Lockheed CEO Pitches Pentagon on Subscription Software
news.usni.org
news.usni.org
Federal contracting rules also limit the profit that can be made selling to the government (you can argue this are not effective).
Some 'agile fans' have been promoted up a few notches at Lockheed, that is all this is.
The USG likes to pay for and own source code and weapon system designs.
Maybe. I definitely do not have first hand knowledge of anything.
What I do remember is that a few months ago SpaceX's Starlink was playing cat-and-mouse games with Russian jamming attempts. They had an issue where the Russians couldn't jam an already established connection, but because of some protocol weakness they could jam the establishment of new connections. Within a very short amount of time they patched it and rolled out a fix both to their satellites and the receivers which negated that particular jamming attack.
When that happened Dave Tremper director of electronic warfare for the Office of the Secretary of Defense commented that they envy that level of agility. He said that current way how the Pentagon responds to such challenges would have taken much much longer.
Not saying that this recent news is directly related to that event, but observations like that can prompt the defense sector to offer solutions like this.
more details on that particular case: https://www.defensenews.com/air/2022/04/20/spacex-shut-down-...
space-X hopefully did a risk review. what is the chance this "100 line fix" is going to break something else? what was the in lab coverage? can it be rolled back?
sometimes you say you want agility; but you would really prefer low risk of greater failure.
They're mechanisms for transferring tax money to gigantic employers that employ people across zillions of different congressional districts. It's a doublespeak to avoid having millions of people on a program that would otherwise be termed "socialism" or "welfare".
SpaceX/Starlink, on the other hand, are more interested in shipping and servicing customers and getting revenue thereby than being approved by a constituency and/or the existing military-industrial complex.
The DoD and SpaceX were built for drastically different purposes, and it's no surprise that they have drastically different change control procedures.
There’s obviously a ton of pork and SLS in particular is a known welfare program, but that’s not a fair critique of F-35 nor of the defense sector writ large.
Fortunately we don't have to be very good to be better than our peers.
Honestly I feel like the F35 is a target in the “information warfare” space. While the project is not above criticism, China/Russia/etc have every reason to promote animosity towards the project.
Surprisingly, the contractors retain copyrights and trademarks somehow. For example, apparently you need a license to put an f-22 or an f-35 in a game (or movie etc?), I think those things necessarily belong to the taxpayers. Weird licensing.
As for licensing to put the f-22/f-35 in a game/movie, I wonder how much of that is actually needed from a legal perspective, and how much is "lets just do it anyway".
The name is covered by trademark so you need a license to call it "F-35 Lightning", but I don't think the design or shape of an aircraft is covered by copyright or trademark law, so it should be fine if you just avoid the name.
[1] https://www.nationaldefensemagazine.org/articles/2016/9/8/in...
There is a separate regime that covers useful articles: design patents, which have a much shorter term. The design patents can AFAIK cover things like toys, I don’t know about game assets. You might have to look at the actual grant of patent rights to see what is claimed. I don’t get the impression that the big aircraft manufacturers care about games for plane nerds.
> Agile is a buzzword of software development, and so all DoD software development projects are, almost by default, now declared to be “agile.”
Who doesn’t love Big Agile?
They work with blocks of sprints, defined as 8-12 weeks, "increments".
12 weeks, that's a quarter. Yeah, that's quarterly planning for you. Only this time it's certified Agile (tm).
It's all very serious, and pointing this out is not always as appreciated as it should be.
And they are really picky about finishing. All stories are tied to a charge code.
Moreover, since they're "agile," they no longer have any kind of dedicated testing teams, so developers are responsible for creating and maintaining all of the test suites. But they're evaluated on velocity, not reliability, and velocity is defined as story points delivered, which is measured by tickets closed, but since labor hours are purchased in fixed buckets, the definition of done becomes "we ran out of hours," so in practice, developers just don't write real tests, everything always passes, and nothing actually works.
More moreover, even though the government has pushed all of its contractors to adapt some kind of "DevOps" branding that implies you have broken down the traditional barriers between development and ops, in practice, development and ops are entirely separate contracts run by separate companies and purchased by separate directorates of your sponsoring agency, so in reality, as a development contractor, your work is purchased and approved by an acquisition office that doesn't actually own the operational environment and you're not even legally allowed to directly communicate with whoever actually does own and run that environment. All communication has to go through the acquisition office. So you end up delivering exactly what was asked for, but it doesn't get accepted because the organization that gets to accept or reject changes to production isn't the same organization that was involved with all of your planning and development activities.
I understand why she still works there. She's spent her entire career working on special access and compartmentalized projects, and I worked at this same place on similar geoint projects even before she did. In spite of what Hacker News would have you believe about the quality of military technology, you really do get to see amazing shit, including satellite capabilities I worked on a decade ago that the likes of Starlink and Google are still not close to matching, but I couldn't do it any more and left years ago.
what a pointless burden
I thought everyone was “doing//going agile” like 10-15 years ago already.
Of course government agencies I expect to be last to be agile but there are still companies I find are claiming to start “agile transformation”.
Maybe should not be surprised as much because I also worked on digitalization projects for some big companies that were still pen and paper in 2020…
Bullshit alarms going off full blast on the first paragraph, that's some brazen BS. I didn't even read the rest of the article since they led with that.
This is management being absolutely horny for recurring payments. I don't know if business schools are really pushing this idea now or if it's in vogue for some other reason but everybody outside of management hates it.
[1] Which was a mess. I mean, even in an industry notorious for messes, it stood out, by a lot. I can think of one other program I've seen that gave the 35 LSA/MTA/ILS/ERP/DOA/InsertYourFavoriteMeaninglessAcronym a run for its money.
I've been involved with a previous project of theirs in that space, possibly even one that led to F-35 one. If not for contractual issues, we would have rewritten it better ourselves in less time, without things like finding two years into supposed deployment that a core module wasn't a fit for the problem domain, because they copied it from a completely different type of vehicle.
The new system is just as bad.
I think most senior leadership at Lockheed are engineers and not MBA-types, certainly the CEO has a bachelors of engineering from the Air Force Academy.
This is probably to try reduce the risk of big projects with one payout causing financial pains during down markets. The government is going to budget for a subscription way in advance so you suddenly see monthly income for the next lifetime of the administration.
One business market a lot of people are enviable is furniture on payment plans - you sell an item and get cash every month for the next couple years. Industries where you have such a clear forecast are much easier to plan around (yes, this was something from Business School lol).
I just don't understand why or how we as a society think it's a great idea to let defense contractors behave as if it's a public corporation that needs to increase revenue every quarter. That to me is a huge national security risk - your top weapons manufacturer bases its decisions off how much money it could make versus how it would affect defense readiness....and ex-military turned lobbyists (or board members) see nothing wrong with that.
> When it does work, it works great.
That, unfortunately, does not currently apply the Lockheed dog though..
Let’s assume that this isn’t some grift like “we hope they’ll pay more than what it costs us in maintenance” or “hope they forget to cancel”.
Don’t subscriptions only make sense if you have lots of customers? If you only have one customer, if they cancel their subscription, your product is dead. You also can’t amortize your fixed development costs across many subscriptions. How many customers can Lockheed have, realistically?
This just sounds like a Government contract with extra steps (read: extra $$$)
1. Predictable income / expenses for each party, many defense projects are cut at the whim of politicians. This is partly why we are having trouble ramping up production for the weapons Ukraine needs.
2. The US military is trying to adopt more agile development practices. They definitely want to move away from waterfall software as part of that.
3. Your 1 customer argument falls apart as soon as you talk about per-seat pricing. The US mil can buy more or less seats as the size of the force changes or is reallocated. It also sets Lockheed up to charge for drone connectivity. The military benefits because they could opt not to buy more if it doesn't turn out well.
note, I have not delved into the details of this article, but base these hypotheses on the developments at large I've seen elsewhere.
That said, I understand why people hate subscriptions. I'm not claiming there are no downsides. And also, yes, as you pointed out, this is the military/industrial complex, so who knows how this will play out.
> 1. Predictable income / expenses for each party, many defense projects are cut at the whim of politicians.
> 3. Your 1 customer argument falls apart as soon as you talk about per-seat pricing. The US mil can buy more or less seats as the size of the force changes or is reallocated. It also sets Lockheed up to charge for drone connectivity. The military benefits because they could opt not to buy more if it doesn't turn out well.
Either they have predictable revenue or the government can cut seats whenever they want. How can it be both?
And for the same reason it doesn't solve the 1 customer problem. If the Marines want something and nobody else does, the contractor has to be charging enough to not go out of business even while only selling it to the Marines. So the rate per seat has to be unreasonably high, which is bad when they're only selling it to the Marines, but even worse if the whole government wants it and they're still paying the same rate.
Say the government agrees to buy 200 planes, so the manufacturer prices the planes in a way that they recoup their development cost (which can be many many billions of dollars), and in the midst of the contract the government decides to say "we don't need more planes" and the manufacturer is screwed.
With "plane as a service" however, should politicians decide to cut delivery of more planes, they can at least recoup their cost with the license fees that apply during the plane's usable life time - it's not like the US government will suddenly go to Eurofighter, and it's not like that there will be much US-based competition in building fighter planes.
In the end there are two problems, the one being that the US government has gone down the drain regarding reliability with multiple shutdowns and a barely avoided default in the last decades and a re-allocation of military budgets looming around the corner, and the other being that there's zero competition to Boeing and Lockheed left so these two companies can milk the government at will.
How does that solve it? If they can charge high enough service fees against 200 planes to recoup their costs then either the license fees aren't fixed per-plane -- in which case the government is effectively prepaying for the lot of them regardless of whether they're any good -- or they are, and then the government is getting soaked in both cases.
> the one being that the US government has gone down the drain regarding reliability with multiple shutdowns and a barely avoided default in the last decades
This was never an actual problem, it's just political posturing. They're not going to default. They just find it politically useful hold out until the last minute. And when a budget ultimately gets passed the contractors get paid.
> the other being that there's zero competition to Boeing and Lockheed left so these two companies can milk the government at will.
This is the actual problem. They should both break these companies up and make it easier for other companies to bid on government contracts.
Whatever excuses they come up with at the root of it is the idea that they will make more money by doing it this way.
If both parties are happier with the arrangement, and it leads to more agility, what exactly is wrong with it. The military is projected to decrease (as % GDP)[1] over the next decade. If they spend more on software and less on overpriced hardware, that might be a good thing.
[1] https://www.statista.com/statistics/217581/outlays-for-defen...
> Regardless of the reasons they spout
They are saying the same thing the military is. The US Military is actively trying out new ways of purchasing, because the old model is not working and not agile enough for today's environment.
I trust that the people leading the military know what they need and that Lockheed cannot force them into this. Everyone is ideating right now, the UA War seems to have woken the West up. No need to derail an idea just because it comes from Lockheed's CEO
If Americans continue to accept non functional government shutdowns, why shouldn't government contractors get in on the grift.
Also, 100$ hammers are basically upfront subscriptions to the means of production so this may actually be more transparent payment type.
I've seen one company go in before getting a sponsor agency, and they got their SaaS environment built and certified as FedRAMP Ready (and listed), but never actually got an agency customer and didn't go the JAB route, and all the money they spent went down the drain.
It's really tough to justify FedRAMP if you don't have the political connections to get a sponsor, and that keeps the space nepotistic and stagnant (as you said).
The US's runaway defense spending isn't just pricey missiles, planes, and salaries; a big chunk is now software. It's a subtle issue, though, and I have yet to hear any politician address it meaningfully.
Many govt contracts are upfront huge capital costs, and then long maintenance contracts which you are locked to that vendor for x years term even though you "own" the software.
Subscription just spreads those fees somewhat evenly across the term
This means that if you're buying something like trucks, it's generally better for the government to buy ordinary trucks from the market than to operate a public sector auto manufacturing industry that would become bloated and inefficient because it captures the government and isn't subject to competitive pressure. Even if the trucks are critical to some government function.
This works fine as long as the bidding process is open to anyone, because then the government just buys cars from carmakers and uniforms from tailors and you ideally don't end up with a voting block of government-exclusive tailors insisting that they need to be able to retire at age 55 with full benefits. Private tailors who insisted on that to a private contractor would be refused because the contractor couldn't give them that while submitting a competitive bid for the government contract.
The problem comes when the government starts making it harder to bid on a government contract, because then the only bidders become companies that specialize in government contracting. At which point the contractors are just government employees by proxy -- they have no private customers, they only work for the government -- so then the government just gets captured again.
What should be happening is that the government puts out bids to design a tank, anyone can submit one, and the winner gets paid for the rights to the design. Then the government takes the design and put out bids to build each of the parts for the tank and gets bids from all of the companies that normally make water pumps and hydraulics for cars and homes and construction equipment. Then they take bids to assemble those parts into a tank. Then they take bids to do maintenance on the tanks. Anyone can bid who is qualified, and because the contracts are each small and as generic as possible, lots of ordinary businesses are qualified.
But instead they put out a single contract to do all of that, end up with a concentrated market of companies who specialize in government contracting, and you lose the benefit of market competition because you no longer have any.
Alongside the above 2 metrics, I'd also add that a lot of the graft is not directly tracked. For instance play ball with the defense industry while in office, and one can look forward to fatly compensated lobbying or advisory positions after leaving office. Similar to how presidents who play ball with banks/wallstreet get to go get paid $500k/speech for say-nothing 30 minute chats.
There's a reason congress has never seen a war it didn't like.
[1] - https://www.businessinsider.com/congress-members-are-trading...
[2] - https://www.opensecrets.org/industries/indus.php?Ind=D
I'm also curious what makes weapons research and production an "essential government function"? Couldn't the same be said about food, technology and bunch of other things?
> I'm also curious what makes weapons research and production an "essential government function"? Couldn't the same be said about food, technology and bunch of other things?
Sure. There's a lot of stuff I wish we would nationalize, personally: basically everything that could reasonably be considered an essential, from healthcare to education to food to broadband, leaving aside things like video games, luxury cars, jewelry, blah blah. What is or isn't an "essential" is a necessarily moving target, evolving with society.
However, I also have zero faith in Washington to execute anything remotely like that. Canada, maybe, or one of the Asian or Nordic governments...
> Sure. There's a lot of stuff I wish we would nationalize, personally: basically everything that could reasonably be considered an essential, from healthcare to education to food to broadband, leaving aside things like video games, luxury cars, jewelry, blah blah. What is or isn't an "essential" is a necessarily moving target, evolving with society.
I for one would prefer essential services not to shutdown because of politics like what almost happened a week ago.
> However, I also have zero faith in Washington to execute anything remotely like that. Canada, maybe, or one of the Asian or Nordic governments...
Canada and europeans have historically failed to keep up with US despite (some would say "because") more government-centric approach. Exercise for the reader - which NATO countries are meeting its declared 2% millitary spending target if you don't include things like pensions.
Yeah, but also seems to happen with big corporations. Power corrupts and all that. It's only human =/
> I for one would prefer essential services not to shutdown because of politics like what almost happened a week ago.
Me too. I'd prefer the government not just randomly shut down every time the Republicans (it's always them, isn't it?) feel the need to virtue signal about some pop culture thing or another. Somehow most governments in the developed world manage to operate day to day without this constant threat. American exceptionalism rearing its ugly head again...
> Canada and europeans have historically failed to keep up with US despite (some would say "because") more government-centric approach. Exercise for the reader - which NATO countries are meeting its declared 2% millitary spending target if you don't include things like pensions.
For real? Who cares about NATO spending? Meanwhile they have functional governments, holidays, healthcare, education... the things that every citizen can use every day, and that you can afford when you're not trying to build up a 100x military that nobody else wants to pay for.
Highly centralized planning creates an enormous vulnerability for corruption, because it creates a small number of high-impact corruption targets.
To quote some US officers - "Soviet equipment works well - in Ukrainian hands".
Something something maintenance and logistics are king, even if artillery is the queen.
US officials say a lot of things... I'm much more inclined to believe actual VSU fighters who say it's better.
Instead We pay 2x the cost of an employee as a contractor, and the company keeps half.
Rarely have I seen such a portrayal of villainy. LM's efforts here seem nearly treasonous, with their attempts to thwart open .mil systems & to sell the American people permission or capability to use the technology they'd funded and purchased, by granting zero rights or control over that technology. The last couple of decades of Lockheed Martin retaining the technology &systems entirely should be a clear enough lesson in what America & (forgive me) anyone with an ounce of patriotic blood should never let happen again. Yet that sure sounds like what's happening here, again, even more boldfacedly.
> General Dynamics and Lockheed Martin considered much of the data and equipment on the LCS proprietary — a problem that the GAO has identified throughout the military. As a result, only their employees were allowed to do certain repairs, former officers said. This sometimes meant that contractors would go overseas to help, adding millions in travel costs and often delaying missions. The Navy recently purchased some of the data. A Navy spokesperson would not disclose the price "due to proprietary reasons."
> Watson and others spent much of their time escorting contractors while on board because so many areas on the ship were considered classified, reducing their ability to do their own jobs, according to interviews with multiple officers who had served on the LCS.
Never have I read such a corrupt inexcusable sad practice in military procurement ever. This is stagnant bullshit that prevents adaption, prevents proper response to situations. We must have open systems architecture, where the government decides how things interoperate, and acquired systems don't dictate down but provide upwards to the defenders.
And it seems like that lesson of what to never allow again has largely been learned & broadly accepted. Almost every faction of mil seems razor focused on open systems architectures, on intercompatible. It top down systems. On making sure this graft corruption & difficulty in doing anything without endless reams of permission & defense contractor hours never happens again. Not only should Lockheed not be able to charge a subscription, they should never own the software funded by & produced for American defense ever again. Open core should be the rule of the realm. It a core threat to American military might to let defense contractors own & keep the special sauce that makes weapons run, and to be the deciders for how defense technology is ongoingly used and adapted. If America pays for the development, America should own the output.
There's similar obvious revelations that have happened elsewhere in the American system. It's not always fully respected or done well. But in principle NIH for example has open access requirements. In principle Americans are entitled to the benefits of health research done by the state, and it's not just sold off or left to be controlled by whomever. Yet Lockheed Martin keeps wanting to be the owner, wants to be the decider, wants to interrupt right of first sale & be forever entailed to & in control of whatever happens next, after some defense tech hardware is purchased.
Yeah this is the playbook for the last half-century. Loot the nation, privatize everything. Procurement and supply are run by the same elite class in a revolving door system. Reagan was the greatest scam artist of all time, and people still worship him (though that is finally changing).
The problem we have right now is that there are so many laws that the only people who know what they do are the lawyers for the industry they regulate.
So the problem is that 80% of them are inefficient and competition-destroying bad laws, but who do you ask to figure out which 80%? The people who know the answer are going to tell you to repeal the ones they don't like, which is most certainly not the same thing.
Reagan saw that we have too many bad laws we need to get rid of, but failed to solve the problem of determining which ones they are. That's the problem we still need to solve.
>> They would be working with a common architecture on existing platforms
>> Taiclet added that in this “big tent” approach to Pentagon buying services for a specified time means the commercial side needs to move away from “vendor lock”
>> He added Lockheed Martin is working with Northrop Grumman to merge their data.
Sounds like they are advocating for something like what you are talking about, not sure how you are reading it as the opposite
But still expect to forever keep control by licensing out subscriptions to the technology they are paid to develop.
It's unclear to me what control they really intend to give up. They can give up ("merge") some data with Northrop Grumman, but are they actually allowing outside powers control over the systems? This feels like an operation to try to maintain some status quo by having nice koombaya words while not changing the actual sharing of power.
One tragedy here is that some of the "open" APIs pushed by Lockheed et al are really about as open as Microsoft's "OpenXML".
https://www.nationalreview.com/corner/f-35-latest-problems-e...