I’ve recently heard some folks state that "we know that US military tech runs 10-15 years ahead of the private sector. . . " Really? Is there solid real evidence to support this? With real examples. Or it this all UAP type speculation?
One example I know is that Space X engineered how to land a rocket upright. That’s not something NASA ever accomplished. Not sure they tried though.
Plus, what does that even mean or look like? 10-15 year ahead? In what areas?
I can’t answer the wider question, but as far as your example is concerned I’d like to point out that NASA is a civilian agency; it’s not part of the military.
OK. I generally think military/government. But, there is a clear line. Mostly.
I’m not really sure what that question means. That’s like when companies advertise their flashlight or cell phone covers as “military grade”. That’s not an actual thing AFAIK.
There are companies like Boeing, Sikorsky, and Lockhead Martin that design and manufacture military systems like aircraft and tank engines and whatnot. They also design versions of various products for civilian applications. AFAIK, one isn’t necessarily more “advanced” than the other. They just tend to have different specifications based on their expected operating environment.
Sometimes military systems make their way to the civilian market (like GPS navigation for example). I believe (at least initially) those systems were degraded to make them suitable for driving to grandmas house vs targeting a cruise missile there.
DARPA also conduct government-funded research for potential advanced military technologies that generally don’t have civilian counterparts (like stealth fighters, bio-material eating robots, and houses that grow themselves (actual DARPA projects) ). It sounds like advanced sci fi shit, but I’m sure Silicon Valley is researching equally advanced stuff for the civilian market.
I mean keep in mind the private sector can often hire the best and brightest at whatever cost with stock options and all sorts of benefits where comp for most government research is limited.
In my experience it’s the opposite – for the US military, most technology can’t just be effective, it has to be extremely reliable. That means NO CRASHES. No screens of death. No unexpected outages. The 688 class submarine I was on had a computer-controlled nuclear reactor, and the computer was (IIRC) an 8086 processor (or its equivalent), and operated well into the 2000s. We said onboard that we knew this computer chip so well we could track every single electron, and it never failed on us – never shut down, never black-screened, etc. Getting the latest and greatest might be nice when it works, but what if it fails in the middle of combat, or at 800 feet below, or 10,000 feet altitude? Better to have something that you can be absolutely sure isn’t going to fail on you at a critical moment.
Same for spacecraft. The computers and chips in spacecraft were years out of date, but they worked.
Not quite true. The DC-X program, eventually handed off to NASA as DC-XA, successfully landed a rocket upright (multiple times) back in the 90s.
What they didn’t accomplish was turning it into an operational orbital launch vehicle, going on to land over 250 times, lowering cost through that reuse, while being among the most reliable rockets in the world as well.
There are some problems where you can get a little ahead of private industry by putting enough money into it. But it’s rare that this results in a commercially viable product. Another example are GPS receivers. GPS was developed for military use, but receivers were very expensive. It’s only once private industry was able to cost-optimize it that it became ubiquitous.
For other problems, it’s not true at all, like semiconductors. Apple gets access to the same tech that the military does. As said above, military use is likely to be behind state of the art, since they want reliability, not bleeding edge.
Also, chip fabs are very large, very expensive, and require some of the most complex, precise high technology on the planet, some of which tools have exactly one source. There is no chance that the military has squirreled away somewhere a whole different secret chip production infrastructure that is 10-15 years ahead of current commercially produced chips. That would be at least a Manhattan-project level infrastructure, possibly beyond.
Actual military stuff complies to Mil-Spec standards. But the more generic term “military grade” is an actual thing.
For example, integrated circuits come in 3 different grades: consumer grade, industrial grade, and military grade. Consumer grade has an operating range of 0 to 70 deg C. Industrial has an operating range of -40 to 85 deg C. And military grade has an operating range of -55 to 125 deg C.
Most of the “military grade” cell phones that I have seen don’t actually claim to conform to any specific Mil-Spec standard. Instead, they’ll conform to something like IP67, which means that they are dustproof and waterproof. To claim conformance to the spec, the cell phone can’t have any ingress of dust during an 8 hour test (time may vary depending on air flow), and has to be completely waterproof when tested at a depth of 1 meter for 30 minutes. An IP69 phone would be similar, except that it has to tolerate actually being sprayed with high pressure water from multiple angles instead of just being dunked for 30 minutes.
Technically though, IP67 is an IEC standard. It’s not a Mil-Spec standard. There probably is an equivalent Mil-Spec standard, though it might be slightly more strict in some areas which is why manufacturers support the IEC standard instead.
That’s mostly because of the added time and effort it takes to radiation-harden the processor. The 8086 that @iiandyiiii mentioned isn’t an off-the-shelf 8086 made by Intel. It’s a special radiation-hardened version that took years and oodles of money to develop.
No, but when I worked in the defense industry, we had absolutely no problem designing a completely new custom chip and paying out the wazoo to have a few thousand of them made. Civilian companies are much more concerned about the bottom line and wouldn’t dare do that unless they had a good financial reason to do so.
A lot of the things we were doing were bleeding edge. My security clearance wouldn’t even let me into the area where they were developing the synthetic aperture RADAR antennas. We had two distinct customers, military and civilian. The bleeding edge stuff was only for the military. The civilian versions of things (like the F4 phantom that we were selling to the Japanese) was only permitted to have last generation technology. F-16s for foreign sales only got the older analog signal processors. Only the military got the high-tech digital processors.
When you are designing something like a fighter jet, you don’t design with current technology. You make guesses about what technology is going to become available and design for that. Your enemies are doing the same, so if you design with what you currently have, your stuff will be years behind the bad guys stuff. Guess wrong about what will be available when, and that part of your project is over budget and behind schedule. But that’s how military development works. In contrast, civilian development is significantly more focused on the bottom line.
On one project I worked on, we got a brand new CPU chip that wasn’t available to the civilian market yet. It came on a development board so that we could start writing software for it. The actual chip wouldn’t be in production until part-way through our development cycle, and the civilian version wouldn’t come out until well after that. The government has deep pockets and is more than willing to pay out the wazoo in a way that most civilian companies aren’t. We got the chip first because we were willing to pay for it. I think Apple was the second customer to get the chip. They also paid out the wazoo to get ahead of other computer manufacturers.
Of course there are some civilian companies like Boeing’s aerospace division that are also on the bleeding edge, and they can also pay out the wazoo for things because ultimately it’s the government that is footing the bill.
Everything that I have worked on in the civilian world has been nowhere close to the bleeding edge. We’ve always designed with what we have currently available. That 10-15 year rule isn’t an absolute, but it’s definitely describing how a lot of things work.
Also, the most advanced military hardware can take years or decades to develop, so even though it was state-of-the-art when spec’d originally, by the time it’s operational the hardware may already be dated.
Starlink is a better example than vertical landing. The military never had any chance of building something like Starlink, but now that SpaceX has, they are leveraging it.
Drone technology of the type used in Ukraine is mostly coming from private industry. The defense industry makes large UAV’s, but the ones you see dropping grenades and such are often just DJI phantoms or homemade drones. Anduril (Palmer Lucky’s company) is making sophisticated military drones. The military is way behind on small drone tech.
Computer controlled! La-de-da!
The reactor on my Madison-class SSBN was controlled with mag amps (a few systems had been upgraded to circuit boards with discrete transistors, but they were crap, and we certainly had no ICs for the reactor). This was years after Macs and PCs were widely available consumer products.
Our torpedo fire control system worked with gears and pulleys (and worked very well, I might add).
The USG is ahead in certain specific areas by at least a decade or more but mainly in areas where there isn’t the civilian resources to drive advancement but there is government resources to dump into classified research.
Two Three specific examples that come to mind:
As Bruce Schneier put it:
So, how good is the NSA at cryptography? They’re certainly better than the academic world. They have more mathematicians working on the problems, they’ve been working on them longer, and they have access to everything published in the academic world, while they don’t have to make their own results public. But are they a year ahead of the state of the art? Five years? A decade? No one knows.
It took the academic community two decades to figure out that the NSA “tweaks” actually improved the security of DES. This means that back in the ’70s, the National Security Agency was two decades ahead of the state of the art.
Today, the NSA is still smarter, but the rest of us are catching up quickly. In 1999, the academic community discovered a weakness in another NSA algorithm, SHA, that the NSA claimed to have discovered only four years previously. And just last week there was a published analysis of the NSA’s SHA-1 that demonstrated weaknesses that we believe the NSA didn’t know about at all.
Maybe now we’re just a couple of years behind.
edit: 3. In 2012, the NRO gifted NASA two surplus spy satellites to be repurposed as deep space telescopes and it revealed that the NRO’s capabilities in the 90s far exceeded that of NASA when they were building the Hubble.
NRO says the telescopes which have high tech lightweight mirrors far more advanced than Hubble’s were manufactured between the late 1990s and early 2000s.
What both have in common is that the work product remains classified and there’s a strong separation between the classified and civilian world. This is distinct from things like fighter jets or rifles where the tech is deployed out in the open and there’s much more of a crossover between the civilian and military contractors.
Comparing the military with private industry is a very wide comparison. Private industry is a lot more than consumer goods. Most of the military stuff is made by private industry. The military has little to no in-house capacity. Whilst there are a lot of pure-play defence contractors, many more do commercial work as well. In the US there are mandated tech transfer rules that prevent tech funded by the government leaking out into the commercial world. Not just for security reasons, but to avoid the government essentially funding a technical advantage for one company over another in the commercial world. This can lead to weird relationships between divisions of a company. Not all countries do this, but many do. So you will see the government, via different arms, fund commercial tech development with a clear objective of this work benefiting the defence sector. Not just DARPA’s occasional madness, but big ticket items like next generation chip fab technology.
If you peel back the layers of a lot of niche commercial endeavours you can find some serious bleeding edge technology. But commercial realities mean the budgets are rarely of the scale of military programmes. Big science is another. The LHC is ridiculous level technology, but unless you stand in the way of the beam it isn’t exactly a weapon. Something like AMSL’s extreme UV litho machines are science fiction levels of tech. At well over a 100 million each, you would hope so.
A clue as to the overlap of technology comes from ITAR controlled technology. ITAR covers stuff that is just plain military, but also dual use. Uses where there are legitimate commercial applications as well as military. You can go to your favourite electronics parts supplier and discover that some parts are so covered, and are subject to regulations controlling export. (And these regulations have teeth.) For instance thermal cameras beyond a given resolution. Or FPGAs beyond a given capacity. FPGAs are a good example of where military applications and budgets swim effortlessly past most commercial applications. The biggest of these come in at eye watering prices (think a really nice car) and are the darling of anyone working in signals processing. But even medium sized ones come with a dual use risk, leaking out to less friendly nations they can enable quite unpleasant weapons systems.
The same mostly applies to space satellites - they take years to develop, then possibly sit around for a year or more waiting for the optimum launch window, then years to get to the outer reaches - and as others point out, have to be hardened versions of the electronics. OTOH, the tech has to be “good enough”. The total electronics probably weigh about the same as the case they’re in, and often don’t need to be super hi-tech. One place I worked at was still using 8080 and 6502 process controllers - they did the job. It’s only when spare parts become an issue that an upgrade is necessary - but then I worked at a place that still used a program based on ACCESS2000 almost until 2020 because it worked. Another was running a program on Windows XP. The military especially, but civilians too, know - if it works, don’t …mess… with it.
Another issue is battle-hardening. A reporter in Gaza documenting the slaughter and destruction mentioned that at one point the IDF stole his drone. They were filming overhead shots and the drone suddenly lost contact and flew away. Ukrainians mentioned that the Russians are using microwaves and other jammers to interfere with drones - fortunately, they could be programmed with a path; so the Ukrainians would launch them, lose contact partway through the mission, then reconnect on the way back. Presumably military grade drones would need to be proof against that.
(Although there were news stories that the original UAV’s watching Afghanistan did not have encrypted video, and that the Taliban could use simple receivers to see what the Americans were focusing on…)
Somewhat similarly, NAVSTAR GPS used to broadcast an intentional error of up to 100 meters in a system called Selective Availability before discontinuing it on May 2, 2000, not because it wasn’t more precise but in order to restrict the availability of that precision to the military. Unlike your examples though, it was openly known that the military had greater precision available and civilian signals were being deliberately degraded.
In undergraduate school, we could buy inexpensive overruns of digital logic ICs that had been hardened against radiation for our labs.
In almost all cases, no. The military is extremely conservative and wants proven tech that can be churned out in large quantities by the lowest bidder.
When it appears that applied science is close to breakthroughs with military significance, however, they can throw money and Top Secret stamps at the problem and beat the civilian sector to the punch. See the Manhattan project, GPS, FLIR, UAVs, etc.
And of course, what does it even mean for a military death robot to be “10-15 years ahead” of the private sector? Maybe GM could have built one ten years before Raytheon did, but why would they?
About ten years ago I read about the comparison online aerial/orbital imaging resolution like Google Earth compared to the best the military had at the time. I don’t recall any actual resolutions but the military was far better.
I found this old photo in the government archive. I don’t know if it has any relevance to this discussion.
There is currently a hilarious kerfuffle in the chip industry where the US is trying to block China from buying high-end microchips and everyone is shocked that China is responding by starting to build their own chip building supply chain from the ground up. Probably won’t take them long to catch up with the “banned” tech, in the long term the ban doing nothing but drying up a market for Qualcomm, nvida, et al.