Pentagon to Significantly Cut CS Research

GabrielF writes "Over the last few decades, DARPA, the Defense Advanced Research Projects Agency has funded some of the most successful computer science research projects in history, such as the Internet. However, according to the New York Times, DARPA has recently decided to significantly cut funding of open-ended computer science research projects in favor of projects that will yield short-term military results. Leading computer scientists, such as David Patterson, the head of the ACM are outraged and worried."

Comments

sigh...

[ocularDeathRay]

I am not surprised but this is kind of sad. Lets stop open ended research that may help people in the future... instead we will spend that money on killing people in the short term.

as great as this country is, it is sometimes frustrating to be an American

Re:sigh...

[Rostin]

There are at least two false dilemmas, here.

First, why do you assume that short-term military spending won't help people in the future? It's not at all obvious that having a powerful, technologically advanced military prevents us from helping people in the future. I would hope that the reverse is true, in fact.

Second, do you think there's a compelling reason to believe that in the absence of military research, people would stop killing one another? Isn't it true that (at least in theory) having better, more accurate weapons means that we kill *fewer* people?

Re:sigh...

[tbo]

Disclosure: I'm a graduate student at a major research university, doing public research that happens to be funded in part by DARPA and the DoD. The research is long-term, but is in a field that will clearly have national security implications.

I'm going to go out on a limb here and say "research." I've never seen an educational institution that was wasteful about it's funding (Maybe Harvard).

Then you've never seen how research happens at a major university. Waste happens *differently* than at major corporations, but it happens in vast amounts, often in the form of wasted time.

At a private company I used to work at, when there was a minor problem with my working environment (too cold), it took a day or two to fix. At a top-rated university, a more serious problem (lights that turn off by themselves every ten minutes) took seven months to fix.

At the same company, security was taken very seriously. When the door to the server room was being repainted, we had a security guard stand there, literally watching paint dry. At the major university, we had five break-ins to our building last semester and yet it's still possible to break in in 15 seconds with nothing more than a newspaper. (The last of those break-ins cost the university about $10,000 in computer equipment, and it took four months to get the computers replaced and running again).

I haven't even started on the amount of time wasted on pointless administrative tasks (e.g. two weeks telling payroll how to do their jobs).

The professors and grad students are paid wages that nobody in the private sector would accept. They don't have crazy offices or private jets or 100,000 dollar golf club memberships.

Professors don't get crazy bonuses, but the top administrators get pretty hefty salaries and bonuses (like a beautiful house on campus). Compensation for administrators is approaching corporate levels.

Plus, universities find lots of ways to sphon off federal grant money. Any major purchase or salary coming from a federal grant gets a ~50% "overhead" charge tacked on--that money goes to the university.

It literally hurts me to see DARPA cut funding to universities (my group took a hit), but I can understand why it's happening.

Well...

[sabernet]

While this does royally suck, we cannot forgot DARPA is a defense agency after all. And in the modern, "Make war, not talk" times of the current administration, this was almost forseeable.

Re:Well...

[flyingsquid]

Well, like they say, it's killing the goose that lays the golden eggs. Maybe this move can get some new weapons system out there a few months earlier, but in the meantime, you're not inventing the technologies which permit whole new classes of weapons systems.

I'll put it in StarCraft terms: you're spending your minerals on upgrading your Zealots, and failing to invest in the pylons and tech structures that would allow you to build a whole frickin' fleet of Protoss Carriers.

Re:Technology

[Anonymous Coward]

It seems to go in a cycle, innovation followed by consolidation. Someone will make a breakthrough somewhere and we'll see the process start over again.

I guess someone important finally watched

[Anonymous Coward]

all 3 Terminator movies in a row and clued in after a night of hard thinking that "Skynet v0.8" was too suspiciouly named to continue to v1.0.

it was an odd arrangement

[Trepidity]

Basic CS research ought to be funded, IMO, but there's no reason completely open-ended CS research should be funded by DARPA---that's what the National Science Foundation is for.

Of course, this cut in DARPA funding is unlikely to be matched by a commensurate increase in NSF funding, which is the real problem...

Twilight of the empire

[Anonymous Coward]

So the gist is that DARPA wants to fund companies, and not universities. And when they do fund .edus, they have outrageous restrictions, like requiring all help on a project be US citizens.

As a CS students, I can tell you: finding hack US coders is easy; find qualified US students who can do research is hard. It's like they don't teach math or science in US schools anymore or something. Kids from Greece or China or wherever come over here, and run circles around US students in formal predicate logic, discrete math, and other subjects that Ken and Barbie found too hard. It's no exaggeration to say that over 70% of all research students are foreign--simply because there are not many qualified US students. (It's a different story if we needed literature or communication students--we've got tons of those.)

America is a country where companies don't make anything anymore. Instead, they just own the IP, and outsource the *production* to China/Taiwan/India. Hell, look at Transmeta, also in /. news today: they are switching to a pure IP model. Exactly what makes use sure that this model is sane for a country? Production capacity is not very mobile, but intellectual talent does not have to stay put in the US. The engineers who invent the IP can just as easily be located (and will soon be born, educated, and working entirely) overseas.

US Companies went through a similar cylce of eating-the-seed corn in the 80s. What happened was they got their asses handed to them by Japan, where R&D was focused on basic science, and not the "short term" deliverables. Now, it seems DARPA is going to try to repeat the same experiment in failure.

Don't get me wrong. This is not the last straw for the US R&D system, but merely one more straw in what has to be the last bundle. It's twilight of the empire, folks. If you're young, start learning another language.

A far better solution is to let all students in US institutions work on projects. (If a project is truly classified, then just use one of the many defense contractors.) When foreign students graduate, most of them (not all) want to become US citizens. What better way to recruit new talented citizens for a country? With the *reeeediculous* DARPA restrictions, many of the foreign students I know are going home. They expect (rightly) that in 10-15 years, their countries will dominate in the industries they've trained for.

Re:Technology

[notque]

Not saying there is anything special about this president but next time try to pick one who has friends in industries you want to see funded because thats how this game works.

I'd rather my president have a combatitive relationship with industry than a friendly relationship.

Re:This Makes Sense

[localman]

You're relatively new to this world, right? :)

We can put a pile of high-tech weapons and defense systems in the hands of our troops. It won't make a spit of difference. The issues there are political and social. Decades of killing hasn't made any progress at all. I just gets worse. If we kill people more efficiently that's not very likely to change.

Why do you think there are so many countries that have been terrorized for decadees? Lack of good enough weapons? I would tend to think it runs deeper than that.

This is different from a regular war where you've got a leader of a cohesive nation invading other nations. In that case you can "win". This stuff is based on centuries of internal religious conflict amont the people themselves. It's unlikely we'll make a high enough percentage of the people there happy in the near future.

Ah well. Let's just nuke the whole area and let God sort them out. Because weapons will help. Right?

Cheers.

Experience of a Governement Contractor

[Paradox]

So, I program for Lockheed, and therefore for the Air Force directly, and I can tell you the kind of feedback we've been getting. I can also tell about the kind of feedback we got when I was hanging around the Computer Security groups at UCSB's graduate labs.

The Government seems fed up with Computers. They need them, they need them incredibly badly, but they can't seem to get exactly what they want. This goes for both contract work and research work. I'll adress it in two parts.

For Research Work: Two major factors are at work here. First is the rule of 80/20. We can do 80 percent of what DARPA (or whatever they're named this week) wants, but that last 20% ("Now make it distributed!" or "Now make it fault tolerant!" or "Now make it cryptographically secure!") needed to make the system usable is really really hard. Lots of research projects have hit dead ends. You expect this to happen in research, of course, but still...

Also, I always got the vibe that DARPA was more than slightly pissed off with us Open Sourcing everything left and right. Maybe it was just us they seemed cross at (and by cross I mean grants and funding tended to shift away from projects with lots of open source offerings), but I've heard other folks doing research mention this too.

I mean, you can easily get the impression that the Government has an attitude of, "You're supposed to be working for us!" Every time a group open sources DARPA-funded stuff (or the components of it, which is usually the case), other people benefit from the research. This may leave a sour taste in the mouth of the accountants over there.

For Contract Work: The US Government's policy is horribly broken. "Cost Plus" contracts may have been great in the 50's for jets and stuff, but we're reaching the point with computer systems and software where we're proving that Design Up Front does not work for large projects.

But, the various millitary branches have so much CYA (Cover Your Ass) paperwork, precedent and process that they cannot disentangle themselves. It's a really bad situation for them, because they have to adapt or die, and they're dying. This is not to say that the Army or Air Force will "go out of business," it's that projects... multi-billion dollar projects... are failing every year now. New projects, huge projects that even a lightweight process would need hundreds of people to deal with, are starting at costs that are so low they'd barely turn a profit for a contractor, because the Army/Navy/Air Force expects to fail.

What I think the Government really needs to do is become more tech-savvy in general. They need to start paying top dollar to hire the best engineers. No more of this "We Give Good Benefits" junk. The Government needs to have its own research groups and they need to be driven by results, technical excellence, and they need to have open-ended budgets (that are limited by results).

Fusion research...

[gnuman99]

Unfortunately, AI is very much like Fusion. It's only 20 years away (for the next century)

No, AI is nothing like fusion. We *don't* know what is required (software-wise) to make a robot alive. We *do* know how to make fusion energy efficient and it was done.

The perception that fusion doesn't work is from the early days of fusion research. Without doing any actual testing, physicists just though if you put the plasma in a magnetic bottle, you get fusion. When they actually done the experiment, they discovered more is going on in the plasma. You can't treat it as a gas. You can't treat it as a liquid. It is kind of a combination of both. Virtually everything in physics with regards to fluid/gas flow, as well as electromagnetism is part of the fusion reactor. Only NOW, after the experiments were done, do we understand WHAT is required to make fusion work and HOW to make it work.

Unlike AI, fusion research has been done. It works. It is here now. All that is needed is money to build a test reactor based on *current* knowledge (no pun intended :), work out final nicks in application of the theory, and then we can build the first commercial fusion reactor.

The obstacle to fusion is not science (or lack thereof), but lack of funding. You see, what people heard in the 60s about fusion, they still think it applies today.

Re:Technology

[Doc Ruby]

The role of government in costly downcycles is to reinvest in stabilizing the cycle. Especially when the cycle has been so integrated with government spending, and when it returns so well on investment. 50% of American economic growth is technology. And American defense depends on retaining our tech edge - so tech investment is an essential role for the DoD. They might have made a more persuasive argument for weaning the tech R&D community from DoD money when it was booming. But cutting it when the DoD budget is booming, and American tech is busting, is to kick this essential industry when it's down.

Re:Technology

[Aix]

Agreed, but the US government should be asking itself whether it can afford to have that breakthrough happen somewhere else. It is extremely foolish (and yet commonplace) to think that Americans have a monopoly on innovation.