Tuesday, September 6, 2016

What happened to NSF and other pillars of US S&T

Some people I know are full of excitement, asking what could be done to make US S&T more innovative. 
They remind me of space enthusiasts who eagerly make detailed plans for Mars, while quietly allowing the basic technology we need to get even so far as earth orbit on an affordable basis slide into oblivion and the past.

In response, I gave a quick summary of the huge recent changes at NSF (which could possibly be repeated at places like the Fed and the NSA, creating even more damage). 

Here is my response: 


First, how do you define "innovation"?

You have one brief definition, "discovery and delivery of...  (new stuff with value added)." When I came back to get the words straight, I clicked on your first box, and saw a definition more on the "delivery" side. The difference between the two is very important, because there have been debates for decades on how to promote discovery,. on the one hand, and how to do translational work to overcome the "valley of death" between discovery and  value added, on the other. 

It will be very easy for a report on this subject to become widely visible and widely quoted. But to make a big improvement would be a huge challenge, and it depends a whole lot on who is really ready to listen and act, to what end. 

In truth, many many organizations all over the world have profited for decades from the discoveries funded by NSF 
by free-wheeling professors at universities and by small businesses. (Earlier on, DARPA and xORs and NASA were more
or less coequal in basic discovery in areas like electronics and photonics, but changes in NASA and by Tony Tether resulted in lots of "refugees" raising proposal counts at NSF in those areas circa 2003. I was there, saw the numbers. Bell Labs also had a great prior history. Some is chronicled in chapters in the book Leadership in S&T edited by William Bainbridge, which should be a must-read and must-cite for this exercise.) But lucrative and useful as this was, it doesn't work any more without the discoveries to translate.  

From 2012 or 2013 to about August 2017 there has been a highly focused change in corporate culture at NSF which many of us believe eliminates the discovery function of NSF, or, to be more euphemistic, puts the agency in harmony with mainline DOE
and the new NASA. The analogy to NASA is more apt, as Lamar Smith has been a lead player (certainly more than Obama) in defining and implementing the changes (though Shelby was there first with NASA). When I was executive VP for policy of the National Space Society, I remember vividly certain ringing words from Smith's office: the US will go to Mars (at least as flyby) using no unproven technology whatsoever." Nothing so novel as a space shuttle, let alone newer than that, when after all we can use the Apollo technology.   

Since I have lived this story, which has many many dimensions to it, it is hard for me to resist the temptation to say a lot more, and to document many, many aspects. But let me focus on the bottom line. From the birth of NSF (or at least since I arrived in 1988)
to 2012 or 2013, most of NSF energetically fulfilled the vision of Vannevar Bush, which I will call "the old NSF" now. The old NSF had many very visible imperfections. I often cited what Winston Churchill said about democracy once -- the worst system you could possibly imagine, except for everything else anyone has ever invented.  (That for discovery let alone translation.)
At times I compared it to an internal combustion engine, about 25% mean efficiency if you are lucky... but still a whole lot better than the 0 percent creativity you get from well-controlled quality control committees and things I saw elsewhere. 
(Could an all-electronic alternative work better for discovery at NSF, as it does in cars? THAT kind of AI is not on the present horizon!) 

In my view, your committee and the folks listening to it would be doing a huge service and performing way above norm if you could so much as restore NSF to the level of discovery it had at peak under the old NSF, restoring the vision of Vannevar Bush.
I could certainly imagine better, at least in principle, and I have certainly spent years agonizing over the huge losses we had in the "valley of death," but the inertia, loss and entropy which have already occurred are a huge matter already, with far-reaching implications, and not easy to reverse.


Just a few vignettes come to mind.

I remember a time, under the old NSF, when an idealistic young program director put together a review panel
which did an utterly incompetent job of reviewing a proposal for a new medical device which I happened to know a lot about.
She had lots of new age ideas about how to do things right, and was equally agile in finding excuses for utter incompetence, for reviews by people who did not even understand the basics of the proposal. I remember thinking: "This is such an obvious travesty,
such an open flaunting of truth and justice, that lawyers could have a field day with it." And then: "But I know how lawyers
and political interventionists work. If we bring them in at this level, this one case may be fixed, but it will all get much worse than it already is." (I worked for 9 years at DOE headquarters, and that helped me see clearly how those slippery slopes can end up.)

At another time, in 2014, I had a long conversation with a friend in the upper management of NSF (still there) who chose to stay and survive the new situation. "We agree that there is a new gestapo in place, and that the words 'transparency and accountability' have become like the 'ministry of peace' in 1984. But the problem is -- whose gestapo is it? Is it Obama's, or Smith's, or ...?"
I opined that Obama could not have wanted a lot of what was happening, and that he must have been playing golf.

In the news, there was a story about a major government employees group (union?) having a big meeting with Obama, where he said to them and to the press: "Don't worry. I have your back.." Many of us laughed.

A year or two ago, I joined an IEEE group meeting with the key science staffer from minority staff of the House Science and Technology Committee. At the time, I had seen retirement figures for those substantive scientific leaders, program directors, 
who were the core front-line people making Vannevar Bush's vision real. About half of them were permanent full-time government employees (strictly speaking, excepted service rather than civil service in most cases), increasing working 'way more than 50 hours a week due to "refugees" and new rules... and I had worried what happens when more than half of them go away. 
I was worried: "The workload to do the job the way we are supposed to will become near impossible for those rotators who make up the other half." I didn't say that at the time... but the staffer told us how Smith was successful in leading a new moral crusade to get rid of the rotators too.

Obvious question: "Who will do the work?" Two answers. First give up the kind of scientific leadership we had in the past, and fill the slots -- especially key "commissar" positions -- with reliable obedient predictable people rotated in from other agencies. Second, redefine the work. (For example, get rid of that old idea that no program director should sign off on funding or declination of a proposal without actually reading and understanding it and the reviews. Use more automated systems like the one developed by George Hazelrigg reported on in Science, a great way to reduce the labor required.) Some said it was a great innovation to reduce the quantity and quality of labor involved in deciding what to fund -- but when that kind of overhead was no more than about 5% of the budget, and the productivity of the rest is very much at issue, I remember "penny wise pound foolish." 
But lazy and incompetent people would rejoice; less work to do in-house, and less real need to work so hard to understand stuff. 

When I was at DOE under Reagan, people joked: "He decided not to abolish it after all, since it's really already been privatized. It's a joint venture, owned 60% by the oil companies and 40% by the nucs." In the old NSF, people would joke at times that it too was a joint venture, 50% owned by the American Physical Society and 50% by the deans. I certainly know a lot about the foibles of both of these, but they were committed to a vision like that of Vannevar Bush, and to principles of free speech as a key to a free market of ideas and creativity. But now the dark hand of the other kind of stakeholder system has appeared more and more. 
Darker than you would possibly believe if you had not seen it, and cross-checked with others.

But that's probably more than enough for a Saturday morning.    

No comments:

Post a Comment