DARPA head resigns, moving on to industry | Hacker News Reader

DARPA head resigns, moving on to industry

defensenews.com

76 pointshhs39 comments

To the folks reaching immediately for the “revolving door” observations:

DARPA is one instance where churn is a feature, not a bug. Or if not a feature, at least an intentional design choice rather than an accident of corruption.

Various rules about the way the USG works make it really challenging to innovate in the way private industry can. DARPA isn’t perfect, but it’s a pretty good effort to engineer an ecosystem within the USG that optimizes for “wild idea pursuit” rather than “Spec’ed hourly labor” (which governs the rest of the defense complex). Part of that ecosystem is regular rotation of leadership to bring in fresh ideas and agendas. Another part of that ecosystem is the acknowledgement that you have to spend money on 99 bad ideas before you find the one nugget of gold. Basically, the same observation that governs VC portfolio math but the DARPA is interested in capability unicorns rather than economic unicorns.

When I worked in DC, the joke was that DARPA was the only agency in town that would pull your funding for SUCCEEDING. The reason being that if you succeeded, it obviously meant they had set their ambitions too low and needed to re-frame the work to be impossible before funding it any further.

Revolving doors are a policy anti pattern when the proverbial door separates industry from industry regulators. DARPA doesn’t fit that category.

"The Dream Machine" by by M. Mitchell Waldrop (https://www.goodreads.com/en/book/show/722412.The_Dream_Mach...) is a really good book about J.C.R Licklider, one of the perhaps lesser-known individuals who helped create the modern computer revolution, in no small part as an ARPA (later DARPA) director. By reading the book, I learned that the changing of directors is by design, and an 8 year tenure looks like a problem with finding a suitable candidate to fill the role. 2 years seems more in line with how it was set up. I really recommend the book for anyone interested in the history of computing, the internet, or DARPA.

What does this mean for DARPA / the state of DARPA? I just searched and see that Walker was head of DARPA for about 2 years, which seems like a short time.

Previous heads' tenures:

1. A. Prabhakar: ~5 years

2. R. Dugan: ~3 years

3. A. Tether: ~8 years

compare that to the heads of other government agencies and the so-called revolving door

https://www.opensecrets.org/revolving/top.php?display=I

Are you saying that this is a longer tenure than other industries? I see in the link you shared that "Internet" has 387 people profiled and a high percentage 78.0% (which I don't quite understand, the page seems to be lacking information.)

I haven't dug into the data that deeply but my hunch is this is probably somewhere in the broadly average range of the distribution

Just for example:

> The most prominent example of the Comcast/FCC revolving door is former FCC commissioner and current Comcast lobbyist Meredith Baker.

> Baker, whose views tended to side with the industry even before she went to the FCC, was appointed to to her FCC position in July 2009 and stayed there for nearly two years, cutting her four-year term short in June 2011 to move to Comcast as its senior vice president of government affairs.

https://www.opensecrets.org/news/2014/04/the-comcast-fcc-rev...

(Check out this comment by JonathonW: https://news.ycombinator.com/item?id=21892818)

I don’t know about the head role specifically, but DARPA generally has short tenures by design.

Each DARPA project lasts two years (with possibility of extension). So a typical DARPA program manager will work in that role for just a couple years before returning to industry.

Walker had been acting director of DARPA since January 2017 (eventually formally becoming director in November 2017), so, when his resignation takes effect, he'll have been in the position for three years-- actually a few months longer than Dugan.

Historically, Prabhakar and Tether's long tenures in the position are unusual [1], with most directors staying anywhere from one to four years. This mirrors how DARPA works at the lower levels: the revolving door for DARPA Program Managers (and their programs) is very much real and 100% intentional: DARPA PMs are typically recruited from academia or industry, serve 3-5 year terms and then return to work in academia or industry. This keeps fresh talent and fresh ideas moving through the organization, and provides a sense of urgency and time-sensitivity to their work (successful projects get transitioned either commercially or into longer-term programs under other branches of the military-- they don't languish forever within DARPA by design).

[1] https://www.darpa.mil/attachments/DARPA_Directors_Sheet-web....

Thank you for clarifying! This approach makes sense.

Seems like a smart way to do business. (Or, research, as it were.)

Too many orgs end up resting on their laurels - either getting too focused on research and not executing on technical advantages generated by that research, or too focused on execution and not enough into research.

This seems pretty normal. I worked at a startup company founded by ex-DARPA people and I got the impression that it’s standard procedure there to move in and out of the agency and while in private business receiving funding from the agency.

This is a bit tangential to the topic at hand, but do you believe that their previous experience at DARPA made the company's culture/workflow distinct from other startups in any significant way?

They were very hands off. Probably the same way they operated as DARPA project managers. It often felt that the CEO and CTO watched us from a distance. But they were excellent at writing, presenting and talking to people.

DARPA is pretty neat. Pitty it's part of the defense department and not something shared between them all. Tech isn't just good for war and ads, ya know.

Make it ARPA again in name and flesh!

There are also ARPA-E[1] for Department of Energy and IARPA[2] for Intelligence Community.

[1] https://arpa-e.energy.gov/

[2] https://www.iarpa.gov/

They do have some programs modeled after ARPA, such as ARPA-E. Unfortunately, they'll likely never get the level of funding that DARPA has enjoyed.

> Walker will move on to a position in industry, though a DARPA spokesman did not reveal where.

Any ideas as to where? Lockheed, Palantir, Google?

How about Boeing

With a PhD in Aerospace Engineering, that would make more sense.

> Walker will move on to a position in industry

It's really hard to read this and remain non-cynical. Were any big contracts signed recently?

There are SO MANY laws preventing this sort of thing. I take the training occasionally; just the refresher takes half a day. The full training to be a Contracting Officer’s Representative (a “COR”) takes a full week.

Suffice it to say: having Walker’s experience and contacts will secure him a nice position, and rightly so. But if that position regards a government action in which he had any role he would likely end up in prison, and the contractor would be heavily fined and maybe barred from competing for government business.

There may be laws but are they actually enforced?

US Federal laws are enforced vigourously enough that I think the burden of proof is on those who claim they are not enforced.

Whether the laws are effective at preventing the behavior they are designed to prevent is another story.

This book https://www.amazon.com/Chickenshit-Club-Department-Prosecute... has lots of proof.

That's not a good point of evidence. The laws in question are regarding government employees self-dealing by awarding contracts in exchange for employment/favors later.

>Why were no bankers put in prison after the financial crisis of 2008?

This question is enough to dismiss the entire book outright. The financial crisis of 2008 was caused by a systemic risk modelling failure. This led to banks relying too heavily on the stability of mortgages. Bad business decisions are not criminal, that's why the bankers weren't arrested.

Here are some other questions for you:

Why were no energy industry employees/CEOs put in prison after the 1973/1979 energy crises?

Why were no tech people arrested after the dot-com bust?

Entire industries have systemic failures that impact broad society. That is not evidence of a crime and implying so in a leading question to your book means pandering to a specific political view or just gullible people.

You're conflating issues here. Fraud is still criminal. Some made bad business decisions, some have committed fraud. Some have made money, some have lost. The ones which made bad business decisions lost money. That's not what the book is about.

> That's not a good point of evidence. The laws in question are regarding government employees self-dealing by awarding contracts in exchange for employment/favors later.

This happens all the time. It's called the 'Revolving door'. It's not that hard to prove, it's just that federal prosecutors don't bother instrumenting the cases, unless there is political gain from doing so. The laws in question I was replying to, were 'Federal laws', not specific to conflicts of interest, but even so..

I'm not advocating for the book, I read it and found it both interesting and representative of the recent enforcement efforts by the government which nobody can deny have been sub-par. Most of this is because of the broken incentive structures for prosecutors. Take for example the Equifax breach in which only two employees have been indicted and convicted of insider trading, with penalties totaling a whopping 160k. That's some vigorous enforcement right there.

Another examples of vigorous enforcement is what happened to the tobacco executives after lying to Congress. Or.. other people who also lied to Congress https://www.nbcnews.com/politics/congress/5-people-who-lied-...

So yeah.. vigorous.

Why do you assume (or know) US federal laws are enforced vigorously? I see almost no evidence of this, and an abundance of evidence to the contrary. As another commenter pointed out - there is an entire book describing this epidemic in detail.

>and an abundance of evidence to the contrary

Then please provide us with a little. It's interesting you're able to see evidence of major federal offenses and no major news outlets or law enforcement agencies have done anything about it. Sounds like quite the conspiracy.

I don't even need to bring up obscure cases or examples, just look to the president. Do you honestly think he's going to be prosecuted (I realize he has been impeached) for any of his offenses?

So you don't need obscure examples but the first one you reach for is highly political and applies to one of the most protected federal workers? No one with any habit of studying us law is under the illusion that the POTUS is bound by the same rules as other federal employees specifically because the law itself has many exceptions for that specific role. If there's so many examples then surely you can find another that isn't tied to the Trump administration.

I gave an example which you don't dispute and I referenced an entire book full of specific examples - it seems unlikely I'm going to persuade you of anything (the fact that you made an account just to reply to this comment doesn't make me think you're going to stick around either) since the goalposts keep moving.

But sure - how about the Scientologists being given a pass for tax fraud, extortion, and a litany of other illegal activities? Or is this another unfair example, because Scientologists are also untouchable by regular laws? My entire point is that the "untouchables" are the problem - there should be no untouchable class of people at all. The fact that the elected leader of the US is considered above the law is moronic. He should face higher scrutiny, not less.

https://en.wikipedia.org/wiki/Tax_status_of_Scientology_in_t...

I'm done responding to comments in this thread.

There are laws against quid pro quo (https://en.wikipedia.org/wiki/Quid_pro_quo), but i'd be curious if someone could point out a clear and obvious case of quid pro quo that is documented, which results in real consequences.

Laws are as good as enforcement. See https://en.wikipedia.org/wiki/Revolving_door_(politics) somehow they are still doing it

It would be worse for society in general if he didn't move back to industry/academia. The entire point of the organization is to encourage research from outside of the government via many of these revolving door relationships.

>Were any big contracts signed recently?

Well if there were any signed with the company he took a position with, he's going to need to lawyer up and provide significant evidence that he didn't know about it lest he spend a significant amount of time in prison.

It's DARPA, not Darpa. It's the Hacker News submission title replacement code again, it's not the first time that the acronyms in the titles has been replaced improperly. To those who didn't know its existence...

1. It properly capitalizes the title, making it more pleasant to read, and it also automatically removes the ALL CAPS titles. But from time to time, it makes the title worse by removing the proper acronyms.

2. It removes buzzwords from the title, which is generally good for anti-karma farming, but sometimes it does this even if such use is technically appropriate. For example, "massively parallel" becomes "parallel".

Submitter, please change the title back manually. And moderators, please update your code to recognize more acronyms, such as DARPA, CSAIL, etc.

https://english.stackexchange.com/questions/51924/proper-cap...

> It's DARPA, not Darpa.

Well, that depends on house style; it's not unheard of to set acronyms (in the narrow sense; that is, abbreviations pronounced as words) in initial caps rather than all caps, though that's (IME, at least) a more common style for English publications outside of the US.

Reply on news.ycombinator.com