The Manhattan project was able uranium purified to get a bomb working without PLCs and networks.
All else being equal it's more expensive, but not compared to being unable to keep your network safe from attack. You need to reconfigure the line? Send eight techs down one by one to set new parameters and octople check the first tech's work.
By enlisting thousands of young women to watch the knobs and dial buttons.
https://outrider.org/nuclear-weapons/articles/women-who-buil...
They didn't even know they were building the bomb. They just had to enter the room, dial knobs back-and-forth for 8-hours/day, and then go home.
Today, that sort of stuff SHOULD be automated. It made sense back then because there weren't computers, but... you really can't just get people to balance knobs 8-hours / day anymore.
-------
EDIT: https://www.energy.gov/articles/five-fast-facts-about-calutr...
Seems like Department of Energy has some historical articles on them too.
So yes, they can probably afford to have a bunch of people tweak knobs in a factory all day.
So the idea might not work for Iran, but it probably would for North Korea.
In the US, I can’t afford a full time maid. In India, I could probably afford 10.
Or more pointedly, in the US you can afford a car far earlier than you can afford a maid. In India, it’s the opposite.
It‘s why, for example, middle class households in the US don‘t have live in maids, because rising living standards made that unaffordable at a middle class income level.
An interesting aspect of poorer economy I’ve noted but haven’t seen anyone discuss otherwise is you can live a much better life being poor in India, than you can in the US — the really low income economy is far healthier (the quality of everything degrades continuously). I’m sure it’s not a novel insight, but I don’t know what others would call it.
In the US you pass under the minimum wage bar of value production and you’re immediately in a tent under an LA bridge — with everyone else ranging from $7.99/hr to $0.01/hr. Of course an economy exists to support them anyways, but it’s largely underground, illegal (bypassing safety and regulations, e.g. selling very old cars; selling cigarettes individually) and actively zoned into oblivion
But what would prevent this system from being compromised?
Seems pretty easy for an adversary to "inject an exploit in these controllers" by paying a few of them to keep the knobs in the wrong position for a bit more time than strictly necessary, or to flip a switch earlier/later than ideal.
Auditing this is seems worse than checking the code for an old-style, non-networked, PLC.
Legend has it that Stalin knew about the American bomb before Truman knew about it. Hiring literally thousands of high-schoolers without much job training has the advantages of being cheap back then, but you don't really know who was or wasn't a Soviet Agent.
Similarly: having literally thousands of modern techs build a bomb manually would only turn the computer-security problem into a human-security problem. Who do you trust? Who do you keep track of?
There's something to be said about trusting fewer people: and checking those few people more vigorously. That simplifies the human-security portion of the puzzle, even if it complicates the computer-security portion.
Why not?
It's interesting that the Manhattan project had to build everything themselves - which if Iran was NOT using off-the-shelf parts and instead was developing everything in-house they'd be less susceptible to certain types of attacks.
This article says they used air-gapped Siemens industrial control systems:
https://www.yahoo.com/entertainment/revealed-how-a-secret-du...
It needed a coordinated 3-axis motor servo motor control, which we initially planned to use PLCs for, however after looking around and seeing the price tag of the stuff that matched our required feature set, we decided to build our own. We were up and running within 2 months, and I can tell you, writing code in C for the control loops is infinitely preferable to whatever monstrosity PLCs can be programmed in. I'd say this is a much more complex scenario than running a single motor at a certain speed depending on some sensor readings.
High tech combat gear with VR and exoskeletons is more of a gun dealer's wet dream(e.g. costs a ton, breaks often) than anything that usually gets deployed.
https://en.wikipedia.org/wiki/Battlestar_Galactica_(fictiona...
I can imagine all of those systems need to communicate within their domain somehow, so they need to be connected. You could make a point that they have “not been connected to the internet”, but then it only takes one node to become a bridge between WAN and LAN and you are Pwned.
It’s even worse now because there’s a tendency in software development to discount security this way: “we don’t need to think about security because we have a firewall”.
I don’t think this is a valid scenario.
It's pretty hard to buffer overflow
if (read_gpio(Ignition_Enable_Pin)) {I agree that this scheme doesn't protect against hardware implants modifying lines, but the threat model discussed thus far was around propagation of software implants throughout a system.
The show created an aesthetic like an old submarine, even using set pieces salvaged from other submarine movies. That was an aesthetic choice, making everything look cluttered and cramped and slow, increasing tension in exactly the way old submarine movies used to. In those movies, the computers weren't connected because they didn't exist. Actions were coordinated by phone.
It's a bit contrived that they had computers but not networks. It was just a conceit of the show, and actually it was really mostly just in that one episode. But the aesthetic carried through and made for great atmosphere.
[1] https://en.wikipedia.org/wiki/Battlestar_Galactica_(2004_TV_...
Reverting to old methods might not work so well when you face that situation.
"Kids these days etc."
Parts of the Russian government have reverted to using mechanical typewriters in lieu of computers, which are too susceptible to malicious hacking.
Also, the US Navy started instructing enlisted seamen how to navigate at sea by the stars, in case GPS stops working.
I think a potential lesson of the story is to use technology that you understand at least as well as your enemy does. The communist bloc maintained registries of typewriter fingerprints so they could identify which typewriters were used to reproduce samizdat; I think it’s fair to assume that the Russians understand typewriters very well.
Is the threat model there that someone hacks the GPS signal, or that someone shoots down the GPS satellites? Lots of countries have ASATs these days.