be reduced. I think we both agree on this.

      Obviously there are some great examples of companies that came out of research labs, my current employer among them. I believe there could be 10x the number of such companies, if we could structure things right. This could have tremendous impact on the economy and provide a much-needed outlet for all of the great work being done in universities that often flounders in the prototype stage with no place to go.

      The trick, as you point out, is how to do this without disrupting the academic system which is *also* responsible for things like educating the next generation, and doing "impractical" long-range research which may not produce a startup this year, or in the next 5, 10 or 20 years, but we should still be doing - despite the lack of a quick reward. There is a risk in changing the expectation of an academic career to be one which produces a string of startups, and judging an academic's value by their commercial impact. I haven't worked out all the kinks in my proposed model just yet :-) My proposal was more along the lines of setting up a different kind of university, not one to replace the current system, but to complement it.

      And you're dead right that I chose the wrong line of research if you care about building real products. Sensor networks have been a total failure in the market. I was enamored with their beautiful properties as an interesting class of distributed system, but ignored how important it would be to work on things that companies (or the NSF for that matter) would care about funding. My colleagues who worked on more practical stuff - like cloud computing - had no problem getting research funding. Over time, the pickings in the sensor network space became kind of slim. But then again, I always sucked at writing grant proposals, so there is that confounding factor...

      Delete
    2. Matt:

      "I never said that the sole function of universities or research should be to *only* produce products. I'm just saying that universities are generally bad at spinning out companies, that there are ways that the impedance mismatch could be reduced. I think we both agree on this."

      I think we can agree on this. I also think, though, that you were saying quite a bit more (and much more negatively) in your initial post, as the quotes I pulled out demonstrate (at least to me).

      "I believe there could be 10x the number of such companies, if we could structure things right."

      I simply don't believe this, and don't see any evidence for it based on what you say. It seems like wishful thinking on your part. I think you're both underestimating the successes of the current system and overestimating the potential successes available.

      "What I'd like to see is a university with a startup incubator attached to it, taking all of the best ideas and turning them into companies, with a large chunk of the money from successful companies feeding back into the university to fund the next round of great ideas. This could be a perpetual motion machine to drive research. Some universities have experimented with an incubator model, but I'm not aware of any cases where this resulted in a string of successful startups that funded the next round of research projects at that university."

      So you're saying (in the last sentence) that some have experimented with this model but you're not aware of any successes. This is part of why I think your 10x number is remarkably optimistic.

      What you're describing is similar in spirit to the idealized "research lab" model. It might be worth looking there for pro and con arguments for our positions. My experience suggests that short-term business fluctuations make it hard for research labs (in the conventional sense) to last long-term.

      But a further concern you simply didn't address initially -- and you've acknowledged somewhat in your response but still don't actually address in any meaningful way -- is that a university with a startup incubator attached will start to look more like a startup incubator and less like a university. As you say, "There is a risk in changing the expectation of an academic career to be one which produces a string of startups, and judging an academic's value by their commercial impact." Indeed. I'd say that point requires more thought.

      Delete
    3. Mike - you have some good points. There are certainly many reasons why this idea might not fly, but that's why I blog about this stuff and let people comment. To be honest, I'm not entirely convinced it could work either, but I think it's good to shake the trees and see what ideas fall out.

      It seems clear that there's almost no reason for a conventional university to take on the risk of a large-scale incubator model for funding. Even private universities are not supposed to be in the game to maximize profit (although I'd love to hear what Harry Lewis thinks about that). So the fact that the limited experiments with incubators have not been successful so far doesn't mean very much: It's never been tried at this scale before, because institutional priorities are misaligned with this concept.

      You say, "... a university with a startup incubator attached will start to look more like a startup incubator and less like a university". That's exactly right, and is exactly my point. I believe it may be possible to hybridize the university and incubator models, getting the best of both: lots of real-world impact coupled with (relatively) safe jobs for faculty and a teaching mission. This in no way replaces traditional universities and would not be for every kind of student or faculty. It is kind of like a vocational version of Stanford on steroids: get your degree while you do your startup, rather than having to pick one or the other (as Peter Thiel would have it).

      I hear your point about whether this concept is tenable in the long term, from a financial perspective. As a counterpoint, consider the successful VC firms with a long track record and substantial reserves. Wouldn't it be possible to direct some of these resources to doing a university-like academic venture? Or do you believe that academic research is fundamentally incompatible with the kind of market volatility that VC firms have to deal with?

      Delete
    4. Matt,

      Care to introspect on why sensor networks research failed as a community? I agree with you as a former researcher in that space. But, curious what your take on it is. (I know its off topic. Everything I would have said on this topic has already been mentioned).

      Delete
    5. Sensor networks research did not fail as a research topic: It has been very successful by every meaningful academic measure (great conferences, journals, lots of government funding, plenty of PhDs graduated and tenured faculty). As far as productization goes, it has kind of fizzled out. I think the real problem is that sensor networks fell into the "chasm" between research and product (http://en.wikipedia.org/wiki/Crossing_the_Chasm). There are a few startups out there and I certainly hope that they are successful, but overall industry interest in the technology has not been as strong as many were predicting.

      Delete
    6. Thanks for replying. My take is slightly different.
      I think sensor networks research started out with a badly hashed out vision (lets throw 1000s of sensors, allow them self organize etc.). In other words, it didn't really start out addressing an actual problem but became a play ground for different communities to extend their research under certain assumptions (extreme low power, fantastic failure rates etc.). By the time I defended, I had come to believe that the community was addressing quasi-invented problems with quasi-invented solutions. My cynicism aside, I do hope that the few remaining start ups in that space are successful.

      Delete
  18. I do agree that focussing on producing real products through academic research towards would contribute to the field in general. Also, startups are not the only way to make a contribution, open source systems like LLVM compiler system is a very good example of a product from academic settings making a difference.

    However, there are various tangible issues which come up when academic research process is geared towards producing real products. Your old post titled "Software is not science" touched upon this issue. If an academic research group is spending resources on creating a real product, which would require extensive amount of time for removing bugs and adding new features, then invariably the corresponding publications would be inclined towards the implementation. And, as you presented in your previous post, this creates an issue while getting accepted at prestigious conferences. This forces research groups to create softwares that just work, so that there is more time for focusing on theoretical foundations and exploring more research methods.

    ReplyDelete
  19. Nonsense, Matt. Nonsense. Mitzenmacher was too easy on you! :-)

    To quote our friend Pascal, I would have written you a shorter note but I didn't have the time. http://databeta.wordpress.com/2012/06/20/an-open-letter-to-matt/

    - Joe

    ReplyDelete
  20. Some time ago, I attended a talk by Randy Katz. One point that he made really stuck with me. He emphasized that it's the students, not the papers, that are the real output of the system.

    Then the argument isn't that academic research is relevant because e.g. the google prototype was developed at Stanford. Instead, the argument is that academic research is relevant because a couple of guys developed certain abilities - ones that would have been difficult to develop without an academic research background - that enabled them to make a certain kind of leap forward.

    Whether that happened pre-thesis at Stanford or some years later at a company or startup isn't the point. It's the abilities that are (uniquely? most often? most effectively?) developed via PhD research (and that professors pass on to their students) that are important.

    Perhaps this has more to do with your earlier post about the value of a PhD in industry than you thought?.

    ReplyDelete
    Replies
    1. Laura - I agree (and have blogged before) that the value of being in academia is not about what you produce but what you learn. I think that's fine for a PhD student, but for a career academic it's still inefficient, in my opinion, to have some of the world's best talent locked up in a system that makes it harder than necessary to do tech transfer.

      Delete
  21. I'm reading an article on Slate about the current goings-on at U Va.
    http://www.slate.com/articles/news_and_politics/hey_wait_a_minute/2012/06/teresa_sullivan_fired_from_uva_what_happens_when_universities_are_run_by_robber_barons_.4.html
    Worth a read on its own. But the last two paragraphs seem particularly apropos. (Spoilers!)

    "The biggest challenge facing higher education is market-based myopia. Wealthy board members, echoing the politicians who appointed them (after massive campaign donations) too often believe that universities should be run like businesses, despite the poor record of most actual businesses in human history.

    Universities do not have “business models.” They have complementary missions of teaching, research, and public service. Yet such leaders think of universities as a collection of market transactions, instead of a dynamic (I said it) tapestry of creativity, experimentation, rigorous thought, preservation, recreation, vision, critical debate, contemplative spaces, powerful information sources, invention, and immeasurable human capital."

    ReplyDelete
    Replies
    1. I think it's unfair to interpret Matt's suggestion for "a university with a startup incubator attached to it" as "believ[ing] that universities should be run like businesses." All Matt seems to be suggesting is that there are better models for getting tech transfer from inherently applied research areas like CS systems research, and that adoption of such alternative models would increase the value and societal relevance of the research.

      Obviously the startup incubator approach is not going to work for German literature or whatever, and I don't think Matt is suggesting to close down the German literature department because you can't build companies from what it teaches.

      It's fine to advocate a "tapestry" or "portfolio" approach, but when there's a specific area where the university system is falling short, and a specific idea on how to improve it, why not try it?

      (As an aside, I found that Slate article to be incredibly one-sided and full of completely unsubstantiated claims such as "most actual businesses in human history" having a "poor record.")

      Delete
    2. Mike, to be sure, there are risks with running a university like a business. The problem in the case of UVa is that it is run like a *bad* business, with a closed, secretive board that need not answer to shareholders - unlike a public corporation.

      If this pattern sounds familiar to you, it should be - it's also the situation at Harvard, which Harry Lewis did such a fine job bringing to light in the press last year -

      http://www.imackgroup.com/mathematics/738432-harvard-professor-harry-lewis-calls-for-bob-rubin-to-resign-from-harvard-board/

      What I'm proposing for Startup U is a risky experiment and would be radically different than the structure of a conventional university. As I've admitted, it might not work. But today, the only option if you want to spin out tech from a university is to *leave the university* to start a company, and let the VCs (rather than the university) reap the benefit.

      If the concern is that such a university might get too motivated by economic concerns, then I'd argue that this is already the case, when you consider the need to grow the endowment, increasing tuition costs, income from patent licensing, and charging ridiculous overhead rates on grants (last I checked, Harvard's overhead rate was 69%). Sure, some departments act as charities funded by the rest of the university, but bringing in money is a really important function of the research mission. Of course, with all this money flying around, we sure hope it's not clouding any judgement....

      Delete
  22. "today, the only option if you want to spin out tech from a university is to *leave the university* to start a company, and let the VCs (rather than the university) reap the benefit."

    Nope, that is neither true nor sensible.

    1) At most schools it's possible (and at some it's easy) to *take a leave* to start a company and then return to the faculty, particularly if you have tenure. One doesn't have to leave for good. I know you can name a long list of people who've done this at various schools. I'm not sure why you feel the need to paint such a stark picture.

    2) Universities are actively involved in tech transfer with financial gain---but typically the monetization model has been targeted at the natural sciences, where the financial framework has been based on monetizing IP via patents. For example, Wisconsin has made a lot of money on patents for Vitamin D and Warfarin ("WARF" stands for "Wisconsin Alumni Research Foundation", the monetization arm of the university). It's true that most universities are not well-equipped to deal with monetizing software startups, where the potential financial gain comes from holding an equity stake in the company. This gets complicated because of (a) precedent (Universities got good at monetizing patents and would like to just keep doing that), and more importantly (b) because IP licenses are aligned with a key mission of the university (IP presumably entails the involvement of Intellectuals, after all). It's less central to a university's mission to invest in risky startups---and if you've seen the recent Kauffman Foundation report, VC seems like a pretty bad business.

    There is a precedent for what you're suggesting though: real estate. Many Universities make good money as landlords. Harvard owns much of Cambridge. Stanford and Wisconsin both profit handsomely off of shopping malls! Running a VC firm out of a university is probably a worse business than running shopping malls, frankly. So I'm more bullish on Mall U than Startup U.

    Go Mall U! (and its sports teams, the MallRats!)

    ReplyDelete
    Replies
    1. Hi Joe, sorry for the confusion - by "leave the university" I meant "leave the university for some period of time, not necessarily permanently". My point was you can't typically do a startup while also doing your professor day job (either due to CoI issues or because it's too much damn work).

      If you read earlier in my post, I acknowledge that the life sciences have a different tech transfer model. My comments here are largely confined to computer science.

      I won't argue that a conventional university should probably not get in the VC business; that's not what I'm proposing though!

      Delete
  23. Hey, Matt, long time no see. As a "one-foot-out-the-door" academic as well, your post really reflects well a lot of the same cynicism I developed myself in academia.

    The start-up life is certainly a heck of a lot more exhilarating than anything past the first couple of years I spent in academia. I ran through a hell of a lot of grants as well and I definitely see elements of the work popping up in industry. Oddly enough, I had nothing to do with those things showing up, but I think I was sometimes on the right track which makes me feel slightly better. I also remain insanely proud of the students I produced.

    But I think what really sealed the deal for me though was serving on NSF funding panels. It is amazing that anyone could assemble a set of more uninspired people than I met while evaluating grant proposals (especially the PMs).

    Personally, I think the secret to "fixing" computer science academia is to fix the funding model. And I don't mean more funding, I mean better funding, because 90% of it might as well have been set on fire given the way it is distributed.

    Here is my two point plan:
    -Get insane program managers with crazy visions like you find in DARPA
    -Make NSF panels no longer anonymous. The reason why NSF panels don't look like PCs of top conferences is because they are blind and academics are basically publicity whores at heart.

    Good luck Matt, hope to run into you sometime soon.

    ReplyDelete
    Replies
    1. Hi Mark,

      I will be frank that one of the things that drove me out of academia was frustration with the funding model. At one point I realized I was spending more time writing grant proposals to support my students than some of those students were spending on research! Since doing my style of systems research requires a good-sized group, it was a never-ending scramble to cobble together enough funding to keep the group going. And, I was never very good at writing proposals. It was always so depressing to pour hours and hours writing a proposal knowing that there was a 90% chance (in my case) that it would not get funded.

      Anyway - I agree that this needs fixing, something I've blogged about before:

      http://matt-welsh.blogspot.com/2010/05/proposal-improving-nsf-review-process.html

      Good luck on your endeavor!

      Delete
    2. You should have moved to Canada. Researchers of your caliber have 80-90% success rates around here.

      Delete
  24. Matt,

    Interesting and highly relevant post, more so given what we do at Osage University Partners - a new venture fund created to invest only in university spin-offs by partnering and profit-sharing with ~45 universities. Please check out osageuniversitypartners.com to lean more. If you are interested, I would love to share more over phone. Let me know.

    Sudip

    ReplyDelete
  25. In terms of incubators, you should consider checking this non-profit out.

    http://opensourceecology.org/

    ReplyDelete

Startup Life: Three Months In

I've posted a story to Medium on what it's been like to work at a startup, after years at Google. Check it out here.