Utopia Talk / Politics / US Air Force has lost it's mind

murder Member	Wed Mar 08 12:08:24 http://www.youtube.com/shorts/kU_lp5Gcf44 200 next generation fighters and 1000 "loyal wingmen" combat drones. This is going to be a disaster that will make the F-35 seem like a good idea by comparison.
Rugian Member	Wed Mar 08 12:21:54 *cue ecstatic Raytheon and Lockheed Martin noises*
murder Member	Wed Mar 08 12:28:43 Why? They were going to make whatever the hell we bought. This is just the insane route the air force has chosen to take. You'd think if they wanted to go this route they would first test this out on a smaller scale before committing to it. But I guess that's not as fun as barreling head first into a brick wall and seeing how it turns out.
Nimatzo iChihuaha	Wed Mar 08 12:44:13 Are air force tests of next gen weapons public and transparent?
Rugian Member	Wed Mar 08 12:48:24 Murder Do you know how much of a financial clusterfuck next-gen fighter orders are? The research, development, and procurement will inevitably go many billion dollars over budget. That's more billings for the contractors.
murder Member	Wed Mar 08 13:10:28 "Are air force tests of next gen weapons public and transparent?" No obviously. But there's just nowhere to hide a bunch of jets flying around, especially is you're asking them to blow stuff up. Even more so if you are testing them vs "opposition".
murder Member	Wed Mar 08 13:15:30 "Do you know how much of a financial clusterfuck next-gen fighter orders are?" It was always going to be next generation. They were going to move past the F-22. It was just a matter of what got built, and we just keep going down goofier and goofier rabbit holes. If you're going to have these purportedly cheaper "loyal wingman" flying just as far and as fast as the next generation fighter, and carrying all the weapons and sensors, then I don't understand the point of the expensive fighter. We can remote control drones halfway around the world from a La-Z-Boy in Arizona and save the money we're going to spend on the jet.
Seb Member	Wed Mar 08 15:33:49 Murder: Really? Look at what you can do with ML right now - I'm pretty sure the tech will handle combat manoeuvres. Building a combat drone with no soft squidgy bits that can't handle high g forces that operate as a picket in front of maned fighters exercising mission control is going to be a game winner. "test this out on a smaller scale" What makes you think they haven't? If the UK and France have been running ucav test beds for over a decade with fully automated combat capabilities I'm pretty sure the US is further ahead still - and we are probably all using technology for the AI part that's well behind what state of the art is capable of.
Seb Member	Wed Mar 08 15:39:40 "then I don't understand the point of the expensive fighter" Human judgement directing and shaping that mission and adjusting to circumstances would be one thought, with the loyal wingman do the actual fighting. But it may also be the case that it's like with tanks when they first came in and they gave them to cavalry units on account of them being things that moved fast on land, but being entirely built around maintaining horses and leading charges rather than supporting infantry the cavalry units did not know what the fuck to do with them. I would suspect keeping a mashed fighter force around is a good hedge until you have really worked out a new operational doctrine with experience.
murder Member	Wed Mar 08 16:08:55 "Look at what you can do with ML right now - I'm pretty sure the tech will handle combat manoeuvres." You can be pretty sure, but all evidence to date is that we can't build them. All drones tend to be slow and/or have to stick to uneventful flying. "What makes you think they haven't?" We haven't seen it. We've seen the B-21 and it hasn't even flown yet.
Nimatzo iChihuaha	Wed Mar 08 17:04:54 Covert ops man, top secret dark site testing. They will not show that stuff. They are still reverse enginnering technology from the Roswell crash.
murder Member	Wed Mar 08 17:12:28 The worst part of this as far as I'm concerned is that we've just been shown what the state of the art looks like (the B-21) ... and it looks a lot like 1980s tech. And that coming on the heels of the F-35 program which despite all the hype is a not particularly impressive fighter wrapped in supposedly improved stealth coating. There's absolutely no reason to believe that some revolution in technology is just over the next hill that is going to redefine combat. It will be a even stealthier than the F-22 and have better engines. And aside from that it will just have the best of whatever the heck everyone else has. No mach 3. No laser weapons. No magic pixie dust.
murder Member	Wed Mar 08 17:13:34 "They are still reverse enginnering technology from the Roswell crash." I know you're joking but it's amazing how many people still believe that nonsense so many decades later.
obaminated Member	Wed Mar 08 21:44:00 I like the idea of a human maned fighter jet being protected by 4 robot controlled drones. It'd make for a fun action adventure movie. Top gun 3 anyone?
Seb Member	Thu Mar 09 01:27:25 Murder: "All drones tend to be slow and/or have to stick to uneventful flying" And how do you know that, exactly? Pretty sure that's regulatory requirement. China just did a round of tests with an AI autonomous done and it best the shit out of manned aircraft (partly because it could do manoeuvres that would kill pilots). I remember new scientist reporting on ai doing this stuff in full simulator when I was a kid back in the 90s when neural nets were new and computers shit. I would not be at all surprised if the capabilities are way more than you are assuming. "We haven't seen it" What do you mean? You've seen UK corax, taranis, US x-47 Pegasus and the ghost Ray?
Seb Member	Thu Mar 09 01:31:47 The revolution is autonomous combat. Planes can already physically pull manoeuvres that would kill their pilots. You don't need Mach 3 and other such performance enhancements - you need to unlock more of the parameternspace you already have access too but can't use because it causes the pilots brain to pop.
Seb Member	Thu Mar 09 02:54:46 Argh ghost bat, not ghost Ray
Seb Member	Thu Mar 09 02:59:25 And I think Lockheed has something along the lines of "just automate modernised f-16s". Anyway, I suspect what you get from drones is: Greater stealth, higher G manoeuvre, lower cost and expendability, simply by taking out the pilot. That's lots of compounded marginal improvement. In terms of tactical advantage this may well translate into way more than some new bleeding edge revolutionary technology targeting some aspect of performance.
jergul large member	Thu Mar 09 04:30:54 So basically, the USAF want recoverable subsonic missiles carrying submunitions. So revolutionary.
Seb Member	Thu Mar 09 05:54:20 Jergul: In the sense that all manned fighters have always been recoverable missiles carrying submunitions; or all missiles have always been unmanned aircraft. I.e. not sure that's a meaningful use of terminology.
jergul large member	Thu Mar 09 06:14:35 Seb Nah. Manned fighters have always had autonomy and only recently have been confined to the role of missile and glide bomb buses almost exclusively. I am pretty sure it is a meaningful use of terminology. It cuts through the hype. A no-brainer to replace manned aircraft with unmanned ones. But why would the USAF even want manned next generation fighters? The concept would do better with an AWAC type control aircraft to manage those recoverable, loitering missiles with submunition payloads.
Sam Adams Member	Thu Mar 09 09:53:22 "There's absolutely no reason to believe that some revolution in technology is just over the next hill that is going to redefine combat." Cheap swarming drones is certainly going to redefine combat and maybe beam weapons.
Sam Adams Member	Thu Mar 09 09:55:05 In an environment woth both... how do you counter ai drones? With other ai drones? With a battlewagon bristling with point defense beam weapons?
murder Member	Thu Mar 09 10:46:28 "I like the idea of a human maned fighter jet being protected by 4 robot controlled drones. It'd make for a fun action adventure movie. Top gun 3 anyone?" The manned fighter flying unmolested as 4 drones clear a path for it. Very exciting.
murder Member	Thu Mar 09 11:02:08 "And how do you know that, exactly?" How do I know that we don't have capabilities that we don't field? How many military budgets do you think we have? Really the only drones that can anything like what is proposed are racing drones, and I don't think those will scale very well. "China just did a round of tests with an AI autonomous done and it best the shit out of manned aircraft (partly because it could do manoeuvres that would kill pilots)." Sure it did. This is the same China that can't produce its own high end chips, can't build "stealth" aircraft without extra control surfaces sticking out everywhere at multiple angles, is still copying or purchasing multiple Russian weapon systems, and is just mastering passenger airliners? Don't become jergul. We've already got one of those. "You've seen UK corax, taranis, US x-47 Pegasus and the ghost Ray?" I haven't seen Corax, but the rest support my point. They all fly like single prop Cessnas. Those things wouldn't even function as clay pigeons for manned fighters let alone aggressor training.
Rugian Member	Thu Mar 09 11:02:15 The entire point of Top Gun: Maverick was that a skilled human pilot is the most important element in a combat scenario, next gen technology be damned. Top Gun 3: Drone Wars would be quite the shift from that lol
murder Member	Thu Mar 09 11:08:59 "The revolution is autonomous combat." Maybe, but not soon. "Planes can already physically pull manoeuvres that would kill their pilots." No they can't. Point to any evidence of this. Any at all. "Greater stealth, higher G manoeuvre, lower cost and expendability, simply by taking out the pilot. That's lots of compounded marginal improvement." High performance drone ... whenever they become available ... will not lower cost right away because the technology will be so cutting edge. In fact cost would initially go up. It would be the premium you pay to keep pilots out of harms way. It would take a while before cost came down.
murder Member	Thu Mar 09 11:11:03 "A no-brainer to replace manned aircraft with unmanned ones. But why would the USAF even want manned next generation fighters? The concept would do better with an AWAC type control aircraft to manage those recoverable, loitering missiles with submunition payloads." Crazy as it sounds, that may be exactly what happens. They keep saying that the next generation air dominance (NGAD) may not resemble a traditional fighter. Although they also get lost in the currently popular system of systems bullshit.
murder Member	Thu Mar 09 11:14:06 "The entire point of Top Gun: Maverick was that a skilled human pilot is the most important element in a combat scenario, next gen technology be damned." Top Gun 3 is going to be cyborg Maverick because he'll be so old that just launching off the deck of a carrier would otherwise kill him.
murder Member	Thu Mar 09 11:21:42 "Cheap swarming drones is certainly going to redefine combat and maybe beam weapons." Come on Sam ... you know that whatever directed energy weapon you can put on an aerial platform of any size, you can put a more powerful one on the ground. The day that directed energy weapons become a useful thing, air defenses win and bombers and fighter of any sort will be obsolete.
Seb Member	Thu Mar 09 11:34:28 Sam: Pervasive small suicide quadcopters are going to be revolutionary. If they can put a thermite payload on a barrel then they can mission kill most things. I suspect the first line of defence will be jamming and the question is how smart you can make on board processor. Hard counters are going to have to be, as you say lasers, but even then I think it favours the drone. They can score
Seb Member	Thu Mar 09 11:38:50 Murder: He's thinking quadcopters. Attack from multiple directions by small, cheap drones going for the optics etc. Saturate defences and then either get a mission kill (wheels, treads, sensors, armaments disabled) or once sufficiently degraded, a bigger missile.
Seb Member	Thu Mar 09 11:39:53 Murder: "No they can't. Point to any evidence of this. Any at all" Pilots pulling manoeuvres that cause blackouts and redouts is a thing. Fast jet pilots already wear suits that try and compensate for this.
Seb Member	Thu Mar 09 11:44:33 "will not lower cost right away because the technology will be so cutting edge" You are using the wrong heuristic. The first autonomous drones will not be radically different in terms of cost than an equivalent manned airframe - they don't use fancy new materials or exotic geometries or clever engine designs. They just take contemporary or previous generation technology and strip out all the stuff needed only for pilots: life support, ejector seat, cockpit etc. and add a bit more compute. Compute is cheap.
jergul large member	Thu Mar 09 12:03:21 Seb That was a very naive view on how the military industrial complex works.
murder Member	Thu Mar 09 12:41:13 Here you go ... http://www.youtube.com/watch?v=dbsyfITResU
murder Member	Thu Mar 09 12:49:25 "Pilots pulling manoeuvres that cause blackouts and redouts is a thing. Fast jet pilots already wear suits that try and compensate for this." Evidence of drones being able to outmaneuver manned aircraft. "Compute is cheap." Compute costs the same for everyone. If you want an advantage then you don't want compute to be cheap. The cutting edge stuff is not cheap. Particularly stuff that can replace a human pilot. Especially one that your opponent can't reproduce. The point of producing a new generation of hardware isn't to produce better weapons than the last generation, it's to gain or maintain an advantage over the enemy.
murder Member	Thu Mar 09 12:51:11 The counter to drone swarms is the same as the answer to piloted aircraft swarms in WWII. Volume of fire. Small drones may be cheap, but so are air bursting shells.
murder Member	Thu Mar 09 12:55:36 "Seb: That was a very naive view on how the military industrial complex works." It has nothing to do with the military industrial complex. It has to do with capability. He likes to say that compute is cheap ... but supercomputers are not. The cost is going to be driven by what you want that computing power to achieve and the ability to cram it into whatever package you need to carry it and the ability to operate it with whatever spare power you can squeeze from the engines.
jergul large member	Thu Mar 09 13:13:07 Murder It has everything to do with the military industrial complex. They are not going to tarp over an f-16 cockpit like som vintage German glide bomb, tack on an antenna to pick up LOS data packets, and call it a day.
murder Member	Thu Mar 09 13:35:14 Of course not. The air force wouldn't for that. They may tarp over the cockpit of an F-35 though. Gotta have stealth.
patom Member	Thu Mar 09 14:27:52 If they want speed, there are probably a thousand or so Phantoms sitting in the Arizona Dessert. They put them out of service because human pilots couldn't handle the maneuvering that these planes might have been capable of.
Seb Member	Fri Mar 10 01:16:45 Jergul: You have confused cost and price. Murder: Give me an example of a city being destroyed by a nuclear bomb delivered by an ICBM. Oh you can't? Clearly they don't work. "Compute costs the same for everyone. If you want an advantage then you don't want compute to be cheap." Nevertheless compute *IS* cheap. So you can stay with manned fighters likely to be outperformed by cheaper to produce, cheaper to operate drones that can pull more extreme manoeuvres and hope nobody else notices compute is cheap or ... "The cutting edge stuff is not cheap." What's cutting edge? Are you moving stuff and compute up? "Especially one that your opponent can't reproduce." Computes a utility. Embrace it. Creating bespoke compute that your enemy doesn't have access to off only useful if it gives a capability you can't get from utility grade. "The point of producing a new generation of hardware isn't to produce better weapons than the last generation, it's to gain or maintain an advantage over the enemy." Ok, sure, then keep looking around for a wunderweapon while your enemies catch up with your old tech, and exploit the comoditisation of all of this. Quantity has a quality all of its own. As does expendability. "but so are air bursting shells." Logistics is expensive though, and ww2 required full mobilisation. And quadcopters are much much smaller and fly much much lower. Sir bursting shells in the middle of your armour formation as a bunch of drones burst out of the tree line seems self defeating.
Seb Member	Fri Mar 10 01:30:35 Jergul: "They are not going to tarp over an f-16 cockpit like som vintage German glide bomb, tack on an antenna to pick up LOS data packets, and call it a day." The phrase is "optionally manned" and yes, they have explored literally that. http://nat...ongside-manned-fighters-196522 But my point is more that previous generations of fighter development normally incorporate some cutting edge material and design aspects with huge overheads and retirement for novel manufacturing techniques. The thing you are leveraging here is comoditisation.
jergul large member	Fri Mar 10 02:13:48 Seb You continually fall into the trap of overoptimizing. The military industrial complex (MIC) is not Hearts of Iron IV. Also. Jargon (technobabble). Geeze. Who cares what the MIC calls stuff. Tarped F-16s is a no-go. The logistic footprint. What the MIC wants is 35 billion to develop something new, then serial production. What makes sense would be AWAC control of Predators or something. Get that lethargy done by controlling the UAVs from 100km instead of from New Mexico via satellite. My caveat here is the war in Ukraine. Turns out repurposing really old stuff from storage is very viable. But wait. I am suggesting doing that as the Predator is retired. Ok, the Reaper then. Loitering capability to match that of the controlling AWAC. The idea that an overtaxed fighter pilot should also be controlling reusable loitering missiles with submunition packages is ludicrous. How about an unmanned controlling aircraft managing cheaper, more expendible reusable missiles? I am sure satellite could allow for human decision inputs somehow. My point is really that nothing here is new.
Seb Member	Fri Mar 10 02:46:26 Jergul: "The ultimate goal of the Loyal Wingman initiative is to pair manned, fifth-generation stealth fighters with unmanned versions of older jets — in order to boost the lethality of both in air combat." This isn't optimization. I mean I literally said he reverse: "The thing you are leveraging here is comoditisation." The model of ever increasingly complex jets exploiting bleeding edge technologies for advantage at ever higher unit prices - that's the optimisation trap. You are being needlessly contrarian again.
Seb Member	Fri Mar 10 02:55:14 "The idea that an overtaxed fighter pilot should also be controlling reusable loitering missiles with submunition packages is ludicrous." Directing, not controlling. The problem is you are mixing up autonomous and unmanned. If we were talking unmanned, then what you say makes sense to a degree*. But the point of loyal wingman programs is the wingmen are autonomous. They see off enemy fighter and missile threats allowing the pilot to focus on the strike element of the mission. They also carry extra munitions, and increase survivability by putting active sensors away from the human. It is a given that unmanned systems are preferable to manned - so the point here is for missions that need manned flights, what can we do to reduce the risk to pilots and enhance their capability given that an unmanned system is not viable due to e.g. need for localised rapid decision making with good situational awareness or a hostile electronic warfare environment. Man plus autonomous systems under direction of man. * To a degree in the sense that controlling from AWACS isn't that robust. Anyone that can jam datalinks to a base way back in the US or Europe or wherever will also jam it to an AWACS 100 km away.
jergul large member	Fri Mar 10 05:26:11 Awacs direct. Overtasking a pilot with directing drones if that is the word you prefer. Then anyone can jam datalinks to a fighter 500m-50km away. Upper range to cover using the "wingman" as an expendable penetration weapon. There is honestly no point to the concept beyond feeding the maw of the MIC. But everybody loves hype. So hype away.
Seb Member	Fri Mar 10 07:12:17 jergul: "Awacs direct." AWACS are not going to be in a great position to do that from hundreds of KM away. Their situational awareness is shit compared to the pilots. After all, pilots already communicate with each other to accomplish missions. "Overtasking a pilot with directing drones if that is the word you prefer." How would it be different to target an enemy fighter with the drone, than to target it with a missile? How would it be different to a pilot directing a human wingman in terms of task load? "Then anyone can jam datalinks to a fighter 500m-50km away." I suspect the kind of thing you would be doing here would be equivalent to painting a target with laser or radar as a signal to an autonomous drone to "go get that bogey while I head in that direction to focus on my actual target, this radar site over here.", perhaps very automated point defence like functions in the same way that CIWS for ships in combat situations are set to engage threats automatically. I don't think it would be so much of the case of the F35 pilot needing to have high bandwidth communication to perform functions that a pilot would perform for that craft if it were piloted. The point of making it autonomous is that you don't need someone doing a lot of thinking - that's been left to the software. Basically, the F35 is there to drop bombs on things. The loyal wingman is there to engage any fighter or missiles trying to engage the F35, provide extra radar source, and a magazine of additional munitions - and do so largely autonomously to leave the F35 pilot to focus on ensuring he drops the bomb on the thing he is trying to drop a bomb on.
Seb Member	Fri Mar 10 07:17:20 "There is honestly no point to the concept beyond feeding the maw of the MIC." Hmm. I can see that the argument is that there is no point to manned fighters. I cannot see the argument here at all that there is no point to reducing the number of highly trained humans that need to be put in harms way and that inherently limit the manoeuvrability, payload and range of aircraft. Especially as not having a pilot means you can start to design aircraft for higher expendability which could give compounding returns as you strip out expensive capabilities that are really just there to ensure survivability and instead double down on cheap and plentiful and accepting losses. Sure, in the limit that means you turn everything into missiles - and maybe that is the way it will reduce down to.
murder Member	Fri Mar 10 08:42:42 "Give me an example of a city being destroyed by a nuclear bomb delivered by an ICBM. Oh you can't? Clearly they don't work." Come on Seb, that's a nonsense argument. We've all seen ICBMs. We all seen evidence of them being tested in public view. The same goes for the warheads. "... cheaper to operate drones that can pull more extreme manoeuvres and hope nobody else notices compute is cheap or ..." You keep insisting on this in the absence of evidence. There is no evidence that the amount of computing power we can load onto drones is capable of the complexity of aerial combat. You literally just witnessed the unveiling of the "6th generation" B-21 bomber ... and it's a manned bomber.
murder Member	Fri Mar 10 08:45:56 "Especially as not having a pilot means you can start to design aircraft for higher expendability ..." Good luck with that. Especially come wartime when weapons platforms start getting blown up ... and it's going to take the factory 3 years to ramp up production to fill the need for the shit you just got blown to hell. You are watching Ukraine, right?
Seb Member	Fri Mar 10 09:15:34 murder: "We've all seen ICBMs. We all seen evidence of them being tested in public view. The same goes for the warheads." Yeah, and we all know that there are planes that can pull higher G manouvres than pilots can safely handle - G-LOC is a thing. We know that we've even developed suits etc. to try and mitigate the fact. Why would you believe that such drones couldn't exist and be part of a design programme? "There is no evidence that the amount of computing power we can load onto drones is capable of the complexity of aerial combat." Why would you think that it wasn't? It takes a supercomputer to train some of these algos, but running them requires far less. You can fit a petaflop of GPU capacity into a single 19-inch rack cabinet these days. "and it's a manned bomber." Indeed. What is your point? The entire loyal wingman programme (and us doctrine) is based on the idea you want a person to be in the decision loop on what gets blown up. You would not want to send an unmanned B21 around the world - the risk is it might get spoofed into landing in the wrong spot, or targeting the wrong thing etc. in ways that a human would not. "and it's going to take the factory 3 years to ramp up production to fill the need for the shit you just got blown to hell." Part of the logic of drones - make them simple and from commodity bits (because you are not chasing advantage through ever more weird specialist materials and architectures) and it is much easier to scale up production or have - like the predator - a steady ongoing production. Not batches of a few hundred exquisitely over engineered air frames. "You are watching Ukraine, right?" Yeah, I am, and I wonder if you are! What person watching Ukraine things "I really ought to double down on a dwindling number of extremely expensive manned platforms with increasingly expensive marginal performance advantages". You want cheap shit you can churn out massively that does not require highly trained staff and preferably has a minimal logistics footprint. And if you can whack a sever rack of GPUS where the cockpit used to be in a few F16s and have them - in extremis - take a missile for an F35 or F22 - isn't that a bloody good deal? It certainly solves the problem in the RAND wargame for the Taiwan strait where the chinese airforce could literally take all the losses the entire F22 fleet could deliver, and still have enough airframes survive to destroy tankers and awacs leaving the F22s to drop into the sea even if they Chinese couldn't hit them. Quantity has a quality all of its own.
jergul large member	Fri Mar 10 09:37:17 Seb No point in manned fighters. I totally agree. My main issue was with that link in the chain.
murder Member	Fri Mar 10 09:38:18 "Yeah, and we all know that there are planes that can pull higher G manouvres than pilots can safely handle - G-LOC is a thing." You keep saying this but there is no evidence that computers can handle those maneuvers. It's that's never happened before thing. "Why would you believe that such drones couldn't exist and be part of a design programme?" Why do you not believe in God? "Why would you think that it wasn't?" Because there's no evidence that it can. And what's more is that there is no advantage to keeping that technology hidden from public view if you can do it. I mean we don't sell the F-22 to other countries, and the tech on the platform is secret, but we fly them at air shows and they do all their fancy maneuvering in plain view. "You would not want to send an unmanned B21 around the world - the risk is it might get spoofed into landing in the wrong spot, or targeting the wrong thing etc. in ways that a human would not." 90% of a bombers targets are going to be fixed. Coordinates would get the job done fine. And how the fuck would AI get spoofed. "Part of the logic of drones - make them simple and from commodity bits (because you are not chasing advantage through ever more weird specialist materials and architectures) and it is much easier to scale up production or have - like the predator - a steady ongoing production. Not batches of a few hundred exquisitely over engineered air frames." Except it's not. Again ... see Ukraine. We can even produce 155mm shells fast enough, and it's not because they are exquisite. "Yeah, I am, and I wonder if you are! What person watching Ukraine things "I really ought to double down on a dwindling number of extremely expensive manned platforms with increasingly expensive marginal performance advantages"." All you're doing is making assumptions. Russia is pounding on Ukraine, not the US. It's not marginal advantages. It's decisive advantages. If Ukraine had the good shit all of Russia's weapons depots would've been smoking 12 months ago. There would be no rail service for them to move material around. Their air defenses would be toast. All the T-62s in the world wouldn't help them.
murder Member	Fri Mar 10 09:42:42 "It certainly solves the problem in the RAND wargame for the Taiwan strait where the chinese airforce could literally take all the losses the entire F22 fleet could deliver, and still have enough airframes survive to destroy tankers and awacs leaving the F22s to drop into the sea even if they Chinese couldn't hit them." Those guys at RAND are brain surgeons. In the real world you blow up their air bases and the planes on the ground so you don't have to worry about it. And the solution to that problem is to provide Taiwan with a shit load of their own air defenses and long range strike capability, not refurbishing a ton of F-4 Phantoms so we can have numbers too. Numbers that of course just create a massive logistics nightmare.
murder Member	Fri Mar 10 14:46:49 WTF is this? What the hell am I looking at??? ========================= USAF Testing ‘Mutant’ Missiles That Twist In Mid-Air To Hit Their Targets The Air Force sees missiles with articulating noses as one way to take on increasingly maneuverable threats. http://www...n-mid-air-to-hit-their-targets
murder Member	Fri Mar 10 17:03:29 May as well drop this here too ... F-35 upgrades http://www.youtube.com/watch?v=7SNALUoybt0
Seb Member	Sat Mar 11 07:06:21 Jergul: I would not bet the farm on unmanned fighters without a hybrid approach first. At least for peer and near peer conflicts, remote is vulnerable and brittle. I'm not sure full autonomy is there yet. Last thing you want is for a flight of "loyal wingman" with nobody locally in the loop getting confused and bombings civilian or shooting down a jet fighter. We may however wind up progressing very rapidly to discovering pilots in practice are not needing to do anything. Hence move to optional manning for next gen manned fighters.
Seb Member	Sat Mar 11 07:35:37 Murder: I'm genuinely perplexed about this "no evidence" line. Back in the 90s they were running neural net dogfighters in simulation. What's the source of your concern here? That the electronics wouldn't stand up to extreme manoeuvres, that the software couldn't figure out how to execute them or that it wouldn't be smart enough to know how and when to pull manoeuvres to generate tactical advantage? Only the last makes any sense to me but that's not really about high g manoeuvres it's about whether the technology for autonomy works. The first two are obvious: the hardware can take high g manoeuvres, cf. Rockets, planes own avionics etc. The second, well like I said this stuff has been done in simulators back in the 90s. The fact that the planes aerodynamics were simulated and not real isn't hugely material. And a giant transformer approach would seem to work here - tokenising flight parameters and control surfaces movements with a training goal of getting certain relationships between the plane and target in a multidimensional parameter space of attitude, relative velocity, energy etc. "Why do you not believe in God?" Because the idea is incompatible with everything else I know and there's no evidence for such a phenomenon. But I would argue that an extreme doubt that drones either in proof of concept or design that fit the description above exist it could be best term development because you have not seen an autonomous supersonic drone execute climbs, dives and turns in excess of limits that would cause G-LoC is more like someone in the 1960s doubting the existence of the far side of the moon. Yes, we've never seen it, but everything we know about the moon suggests that the bit we can't see is going to be like the bits we can see, and nothing radical is required to extrapolate it. "Because there's no evidence that it can" But that explicitly isn't true. *There are plenty of examples of airframes that can pull manoeuvres that cause pilot G-LoC *Prior approaches to AI and autonomous flight have worked just fine on simulation *There is nothing special about supersonic flight conditions that would invalidate those results *The hardware such things run on can almost certainly fit in an airframe *The hardware stuff things run on would not be affected by high g manoeuvres *More modern approaches to AI would appear to be applicable to this problem too *Internationally, there are a lot of countries and defense firms approaching this problem this way "but we fly them at air shows and they do all their fancy maneuvering in plain view." We've been conducting fight tests of autonomous vehicles in plain view too. https://arstechnica.com/cars/2023/02/the-us-air-force-successfully-tested-this-ai-controlled-jet-fighter/ https://www.wired.co.uk/article/ai-fighter-pilots-automated-warfare N.b. there are good reasons why you might not demonstrate vehicles performing over and above the limits imposed manned craft. It would give adversaries information to design missiles and missiles software to. Especially if - cf. Move 37 in Alpha go - a transformer based AI for tactical combat has generated novel manoeuvres in simulation by virtue of being able to access parameter spaces. IIRC one of the interesting things in the 90s article I remember was how the algo figured out that rolling the plane helped stabilise it for acquiring a target, or some such. Its also possible that while this tech is very plausible, we are still in integration phase and there's nothing yet to demonstrate. Put it this way, if there are not AI pilots that can pull manoeuvres in existing airframes humans can't due to g forces or reaction times etc, then you can guarantee that future airframes will be designed without human constraints will be able to.
murder Member	Sat Mar 11 07:50:33 "Hence move to optional manning for next gen manned fighters." Optionally manned is manned. No one is going to produce an aircraft with a cockpit if there aren't going to be cocks in it. That's a whole lot of extra cost and weight down the shitter.
murder Member	Sat Mar 11 07:50:53 "Hence move to optional manning for next gen manned fighters." Optionally manned is manned. No one is going to produce an aircraft with a cockpit if there aren't going to be cocks in it. That's a whole lot of extra cost and weight down the shitter.
Seb Member	Sat Mar 11 07:52:14 "90% of a bombers targets are going to be fixed. Coordinates would get the job done fine. And how the fuck would AI get spoofed." SEAD would be at fleeting targets. Fixed targets are what cruise missiles are for. "Except it's not. Again ... see Ukraine. We can even produce 155mm shells fast enough, and it's not because they are exquisite." We could produce 155mm shells, we just haven't for other reasons. The limiting factor is probably casting and the explosives and explosive handling. If you wanted to start mass producing f-35s you certainly couldn't. The supply chain is far too complex. And f-22s, well the production facilities have been dismantled. You certainly could do predators though. Mass producing quad copters - you are drawing on existing and robust consumer grade supply chains that are much easier to scale. Hence Ukraine are building them from commercial grade imports. "It's not marginal advantages. It's decisive advantages. If Ukraine had the good shit all of Russia's weapons depots would've been smoking 12 months ago." Fair point, but that's largely because Russian weapons don't work as advertised, aren't operational in the scale we expected, aren't properly integrated. As such, looking just at Russia, 5th generation was arguably a waste of r&d let alone a 6th. The point about the 6th is you are planning for a war with China on the assumption China has caught up quite a lot. China has a huge manufacturing advantage. Now if Chinas technology is overrated, that's great, we beat them or deter them either way. But if it's not over-rated. Do you double down on finding some technological edge that gives comparative advantage, or seek to erode theirs? Each generation of tech so far has been higher and higher unit cost and eventually that fucks you in a numbers game. So I would think the latter. My gut is that if China's kit performs as advertised, even if it is more 4.5 than 5th gen, there's nothing coming in 6th generation that is enough to offset their numbers game. We need large numbers of commodity kit that can be produced quickly at scale by Western supply chains and performs "good enough". Plus, the West wins strategically if we can deter a challenger from fighting. So we have a strategic advantage there. So for me: drones drones drones.
Seb Member	Sat Mar 11 07:54:31 Also it's obvious that China is going to look to exploit manufacturing capacity as part of its military strategy.
murder Member	Wed Mar 15 11:54:13 The US Air Force ... "We don't have enough air dominance fighters!" Also the US Air Force ... "Lets retire 32 of our 183 F-22 Raptors!" http://www...he-boneyard-in-new-usaf-budget
Seb Member	Wed Mar 15 14:19:26 That seems odd. I assume they are an early batch that's burned though the airframes lifetime through e.g. training and they reckon they can't extend. Once the production line is shut down it can get expensive to do LEP.
murder Member	Wed Mar 29 20:13:25 http://www.youtube.com/shorts/G7taJ-yD5EY
	show deleted posts

Your Name:
Your Password:
Your Message: