7 ms·

Sure, but that's not the issue. That problem can be solved in a hackish fashion for a LOT less than $200mm and perhaps even in an elegant fashion for less. If

Sure, but that's not the issue. That problem can be solved in a hackish fashion for a LOT less than $200mm and perhaps even in an elegant fashion for less. If every Predator gets a 100W wideband transmitter and SOME of them get instead of weapons the equipment necessary to process the reflections and do the calcs then the others can get that target information and blammo. If there are 100 Predators for every F35 in the combat zone and they ALL turn on their transmitters simultaneously at some kind of interval you're never going to know which one is carrying weapons and which one is command and control. So you can't do prioritized targeting and thus you're shooting blind. Yeah you can shoot down some of them but once you're out of missiles that's it. Do we have to worry about 3rd world countries being able to muster this kind of response? No. But there are plenty of industrialized countries that could, and they could bleed us dry one $200mm plane at a time.

greedo 12y ago
Again, it's a cost issue. If you think you can build a super Predator with some type of targeting radar, and a swarm type of AI for under $10M, you're not being realistic. Current build Predators are already over $4M a copy, and even with this fantasy equipment you've described, wouldn't stand a chance against a decent human in an F-15. It's important to understand the limits of radar too; it's not like in a video game where you get this god view of the airspace around you. You are always making tradeoffs in flight about where your radar is focused, it's range, and how many targets it can track while scanning. That's why the US and it's allies rely so heavily on an AWACs system. UCAS aircraft that can drop bombs in a contested airspace are going to cost north of $50M per copy when all is said and done, and UCAS that can do A2A combat will be double that. Once you start putting high performance jet engines in drones, the costs start to rise. Same with avionics. Today's drones are very very simple aircraft compared to high performance fighters.
- msandford 12y ago
  Agreed, cost is what would drive using drones instead of high performance fighter aircraft. I am not suggesting that you build unmanned F-15s for the $4mm each that you can build a Predator for. That's patently ludicrous I agree with you! So no fancy engines, nor fancy avionics. Also don't bother with swarm AI, have people on the ground handling targeting. Don't develop a billion dollar solution when $50k/year (roughly) works just fine. What I was suggesting is that you build a bunch of aerial SAM launchers. Take a Predator, put something that can receive targeting info on it, arm it with air to air missiles. Deploy many of these in the sky. Then for every 10 missile mules, you put up a single command and control Predator. It has the fancy equipment to ensure that you can actually do targeting. Then put a transmitter on all Predators which can be turned on such that you can't detect which Predator is the command and control one. Let's suppose that you could create this fleet of 11 for $4mm * 11 + $2mm * 10 + $20mm * 1 = $84mm This is very conservative in my mind because at this point we're basically paying US prices for the gear. Again, we're not going for the fanciest, best, most high performance stuff ever built. Use old designs, don't worry about the best survivability, don't make them too rugged and thus don't drive up the cost. We're going for The Innovator's Dilemma style worse-is-better so long as it's functional enough to mule up the ordinance. You can now afford at least 2x of these 11 unit Predator swarms for every 1x F-35 that can be deployed. An F-35 only carries 10 air to air weapons and that's when loaded with nothing but. The odds they get both command and control Predators on every engagement is statistically slim provided you're talking about at least a few dozen engagements. That means that just playing a numbers game you're going to shoot down the F-35s and because you've got so many damn drones in the air enjoy quite a bit of air superiority. It's not exceedingly difficult to imagine that if the sky was thick with these things that the F-35s might not get all their ordinance away before being shot at making the grinding more painful. Could the US manage such a feat of engineering discipline and fast-track development? NOT A CHANCE!! If you're thinking that this is unrealistic based on how things work in the US you're 100% correct. But to suppose that no other industrialized country in the world could embark on such a venture and succeed? If you believe that please stop reading HN because startups could never work/win against the entrenched giants who have all the advantages.
  - greedo 12y ago
    It's not a matter of startups, or entrenched giants or any such thing, or the Innovator's Dilemma. It's a matter of communications, cost and performance. A Predator can't fly as high as an F-35 or F-22. It can't maneuver nearly as well either. Can't fly as fast. It can't be refueled in flight, so it's range is lower than the F-35. So you'd have to hope that the missile it carried could compensate for these deficiencies. However, missiles are very finicky, and have launch parameters that help them improve their odds. The military looks at it in terms of PoK (probability of kill). So to be cheap (so you can afford to have swarms) you have to give up performance. This will affect PoK. Next you have to have robust, un-jammable commlinks. So far, no one has those. That's going to be expensive to develop. Humans can make decisions when the comms are degraded, or non-existent. AI has a long ways to go, and AI is expensive. To remain cheap, you'll have to give that up. With you napkin/elevator calculations, you would be trading 22 Predators for each F-35. 22 Predators is $88M, and the F-35 will eventually drop close to $100m as it reaches IOC. Still, not a bad bargain if it works. But now, since you've gone cheap and off the shelf, you've got a slow, unmaneuverable aircraft that can't detect a target (no radar), can't communicate well with it's controllers, and can't climb very high, limiting the range of its missiles. Then, even if you outfit this Predator swarm with an UberMissile, you'll have blown your cost equation out of the water when this swarm comes up against SAMs, or a squadron of F-16s. The swarm won't even see the F-35s due to frontal aspect stealth, and they'll be able to target the Predators and kill them like baby seals. They'll be able to engage, kill, and disengage before the low performance Predators will be able to react.
    - RogerL 12y ago
      I want to write to back you up on this. I've spent most of (not all of) my career in military avionics and aerospace. And this was at small companies. The big companies win the big contracts, but so much is contracted out, and there are plenty of other options. But I digress... This is a really hard, multi-dimensional optimization problem. The engineers at Lockheed Martin, Rockwell, etc., are not dumb. They are the opposite of dumb. They are rocket scientists, in both the colloquial and literal sense. The same holds true for us at the smaller companies. Heavy STEM educations, a lifetime of engineering and research. I've done a lot of work on next-gen battlefield stuff. It's not easy. It's not. It's control theory, filter theory, AI, computer vision, RF design, and so much more. If you can't play in at least 3-4 of those fields at a pretty high level you aren't invited to the party. I've seen far more professional, intelligent, and pragmatic engineers in that work than what I've seen pass for engineering in Silicon Valley. You really aren't going to sit in an armchair and design up a system that'll work. OTOH, if you make something that does work, the military will beat a path to your door. Look into SBIRs, for example. I've worked on a lot of them. The door is not closed. The door is closed to half-baked schemes, but if you have an idea for putting some low cost hardware and software together, you can do something with it. I'm being more than a bit facile here; it isn't so hard to get through a Phase I and Phase II SBIR; getting your system deployed is quite a bit harder. Still, there are options. Just be prepared for the fact that they are going to take your product, stick it in a field, and shoot (literal) bullets at it. Ohh, failed the projectile test.... Yes, I take the point about swarms, but war is an extremely hostile environment (no pun intended). And as soon as you have tipped the equation towards "disposable", expect the enemy to change their systems to help you 'dispose' of them as soon as possible. Maybe your drone can operate in the theater today; what about 10 years from now? I honestly don't know, the calculus is anything but clear to me.
branchan 12y ago
I think there's a general disconnect between HN users and how the defense industry works. When you are designing a weapons system designed to kill people, you cannot just design something 'in a hackish fashion'. No, this shit has got to work correctly all the time. The cheapest Predator drone cost about $5M and top flight Predators and Global Hawks cost between $20M-50M. This means flying 100 of them is not reasonable. UAVs are also not autonomous and need manned pilots. Even if you were to build 100 drones, are you going to pay for 100 extra pilots, create some kind of airport that can house 100 drones, keeping them all fueled and maintained in the desert? Kinda naive to think that you can just use SV start up philosophies in this application.
- owenmarshall 12y ago
  > Kinda naive to think that you can just use SV start up philosophies in this application. Pick your response: * Welcome to Hacker News, please enjoy your stay! * If I had a nickel for every time someone tried to shoehorn start up philosophy into an unrelated field, I'd be posting from my yacht!
- msandford 12y ago
  > I think there's a general disconnect between HN users and how the defense industry works. Agreed, and I stated EXACTLY that in another portion of the thread: https://news.ycombinator.com/item?id=7907637 https://news.ycombinator.com/item?id=7907637 > When you are designing a weapons system designed to kill people, you cannot just design something 'in a hackish fashion'. No, this shit has got to work correctly all the time. No, with that part of engineering as with EVERY part of engineering: it depends. Do you want UNIT reliability or SYSTEM reliability? Does every plane have a literal 100% chance of intercepting every target it is tasked with? No, of course not. But the missiles do, you say? Again wrong. But surely SOME portion of the entire system will work ALL THE TIME! Nope, no engineer worth his salt will give you a 100% guarantee for anything as unknown as "stopping all possible threats". Furthermore plenty of systems that are supposed to be very reliable are made up of less reliable parts. The whole Star Wars program was based on defense in depth where no one layer of system was going to stop 100% of the warheads but multiple layers acting in concert would be able to (presumably) stop them all or almost all. Hackish is fine provided that you've got substantially overlapping coverage from multiple command and control drones. No one drone has to work 100% guaranteed because even at only a 95% success rate (which is abysmal compared to the "all the time" demand you're making) with three overlapping zones gives you .05 * .05 * .05 = .0125% chance of failing. I'll take 1/8 of a chance of failure per thousand incidents as successful enough. Again, I'm not talking about having 100 drones with 100 pilots but perhaps having 100 drones with 9 pilots for the command and control drones (one for every 10 mules) and some kind of a very rudimentary, randomized loiter algorithm for the mules. > Kinda naive to think that you can just use SV start up philosophies in this application. I want to say a bunch of really snarky stuff in response. I'm an outsider to SV; I grew up in MN, went to school in FL and now I live in TX. None of these places ever get accused of being even SLIGHTLY SV-ish in nature so I don't see how I deserve that kind of comment. Ultimately though you've made a lot of assumptions that don't necessarily hold. The idea that you're going to defeat the US using the same kind of procurement and whatever that the US uses is a non-starter. You don't try and beat the US at a symmetric war but you can defeat them with an asymmetric one. We're losing how many lives and how much money in Iraq and Afghanistan right now to IEDs which are what, 10 notches below the fancy shit we have? And yet all our fancy airplanes haven't saved a single soldier from an IED that I'm aware of. I'm not necessarily saying that you absolutely 100% can use SV philosophies in war and win, but I am saying that the notion that the way the US military does it is the ONLY way to do it isn't correct either.
msandford 12y ago
Most of the arguments fall into one of two categories: 1. Performance is linear(ish) with cost 2. The US operates at a global maximum in price/performance I disagree with both. Performance is roughly linear with cost given a certain operational parameter envelope. Things change when you change the envelope though. Dramatically. And the idea that the US has chosen the ultimate operational parameter window which cannot be improved on under any circumstances isn't believable. I do appreciate everyone arguing with me long enough that I could distill it so succinctly. You may still disagree with me and that's OK. But at least my crazy heretical thoughts are more easily digested now.

greedo 12y ago

Again, it's a cost issue. If you think you can build a super Predator with some type of targeting radar, and a swarm type of AI for under $10M, you're not being realistic. Current build Predators are already over $4M a copy, and even with this fantasy equipment you've described, wouldn't stand a chance against a decent human in an F-15. It's important to understand the limits of radar too; it's not like in a video game where you get this god view of the airspace around you. You are always making tradeoffs in flight about where your radar is focused, it's range, and how many targets it can track while scanning. That's why the US and it's allies rely so heavily on an AWACs system. UCAS aircraft that can drop bombs in a contested airspace are going to cost north of $50M per copy when all is said and done, and UCAS that can do A2A combat will be double that. Once you start putting high performance jet engines in drones, the costs start to rise. Same with avionics. Today's drones are very very simple aircraft compared to high performance fighters.

msandford 12y ago
Agreed, cost is what would drive using drones instead of high performance fighter aircraft. I am not suggesting that you build unmanned F-15s for the $4mm each that you can build a Predator for. That's patently ludicrous I agree with you! So no fancy engines, nor fancy avionics. Also don't bother with swarm AI, have people on the ground handling targeting. Don't develop a billion dollar solution when $50k/year (roughly) works just fine. What I was suggesting is that you build a bunch of aerial SAM launchers. Take a Predator, put something that can receive targeting info on it, arm it with air to air missiles. Deploy many of these in the sky. Then for every 10 missile mules, you put up a single command and control Predator. It has the fancy equipment to ensure that you can actually do targeting. Then put a transmitter on all Predators which can be turned on such that you can't detect which Predator is the command and control one. Let's suppose that you could create this fleet of 11 for $4mm * 11 + $2mm * 10 + $20mm * 1 = $84mm This is very conservative in my mind because at this point we're basically paying US prices for the gear. Again, we're not going for the fanciest, best, most high performance stuff ever built. Use old designs, don't worry about the best survivability, don't make them too rugged and thus don't drive up the cost. We're going for The Innovator's Dilemma style worse-is-better so long as it's functional enough to mule up the ordinance. You can now afford at least 2x of these 11 unit Predator swarms for every 1x F-35 that can be deployed. An F-35 only carries 10 air to air weapons and that's when loaded with nothing but. The odds they get both command and control Predators on every engagement is statistically slim provided you're talking about at least a few dozen engagements. That means that just playing a numbers game you're going to shoot down the F-35s and because you've got so many damn drones in the air enjoy quite a bit of air superiority. It's not exceedingly difficult to imagine that if the sky was thick with these things that the F-35s might not get all their ordinance away before being shot at making the grinding more painful. Could the US manage such a feat of engineering discipline and fast-track development? NOT A CHANCE!! If you're thinking that this is unrealistic based on how things work in the US you're 100% correct. But to suppose that no other industrialized country in the world could embark on such a venture and succeed? If you believe that please stop reading HN because startups could never work/win against the entrenched giants who have all the advantages.
- greedo 12y ago
  It's not a matter of startups, or entrenched giants or any such thing, or the Innovator's Dilemma. It's a matter of communications, cost and performance. A Predator can't fly as high as an F-35 or F-22. It can't maneuver nearly as well either. Can't fly as fast. It can't be refueled in flight, so it's range is lower than the F-35. So you'd have to hope that the missile it carried could compensate for these deficiencies. However, missiles are very finicky, and have launch parameters that help them improve their odds. The military looks at it in terms of PoK (probability of kill). So to be cheap (so you can afford to have swarms) you have to give up performance. This will affect PoK. Next you have to have robust, un-jammable commlinks. So far, no one has those. That's going to be expensive to develop. Humans can make decisions when the comms are degraded, or non-existent. AI has a long ways to go, and AI is expensive. To remain cheap, you'll have to give that up. With you napkin/elevator calculations, you would be trading 22 Predators for each F-35. 22 Predators is $88M, and the F-35 will eventually drop close to $100m as it reaches IOC. Still, not a bad bargain if it works. But now, since you've gone cheap and off the shelf, you've got a slow, unmaneuverable aircraft that can't detect a target (no radar), can't communicate well with it's controllers, and can't climb very high, limiting the range of its missiles. Then, even if you outfit this Predator swarm with an UberMissile, you'll have blown your cost equation out of the water when this swarm comes up against SAMs, or a squadron of F-16s. The swarm won't even see the F-35s due to frontal aspect stealth, and they'll be able to target the Predators and kill them like baby seals. They'll be able to engage, kill, and disengage before the low performance Predators will be able to react.
  - RogerL 12y ago
    I want to write to back you up on this. I've spent most of (not all of) my career in military avionics and aerospace. And this was at small companies. The big companies win the big contracts, but so much is contracted out, and there are plenty of other options. But I digress... This is a really hard, multi-dimensional optimization problem. The engineers at Lockheed Martin, Rockwell, etc., are not dumb. They are the opposite of dumb. They are rocket scientists, in both the colloquial and literal sense. The same holds true for us at the smaller companies. Heavy STEM educations, a lifetime of engineering and research. I've done a lot of work on next-gen battlefield stuff. It's not easy. It's not. It's control theory, filter theory, AI, computer vision, RF design, and so much more. If you can't play in at least 3-4 of those fields at a pretty high level you aren't invited to the party. I've seen far more professional, intelligent, and pragmatic engineers in that work than what I've seen pass for engineering in Silicon Valley. You really aren't going to sit in an armchair and design up a system that'll work. OTOH, if you make something that does work, the military will beat a path to your door. Look into SBIRs, for example. I've worked on a lot of them. The door is not closed. The door is closed to half-baked schemes, but if you have an idea for putting some low cost hardware and software together, you can do something with it. I'm being more than a bit facile here; it isn't so hard to get through a Phase I and Phase II SBIR; getting your system deployed is quite a bit harder. Still, there are options. Just be prepared for the fact that they are going to take your product, stick it in a field, and shoot (literal) bullets at it. Ohh, failed the projectile test.... Yes, I take the point about swarms, but war is an extremely hostile environment (no pun intended). And as soon as you have tipped the equation towards "disposable", expect the enemy to change their systems to help you 'dispose' of them as soon as possible. Maybe your drone can operate in the theater today; what about 10 years from now? I honestly don't know, the calculus is anything but clear to me.

branchan 12y ago

I think there's a general disconnect between HN users and how the defense industry works. When you are designing a weapons system designed to kill people, you cannot just design something 'in a hackish fashion'. No, this shit has got to work correctly all the time. The cheapest Predator drone cost about $5M and top flight Predators and Global Hawks cost between $20M-50M. This means flying 100 of them is not reasonable. UAVs are also not autonomous and need manned pilots. Even if you were to build 100 drones, are you going to pay for 100 extra pilots, create some kind of airport that can house 100 drones, keeping them all fueled and maintained in the desert? Kinda naive to think that you can just use SV start up philosophies in this application.

owenmarshall 12y ago
> Kinda naive to think that you can just use SV start up philosophies in this application. Pick your response: * Welcome to Hacker News, please enjoy your stay! * If I had a nickel for every time someone tried to shoehorn start up philosophy into an unrelated field, I'd be posting from my yacht!
msandford 12y ago
> I think there's a general disconnect between HN users and how the defense industry works. Agreed, and I stated EXACTLY that in another portion of the thread: https://news.ycombinator.com/item?id=7907637 https://news.ycombinator.com/item?id=7907637 > When you are designing a weapons system designed to kill people, you cannot just design something 'in a hackish fashion'. No, this shit has got to work correctly all the time. No, with that part of engineering as with EVERY part of engineering: it depends. Do you want UNIT reliability or SYSTEM reliability? Does every plane have a literal 100% chance of intercepting every target it is tasked with? No, of course not. But the missiles do, you say? Again wrong. But surely SOME portion of the entire system will work ALL THE TIME! Nope, no engineer worth his salt will give you a 100% guarantee for anything as unknown as "stopping all possible threats". Furthermore plenty of systems that are supposed to be very reliable are made up of less reliable parts. The whole Star Wars program was based on defense in depth where no one layer of system was going to stop 100% of the warheads but multiple layers acting in concert would be able to (presumably) stop them all or almost all. Hackish is fine provided that you've got substantially overlapping coverage from multiple command and control drones. No one drone has to work 100% guaranteed because even at only a 95% success rate (which is abysmal compared to the "all the time" demand you're making) with three overlapping zones gives you .05 * .05 * .05 = .0125% chance of failing. I'll take 1/8 of a chance of failure per thousand incidents as successful enough. Again, I'm not talking about having 100 drones with 100 pilots but perhaps having 100 drones with 9 pilots for the command and control drones (one for every 10 mules) and some kind of a very rudimentary, randomized loiter algorithm for the mules. > Kinda naive to think that you can just use SV start up philosophies in this application. I want to say a bunch of really snarky stuff in response. I'm an outsider to SV; I grew up in MN, went to school in FL and now I live in TX. None of these places ever get accused of being even SLIGHTLY SV-ish in nature so I don't see how I deserve that kind of comment. Ultimately though you've made a lot of assumptions that don't necessarily hold. The idea that you're going to defeat the US using the same kind of procurement and whatever that the US uses is a non-starter. You don't try and beat the US at a symmetric war but you can defeat them with an asymmetric one. We're losing how many lives and how much money in Iraq and Afghanistan right now to IEDs which are what, 10 notches below the fancy shit we have? And yet all our fancy airplanes haven't saved a single soldier from an IED that I'm aware of. I'm not necessarily saying that you absolutely 100% can use SV philosophies in war and win, but I am saying that the notion that the way the US military does it is the ONLY way to do it isn't correct either.

msandford 12y ago

Most of the arguments fall into one of two categories: 1. Performance is linear(ish) with cost 2. The US operates at a global maximum in price/performance I disagree with both. Performance is roughly linear with cost given a certain operational parameter envelope. Things change when you change the envelope though. Dramatically. And the idea that the US has chosen the ultimate operational parameter window which cannot be improved on under any circumstances isn't believable. I do appreciate everyone arguing with me long enough that I could distill it so succinctly. You may still disagree with me and that's OK. But at least my crazy heretical thoughts are more easily digested now.