Tuesday, September 14, 2010

How Can You Spend $150 Million?

Suppose you happen to have $150 million burning in your pocket that you want to use to help the theoretical computer science community ($150 million endowed will bring in about $6 million/year) How would you use it?
  • Create an Institute for Theory of Computing?
  • Endow a few theory conferences?
  • Create a 150 millennium prizes?
  • Build a fancy new CS building?
  • Endow 25-30 professorships or a 75-100 postdocs?
  • Hire lots of Fields medalists at ridiculous salaries only if they spend all their time working on the P v NP question?
  • Something else?
This is all hypothetical, I don't known anyone, outside of the Simons Foundation, who has $150 million ready to spend.

$150 million doesn't go as far as it used to. In 1992, Henry Rowan gave $100 million to Glassboro State College in New Jersey, now called Rowan University. How much would it cost to get your name on a university today (say at a large private Midwestern university with a geographically confusing name)?

46 comments:

  1. Use it to build and subsidise a prediction market in science. To reward those that know the correct answer and are willing to bet on their knowledge

    ReplyDelete
  2. One should be able to create many more than 150 Millenium prizes with $150M, since only some small fraction will be paid out.

    Maybe TCS needs an easier Millenium prize?

    ReplyDelete
  3. i gotta be honest, i am getttin a little bit sick and tired of hearing china here, china there china everywhere. For instance china theory week and lots of propaganda on various threads ..... it just blows my mind ..... why dont we get a grip on this delicate issue of leadership.

    I admire people in beijing who try to build/mimic our engagement and want to challenge our leadership in theory of computation but at the same time, i am american and would like to see a definite resurrection of our old image in that field.

    At the end of the day, it is a zero sum game.

    ReplyDelete
  4. Who mentioned China?

    ReplyDelete
  5. > Who mentioned China?

    the previous anonymous who pretends to be sick of it...

    ReplyDelete
  6. An unrelated comment:

    Why are soda 2011 registration fees so high? And why isn't there a postdoc/other registration fee for people who are not students/professors/employed by big corporate lab? This is *not* an inclusive gesture by our community.

    The *early* fee is $570.

    ReplyDelete
  7. I would build a new Dagstuhl.

    ReplyDelete
  8. Historically, one can see that a lot depends on how much existing funding already is available. If there is a true TCS funding shortage these funds would allow to hire otherwise wasted and unused talent. On the other hand if theory of computing is already firing in all cyllinders, the extra money will simply imply more funding for researchers who were not going to do good work anyways.

    Lastly, in terms of broad impact how about providing teaching relief down to one course every two years to the top 250 theory professors in the world? If a non-overhead rate is negotiated 6 million would be more than enough to cover the expense.

    ReplyDelete
  9. I think the money should be used to employ people.

    If you look at the statistics from last year, you basically had to win a major award to get a TT Asst Prof.

    There are many successful researchers who were hired approximately a decade ago who never won such awards and nevertheless went on to productive careers. So yes, much talent is currently being wasted by the current jobs situation.

    ReplyDelete
  10. So yes, much talent is currently being wasted by the current jobs situation.

    But most of these people landed post-docs did they not? This seems to be very much the case around here: anyone worth hiring landed a postdoc somewhere.

    ReplyDelete
  11. i think money should be given to kamouna, whose talent i got to know via this blog. he can produce more than 250 professors combined.

    ReplyDelete
  12. I don't think a postdoc is an optimal use of one's talent, if they are as qualified to be a prof as people who are currently professors.

    Also, unfortunately, there are people who didn't land postdocs.

    ReplyDelete
  13. but kamouna is removing posts from his blog comments section so maybe would not be deserving of this money as much as lance and bill for not removing his comments

    ReplyDelete
  14. Try to bring tcs to high school children. They study physics and chemistry. Why not tcs.

    ReplyDelete
  15. but lance and bill have not solved the simplest of all problems np vs p.

    kamouna has solved this two times over. if you can give $150 million dollars to kamouna, as a bonus, i will put my own 150 pennies to go with it. may be kamouna will prove that 150 pennies is bigger than $150 million or else the world is a paradox.

    ReplyDelete
  16. also, it is quite a bit satisfying to kamouna that somebody is actually visiting his blog to post a comment.

    ReplyDelete
  17. Fund a children's show introducing basic computational concepts and get it on PBS.

    This is admittedly a bit of a long term plan.

    ReplyDelete
  18. Your naivety amazes me Lance!

    ReplyDelete
  19. "Create a 150 millennium prizes?
    Hire lots of Fields medalists at ridiculous salaries only if they spend all their time working on the P v NP question?"
    ..i'd rather burn the money(literally) for heat...

    ReplyDelete
  20. Here is another question: say you have $150 million and you want to help science. Would you invest in TCS, another field of CS, or another area of science altogether?

    ReplyDelete
  21. 20% of posts already mentioned Kamouna even though he was not here, yet. This is called impact.

    ReplyDelete
  22. "But most of these people landed post-docs did they not? This seems to be very much the case around here: anyone worth hiring landed a postdoc somewhere."

    Many people, well worth hiring, have given up. Partly it depends on your timing. If you were just finishing up a first postdoc a year or two ago, and are now looking at several more years of postdocs before the job market improves, you are probably better off leaving the field. There is a strong bias toward hiring newer people, and this bad timing will severely hurt your chances.

    ReplyDelete
  23. Can anyone give me a good estimate of what is being spent on CS research (with or without Lance's generous $150 million)? Of course this is an open-ended questions, since you can be very broad or narrow in what counts. Therefore, I will be thrilled at any well-reasoned answer.

    ReplyDelete
  24. come on guys, can you get serious here. If you have something serious to say, then comment on it, otherwise don't waste our time. This used to be a serious blog with enlightening comments. What happened to this standard.

    LANCE, do something to maintain the standard ... or else, I won't be browsing through ...

    Another thing, if you want to criticize someone's work or work ethics, this is fine. Do it in ONE COMMENT with something insightfully argumentative.

    For instance, the comment on Leadership and Competition in China was somewhat worthwile reading. Comments on kouma were, less worthy reading.

    maybe i am simply asking for too much ... should we be thinking of retiring this blog altogether, then ?

    ReplyDelete
  25. Anonymous #24, I agree. But I think part of the reason such inane commentary is here is because the blog posts themselves have become much less technical. I can't recall any serious post on the purported main topic of this blog, computational complexity, from the recent past. There are useful announcements, some fun philosophical discussions (like the one about the Rubik's cube) but hardly any serious posts on theory. I really miss those.

    ReplyDelete
  26. Kamouna is from Egypt. African average IQ is around 70, with northern Africa about 80-85. That is why he is so dumb.

    ReplyDelete
  27. I think researchers who want to elevate the complexity blog should start offering to write guest posts -- and these guest posts need to include LaTex (or whatever mathematical formula-rendering Blogger supports). I'll email Bill with a suggested topic of my own, but seriously, I'm the juniorest kid on the block here. Why not tell other people about the cool paper you just read?

    ReplyDelete
  28. Just to point-out the obvious, the STEM community's posters stand on mighty shaky grounds when they criticize the STEM community's bloggers.

    The entire "first-name" cadre of mathematics bloggers—Bill, Lance, Scott, Dick, Gil, Terry, Tim, Dave (and others)—is doing an outstanding job IMHO.

    In contrast, of 27 responding comments on Lance's important topic, so far the overall quality has been ... well ... not what Lance must have hoped to elicit.

    Posts that are irrelevant and/or toxic are easy to ignore ... but the lack of good posts (to date) is impossible to ignore ... and perhaps signals to the Simon Foundation that investing their resources wisely won't be easy.

    ReplyDelete
  29. I agree with the professor here. Miami University (OH) is a very misleading name. Some rich SOB should do something about it.

    ReplyDelete
  30. Lance, I think you really need to make the comments moderated. Too much graffiti on the playground.

    Free registration for conferences would be great and not too expensive. The Simons Symposium on Theory of Computing. SSTOC.

    ReplyDelete
  31. Perhaps the best investment would be to create a substantial middle school / high school combinatorics and computing curriculum, then train some of the theoreticians who did not get a postdoc but are good and like to teach to get positions in schools to teach it. Subsidize these teacher salaries (in many school districts salaries for people with PhDs are already quite high). Publicize competitions like the Math Olympiad and support it.

    [In some states: Use the bucks to force through reforms that allow these folks to become teachers.]

    It is far from impossible to do science -- especially CS and Math -- while teaching school. You have three months to do research (yes, the 150M could also support such activities, or better, leverage NSF/DOE/DEd funds to sponsor it.)
    If teaching became respectable -- and at least mildly decent-paying, it would open a huge labor market. And, in a nice positive feedback, the presence of a large number of well-educated and very smart people in teaching would change the image of the profession.

    As for the cost of a university today, how much did Stanford or Duke pay originally? I believe 300M made Booth the name of a business school.

    ReplyDelete
  32. I would use the money to add new tenure track positions and hire recent grads and post docs. Like someone else said too many good candidates end up leaving the field because a lack of opportunity.

    ReplyDelete
  33. Give money to develop computational complexity in subsaharan Africa. There is a lot of unchecked talent in the bushes. Establish a special unnamed award for African-origin contributions to the computation theory, and name it after the first negroid researcher who publishes a notable result.

    This is very cost effective. With salaries 300$ a year, Africa has very cost effective population. If you pay your fellows three times that amount, you can have over 6000 fellows. That is, assuming budget 6 million per year, 6000 top brains, a lot more than what you can do in USA. The field will explode, and P vs NP will be solved in no time.

    ReplyDelete
  34. lance why aren't you getting any quality feedback ? Should we retire weblog.fortnow.com ....

    i dont know anymore, i am worried about the deteriorating quality replies ... is there something that could be done here ?

    ReplyDelete
  35. Anonymous asks: Lance why aren't you getting any quality feedback?

    With the most serious intent ... and therefore (in my case) a pretty substantial risk of achieving only comedy ... I will suggest that this is among the hardest questions ever posed on the Fortnow/GASARCH weblog.

    Lance's question invites us to contemplate the sobering reality of a North American STEM enterprise that, by almost any per-capita measure, has declined steadily decade-by-decade, since (very broadly) the late 1960s.

    That's two full generations of steady, genteel decline. If we interpret the Simon Foundation Endowment as an opportunity to contribute substantially to reversing that decline ... well .. that is a pretty undertaking. Because (obviously) the obvious, easy, and conservative answers have all been tried, without notable efficacy.

    Twice I've tried to write a thoughtful response to Lance's challenge ... twice I've been dissatisfied with the result & so did not post it ... and perhaps other folks are having the same experience.

    I'll try a third time this weekend ... because Lance's question *is* serious and important ... and will post the result ... *provided* that it turns out better than the first two attempts.

    ReplyDelete
  36. I would use the money to make lasting effect. I think we have related two main issues in TCS, one is job market for graduates and the other is recruiting bright students. I don't have any ideas how
    to change the first issue. For the second issue, organize an annual TCS boot camp for high school students, and organize TCS tours for high school students around the country.

    Start a project to study and develop next generation methodologies and tools for TCS research.

    ReplyDelete
  37. Shame on all of you for suggesting that we should spend this on TCS.
    Spend it on Cancer or AIDS research. The world would be better off if the kids know more about gene regulation, than ramsey theory.

    ReplyDelete
  38. Shame on all of you for suggesting that we should spend this on TCS.
    Spend it on Cancer or AIDS research. The world would be better off if the kids know more about gene regulation, than ramsey theory.


    I believe just the opposite: diseases like cancer and AIDS will only be cured if we achieve significant advances in TCS that can be applied to the biological sciences. The lack of ability to interpret the biological data we already possess is a far greater obstacle than our inability to produce new, "better" biodata. I'll give one examples.

    Suppose you want to design a particular drug. You can start with certain compounds, obtain intermediate compounds through controlled reactions, and, at the end, you obtain your objective. If you make each compound a vertex, and each reaction an arrow, you obtain a Directed Acyclic Graph known as a synthesis graph. How to compare complexities of different synthesis graphs, and how to generate a low (lowest?) complexity synthesis graph from a desired final product, are wide-open chemical engineering problems. The science write David Bradley once wrote, "It is rather astonishing how little chemists really know about designing the shortest and simplest synthesis of a new molecule." Note that the primary barrier here is a lack of algorithmic understanding. Such questions fundamentally relate to TCS. In fact, one of the main measures of complexity of graphs that appear in chemistry is to count the number of distinct spanning trees of that graph. An identical complexity measure is used in the theory of computer networks.

    Also, about Ramsey theory: the first paper I ever published about nanomolecular self-assembly applied a theorem whose proof required Ramsey theory. Bill Gasarch even posted my paper on his "Applications of Ramsey Theory" web page. I am confident there are far more applications of Ramsey theory to the complexity of molecular programming. When working with enough molecules to make something that can form something as thick as a human hair, the numbers are large enough that "Ramsey phenomena" are bound to emerge. I have no rigorous proof of this at the moment, and perhaps the future will show that I am totally whacked. However, it's a hunch I plan to follow until further notice.

    ReplyDelete
  39. Aaron Sterling's post (IMHO) exemplifies what folks hoped to see posted on this topic.

    Uhhh ... possibly this is because I agreed with its main points ... yet none-the-less ... it was a well-thought-out, well-written post ... for which thanks and appreciation are extended.

    ReplyDelete
  40. $1 million could commission a whole pile of free textbooks to be published online, at every level from junior highs chool through postgraduate collegiate. Most of the CS that I know came from Wikipedia which was written by folks writing recreationally in their spare time. There are also some good theory textbooks online written by real researchers, but nowhere near enough. Imagine if there were dozens or hundreds.

    ReplyDelete
  41. Another thing to do: endow some modest prizes for best results whose publication of record is in open-access journals. And create and fund a journal like "PLOS Theoretical Computer Science" with a prestigious editorial board and high standards, so people could get tenure by publishing in it.

    ReplyDelete
  42. Give 150 million to Iraqi children, to develop their skills, as part of war reparation effort. Alternatively, give 75 milion to Iraqi children and 75 milion to Vietnamese children, that gave us new Fields medalist. I suspect agent orange didn't help that happen, so there is a moral obligation to do something in these countries other than testing military equipment.

    ReplyDelete
  43. @Aaron Sterling

    Nice argument, can we pursue it a little...
    What would allow more efficient research in mathematics the same way that mathematics allows more efficient research in chemistry/biology?

    ReplyDelete
  44. What would allow more efficient research in mathematics the same way that mathematics allows more efficient research in chemistry/biology?

    Not having to worry about living expenses.

    I am certain that I could have been more focused and productive over the last few years if I hadn't been worried about food and shelter. Beyond that, professional development costs money. Most recent example: the PODC 2010 Organizing Committee (very graciously) offered me significant student support to attend their conference, and I turned down the award because I just couldn't swing the rest of it. Previous to 2007, my academic experiences had been in "top" places, and, before I arrived here, I had no idea of the financial gulf separating Ivy League institutions from schools that are "merely" good. (Let me say that I am *extremely* grateful to the Center for Computational Intractability, which has made it possible for me to attend a few workshops there at this point -- thanks very much.)

    Also, sucky as things have been for me, my own problems are small compared to foreign students in the department, who have a visa-gun to their heads each semester, and greater expenses.

    I would probably make the same choices again, even if I had a time machine, just to be clear. I chose a school based on faculty expertise in areas I found most exciting. I already, in 2006, understood/believed the connection I mentioned above between TCS and the biological sciences, and I doubt I would have enjoyed myself as much (or been able to produce things as off-the-wall/"conceptual") if I'd been a student at, say, Princeton or CMU (if I'd even gotten in). However, I'll be moving to a different school's theory group for a year, starting in January, because they have funding for me, and my advisor and home department do not. I'm grateful for the opportunity, that is for sure, but I've become a gypsy-researcher, and I'm not even a postdoc.

    ReplyDelete
  45. Sometimes foreign students get fewer expenses. For example, if they live in a dorm, then I believe they do not have to pay state income tax.

    ReplyDelete
  46. @Aaron Sterling
    Not having to worry about living expenses.

    Of course, makes sense, but that is not what I meant.
    What did I meant? :-)

    ReplyDelete