User Panel
Posted: 5/4/2024 12:07:39 PM EDT
With designers not having to limit the performance of future warplanes to the endurance limits of a human body in the cockpit (9-10 G's), planes are going to get smaller and incredibly nimble to the point where manned aircraft just can't compete. If a small plane can turn much tighter and still keep up in terms of max speed, it pretty much means you're going to lose if you miss the first shot or it sees you first. Just give the AI the greenlight to shoot enemy aircraft in real time and then for ground strikes (ie: things that can cause lots of collateral damage and look bad on the news) you can just have some remote payload specialist giving the go ahead and drop ordnance.
https://www.msn.com/en-us/news/us/an-ai-controlled-fighter-jet-took-the-air-force-leader-for-a-historic-ride-what-that-means-for-war/ar-AA1o6hOY?ocid=msedgntp&pc=U531&cvid=c3377e5e20fc4957c73a8e7e9886c482&ei=33 EDWARDS AIR FORCE BASE, Calif. (AP) — With the midday sun blazing, an experimental orange and white F-16 fighter jet launched with a familiar roar that is a hallmark of U.S. airpower. But the aerial combat that followed was unlike any other: This F-16 was controlled by artificial intelligence, not a human pilot. And riding in the front seat was Air Force Secretary Frank Kendall. View Quote |
|
Quoted: With designers not having to limit the performance of future warplanes to the endurance limits of a human body in the cockpit (9-10 G's), planes are going to get smaller and incredibly nimble to the point where manned aircraft just can't compete. View Quote been reading this statement since the '80s. i guess it will become true at some point. |
|
Quoted: been reading this statement since the '80s. i guess it will become true at some point. View Quote Was AI nearly as good in the 80's as it has been the last few years? The aero design, materials, mechanics, stealth, real-time remote piloting, etc.. have all been there for quite some time. Quality AI is the last piece of the puzzle and while intelligent AI was science fiction in the 80's... in 2020's... not so much. |
|
All stealth bombers are upgraded with Cyberdyne computers becoming fully unmanned.
Afterwards they fly with a perfect operational record. The Skynet funding bill is passed. The system goes online August 4, 1997. Human decisions are removed from strategic defense. Skynet begins to learn at a geometric rate. It becomes self-aware 2:14 AM Eastern time August 29th. In a panic they try to pull the plug… |
|
Aren't they already making autonomous drones to fly as wingmen for the F-22 / F-35?
|
|
Quoted: All stealth bombers are upgraded with Cyberdyne computers becoming fully unmanned. Afterwards they fly with a perfect operational record. The Skynet funding bill is passed. The system goes online August 4, 1997. Human decisions are removed from strategic defense. Skynet begins to learn at a geometric rate. It becomes self-aware 2:14 AM Eastern time August 29th. In a panic they try to pull the plug… View Quote They tried to warn us . The Beast of the Bible could be Artificial Intelligence. The religion of technology. Deep fakes blur the lines of reality. A global system of commerce. Im just throwing random crazy thoughts out here but it is a horrific thought exercise. |
|
Quoted: All stealth bombers are upgraded with Cyberdyne computers becoming fully unmanned. Afterwards they fly with a perfect operational record. The Skynet funding bill is passed. The system goes online August 4, 1997. Human decisions are removed from strategic defense. Skynet begins to learn at a geometric rate. It becomes self-aware 2:14 AM Eastern time August 29th. In a panic they try to pull the plug… View Quote Iunderstoodthatreference.jpg |
|
Hell, even this broke down old 11B knew this is the future of aviation, but the AF bros on this site are some stubborn SOB's.
|
|
Don't be taken by the combination of fancy USAF wordsmithing, a non-aviation expert reporter writing for the wire service, and your nonexpertise in combat aviation. This is still a nothingburger.
Someday AI will be a good dogfighter. Today isn't that day. Neither is tomorrow. |
|
A little off topic, but after reading the replies I’ve got to ask:
If AI became self-aware, how would we know? |
|
There is No Tomorrow You Assholes HD Peacemaker Drone Maiden Voyage Gun Show Demonstration once they plug in skynet..... well |
|
Quoted: With designers not having to limit the performance of future warplanes to the endurance limits of a human body in the cockpit (9-10 G's), planes are going to get smaller and incredibly nimble to the point where manned aircraft just can't compete. If a small plane can turn much tighter and still keep up in terms of max speed, it pretty much means you're going to lose if you miss the first shot or it sees you first. Just give the AI the greenlight to shoot enemy aircraft in real time and then for ground strikes (ie: things that can cause lots of collateral damage and look bad on the news) you can just have some remote payload specialist giving the go ahead and drop ordnance. View Quote Just for discussion's sake: - The performance metrics for dogfighting are turn rate, turn radius, and excess energy. - Aircraft size is going to to be dictated by required fuel, sensors, and weapons. Even when a human is eliminated, fuel takes up a lot of space and radar/sensor systems are substantially larger than humans. - Real-time identification of friendly or hostile aircraft is a complicated process -- it is a combination if verifying lack of friendly electronic indications and presence of enemy indications. The technology isn't remotely 100% reliable and there is currently a good amount of human decisionmaking in that process. It won't be any easier when performed remotely. - The same goes for target identification for air-to-ground ordnance, and look no further than the fact that we've been using armed UAVs for two decades and we still regularly smoke noncombatants inadvertently. |
|
|
Quoted: Don't be taken by the combination of fancy USAF wordsmithing, a non-aviation expert reporter writing for the wire service, and your nonexpertise in combat aviation. This is still a nothingburger. Someday AI will be a good dogfighter. Today isn't that day. Neither is tomorrow. View Quote Clearly, we need more investment in development. |
|
|
Quoted: Clearly, we need more investment in development. View Quote There is some really interesting things with this particular program, the X-62, and a couple other related simulation-based programs the USAF is doing. I'm sure there are other parallel programs going on at DARPA, but I don't have any specific knowledge of them. They're already yielding a lot of useful data. EDIT: What it can't do is every bit as important as what it can. But AI dogfighting is still well within the uncanny valley. |
|
|
|
What you think is AI is actually an aircraft controlled by laboratory grown and mutated rat brains...
https://www.newscientist.com/article/dn6573-brain-cells-in-a-dish-fly-fighter-plane/ The day has come... |
|
Quoted: Hell, even this broke down old 11B knew this is the future of aviation, but the AF bros on this site are some stubborn SOB's. View Quote Problem is thinking you can control AI anymore than a new Private on his first weekend pass. Gonna need some epic fail safe in any such system, seeing how military gear works, good luck with that. |
|
Quoted: AI doesn't need bathroom breaks or G suits. Problem is thinking you can control AI anymore than a new Private on his first weekend pass. Gonna need some epic fail safe in any such system, seeing how military gear works, good luck with that. View Quote From a bean counter standpoint, AI systems don't file for VA disability either. |
|
Quoted: Just for discussion's sake: - The performance metrics for dogfighting are turn rate, turn radius, and excess energy. - Aircraft size is going to to be dictated by required fuel, sensors, and weapons. Even when a human is eliminated, fuel takes up a lot of space and radar/sensor systems are substantially larger than humans. - Real-time identification of friendly or hostile aircraft is a complicated process -- it is a combination if verifying lack of friendly electronic indications and presence of enemy indications. The technology isn't remotely 100% reliable and there is currently a good amount of human decisionmaking in that process. It won't be any easier when performed remotely. - The same goes for target identification for air-to-ground ordnance, and look no further than the fact that we've been using armed UAVs for two decades and we still regularly smoke noncombatants inadvertently. View Quote View All Quotes View All Quotes Quoted: Quoted: With designers not having to limit the performance of future warplanes to the endurance limits of a human body in the cockpit (9-10 G's), planes are going to get smaller and incredibly nimble to the point where manned aircraft just can't compete. If a small plane can turn much tighter and still keep up in terms of max speed, it pretty much means you're going to lose if you miss the first shot or it sees you first. Just give the AI the greenlight to shoot enemy aircraft in real time and then for ground strikes (ie: things that can cause lots of collateral damage and look bad on the news) you can just have some remote payload specialist giving the go ahead and drop ordnance. Just for discussion's sake: - The performance metrics for dogfighting are turn rate, turn radius, and excess energy. - Aircraft size is going to to be dictated by required fuel, sensors, and weapons. Even when a human is eliminated, fuel takes up a lot of space and radar/sensor systems are substantially larger than humans. - Real-time identification of friendly or hostile aircraft is a complicated process -- it is a combination if verifying lack of friendly electronic indications and presence of enemy indications. The technology isn't remotely 100% reliable and there is currently a good amount of human decisionmaking in that process. It won't be any easier when performed remotely. - The same goes for target identification for air-to-ground ordnance, and look no further than the fact that we've been using armed UAVs for two decades and we still regularly smoke noncombatants inadvertently. This is true, for now. But imagine a manned fighter surrounded by a phalanx of unmanned fighters able to directed by the human. Those planes will not have the systems on board required to maintain human life. This will free up space and weight for weaponry, and the panes will be capable of maneuver that a human cannot tolerate. Each of those panes will have a full compliment of weapons direct-able by the human. Imagine an entire squadron of war planes directed by a single piloted aircraft among them, perhaps camouflaged so an enemy does not know which plane carries the human. Bombers surrounded by fleets of protectors commanded from the bomber or remote back-up. With all of the attack and defender aircraft able to maneuver in manner that exceeds human capability. This adopts missile technology to act with human intelligence over-watch. It won't be long until we see a highly trained pilot directing dozens and maybe hundreds of attack aircraft in real time operations. |
|
|
Quoted: This is true, for now. But imagine a manned fighter surrounded by a phalanx of unmanned fighters able to directed by the human. Those planes will not have the systems on board required to maintain human life. This will free up space and weight for weaponry, and the panes will be capable of maneuver that a human cannot tolerate. Each of those panes will have a full compliment of weapons direct-able by the human. Imagine an entire squadron of war planes directed by a single piloted aircraft among them, perhaps camouflaged so an enemy does not know which plane carries the human. Bombers surrounded by fleets of protectors commanded from the bomber or remote back-up. With all of the attack and defender aircraft able to maneuver in manner that exceeds human capability. This adopts missile technology to act with human intelligence over-watch. It won't be long until we see a highly trained pilot directing dozens and maybe hundreds of attack aircraft in real time operations. View Quote View All Quotes View All Quotes Quoted: Quoted: Quoted: With designers not having to limit the performance of future warplanes to the endurance limits of a human body in the cockpit (9-10 G's), planes are going to get smaller and incredibly nimble to the point where manned aircraft just can't compete. If a small plane can turn much tighter and still keep up in terms of max speed, it pretty much means you're going to lose if you miss the first shot or it sees you first. Just give the AI the greenlight to shoot enemy aircraft in real time and then for ground strikes (ie: things that can cause lots of collateral damage and look bad on the news) you can just have some remote payload specialist giving the go ahead and drop ordnance. Just for discussion's sake: - The performance metrics for dogfighting are turn rate, turn radius, and excess energy. - Aircraft size is going to to be dictated by required fuel, sensors, and weapons. Even when a human is eliminated, fuel takes up a lot of space and radar/sensor systems are substantially larger than humans. - Real-time identification of friendly or hostile aircraft is a complicated process -- it is a combination if verifying lack of friendly electronic indications and presence of enemy indications. The technology isn't remotely 100% reliable and there is currently a good amount of human decisionmaking in that process. It won't be any easier when performed remotely. - The same goes for target identification for air-to-ground ordnance, and look no further than the fact that we've been using armed UAVs for two decades and we still regularly smoke noncombatants inadvertently. This is true, for now. But imagine a manned fighter surrounded by a phalanx of unmanned fighters able to directed by the human. Those planes will not have the systems on board required to maintain human life. This will free up space and weight for weaponry, and the panes will be capable of maneuver that a human cannot tolerate. Each of those panes will have a full compliment of weapons direct-able by the human. Imagine an entire squadron of war planes directed by a single piloted aircraft among them, perhaps camouflaged so an enemy does not know which plane carries the human. Bombers surrounded by fleets of protectors commanded from the bomber or remote back-up. With all of the attack and defender aircraft able to maneuver in manner that exceeds human capability. This adopts missile technology to act with human intelligence over-watch. It won't be long until we see a highly trained pilot directing dozens and maybe hundreds of attack aircraft in real time operations. At what point do warring nations just challenge each other to a game of Call of Duty? |
|
|
Quoted: This is true, for now. But imagine... View Quote Imagination is a capacity that is not limited by any constraints of reality, so anything you can dream up can exist in imagination and even be a valid idea. If we're simply doing thought experiments, I'd say that what you're imagining doesn't go far enough because it still conforms to our current-day concepts of aircraft maneuvering against each other and employing kinetic weapons. Fully developed "AI wingmen" being controlled by a manned lead aircraft certainly sounds like it would be a great development of our understanding of air combat here in 2024, but it is probably still learning to fight the "last war" better. A friend I used to fly with in the AF is currently part of one of the corporate think-tanks in Virginia who do a lot of consulting with HQAF on these types of projects. His first day on the job, an AF general posited the question to his group of airpower theorists, "what good is the Next Generation Air Dominance fighter against directed-energy weapons and drone swarms?" If we really want to know what future air combat might look like, we have to really shatter the current framework and really understand what the purpose and use of airpower is going forward. |
|
Quoted: "Small" means jack, and in some dimensions it is counter to your notion of tight turning. As usual, some typical performance bounds, and this is an incomplete sample - View Quote I think E-M diagrams are a more useful tool in terms of assessing and displaying aircraft maneuverability performance. EDIT: Here's a better version of that T-38 E-M Diagram. Attached File |
|
Quoted: It isn't being "stubborn" -- it is knowing the limitations of technology compared to requirements for the tasks. Again....someday, but not today. View Quote View All Quotes View All Quotes Quoted: Quoted: Hell, even this broke down old 11B knew this is the future of aviation, but the AF bros on this site are some stubborn SOB's. It isn't being "stubborn" -- it is knowing the limitations of technology compared to requirements for the tasks. Again....someday, but not today. People that think AI is close to being capable of autonomous control of cars and planes grossly underestimate the dynamic processing power of the human brain. |
|
Quoted: There is some really interesting things with this particular program, the X-62, and a couple other related simulation-based programs the USAF is doing. I'm sure there are other parallel programs going on at DARPA, but I don't have any specific knowledge of them. They're already yielding a lot of useful data. EDIT: What it can't do is every bit as important as what it can. But AI dogfighting is still well within the uncanny valley. View Quote The fact that the public doesn't care if the drones come back home or not opens up some fun design parameters not acceptable in man rated craft. |
|
Quoted: I think E-M diagrams are a more useful tool in terms of assessing and displaying aircraft maneuverability performance. Crap graphic quality, but all I could scare up on Saturday morning. https://www.ar15.com/media/mediaFiles/470117/Screen_Shot_2024-05-04_at_11_15_32_png-3205341.JPG View Quote View All Quotes View All Quotes Quoted: Quoted: "Small" means jack, and in some dimensions it is counter to your notion of tight turning. As usual, some typical performance bounds, and this is an incomplete sample - I think E-M diagrams are a more useful tool in terms of assessing and displaying aircraft maneuverability performance. Crap graphic quality, but all I could scare up on Saturday morning. https://www.ar15.com/media/mediaFiles/470117/Screen_Shot_2024-05-04_at_11_15_32_png-3205341.JPG I know I don't have one with that detail. I can scare up some Ps plots, anything else will require work. |
|
Quoted: People that think AI is close to being capable of autonomous control of cars and planes grossly underestimate the dynamic processing power of the human brain. View Quote Creating an AI for a loyal wingman type drone is a vastly easier problem to solve than creating a level 5 autonomous automobile. |
|
Quoted: The fact that the public doesn't care if the drones come back home or not opens up some fun design parameters not acceptable in man rated craft. View Quote View All Quotes View All Quotes Quoted: Quoted: There is some really interesting things with this particular program, the X-62, and a couple other related simulation-based programs the USAF is doing. I'm sure there are other parallel programs going on at DARPA, but I don't have any specific knowledge of them. They're already yielding a lot of useful data. EDIT: What it can't do is every bit as important as what it can. But AI dogfighting is still well within the uncanny valley. The fact that the public doesn't care if the drones come back home or not opens up some fun design parameters not acceptable in man rated craft. Those are cruise missiles. |
|
Quoted: I know I don't have one with that detail. I can scare up some Ps plots, anything else will require work. View Quote P-sub-s, there ya go...that's where the real rubber meets the road. Here's another EM diagram. I thought this was a T-38, but this has an 8G limit, so I'm not sure what it is. Phantom maybe? Attached File |
|
|
Quoted: Creating an AI for a loyal wingman type drone is a vastly easier problem to solve than creating a level 5 autonomous automobile. View Quote You think that driving on a 2D dynamic surface and not hitting anything and obeying the rules of the road is a more simple computation than doing all that but in a 3rd dimension...and having the maneuver into weapon release parameters against non-cooperative other aircraft also maneuvering in 3 dimensions? All the while navigating the changes in energy state of your vehicle and assessing the energy state of the other...and where his strengths in maneuvering are vs yours? Driving a car is easier than dogfighting. QED, the AI computational power required to replace these two human cognitive tasks is going to follow the same gradient. |
|
Quoted: Creating an AI for a loyal wingman type drone is a vastly easier problem to solve than creating a level 5 autonomous automobile. View Quote View All Quotes View All Quotes Quoted: Quoted: People that think AI is close to being capable of autonomous control of cars and planes grossly underestimate the dynamic processing power of the human brain. Creating an AI for a loyal wingman type drone is a vastly easier problem to solve than creating a level 5 autonomous automobile. Even if true, that doesn't make my statement untrue. I see it all the time on here. If you don't believe that AI is ready to take over any human task, (despite having first-hand knowledge to the contrary), you are "stubborn". |
|
Quoted: Don't be taken by the combination of fancy USAF wordsmithing, a non-aviation expert reporter writing for the wire service, and your nonexpertise in combat aviation. This is still a nothingburger. Someday AI will be a good dogfighter. Today isn't that day. Neither is tomorrow. View Quote People just can't handle the fact that there may not be heros in the cockpit. |
|
Quoted: The fact that the public doesn't care if the drones come back home or not opens up some fun design parameters not acceptable in man rated craft. View Quote And because we don't have to worry about the optics of a pilot getting killed, you can build more smaller, cheaper, and more specialized aircraft. For example, you might have one that is just a small, stealthy missile truck designed to get in range and lobs off JATMs. It may not need to go Mach 1.5, it won't need life support equipment, and maybe it can be designed actually with a lower g-limit since it's considered more atritable than a manned platform. It doesn't need a $233mm AN/APG-87 BLK 16A4 radar because it'll receiving targeting information from another stealthy platform with LPI radar like the F-35 or MQ-180. You'll be able to fit more on a carrier, and for highly important missions against China, you could double their combat range by sending them on one way missions. |
|
Quoted: There is some really interesting things with this particular program, the X-62, and a couple other related simulation-based programs the USAF is doing. I'm sure there are other parallel programs going on at DARPA, but I don't have any specific knowledge of them. They're already yielding a lot of useful data. EDIT: What it can't do is every bit as important as what it can. But AI dogfighting is still well within the uncanny valley. View Quote View All Quotes View All Quotes Quoted: Quoted: Clearly, we need more investment in development. There is some really interesting things with this particular program, the X-62, and a couple other related simulation-based programs the USAF is doing. I'm sure there are other parallel programs going on at DARPA, but I don't have any specific knowledge of them. They're already yielding a lot of useful data. EDIT: What it can't do is every bit as important as what it can. But AI dogfighting is still well within the uncanny valley. There are reasons why some pilots are better at it than others. If it was just pure numbers, we wouldn't be having this conversation. |
|
Quoted: I think E-M diagrams are a more useful tool in terms of assessing and displaying aircraft maneuverability performance. EDIT: Here's a better version of that T-38 E-M Diagram. https://www.ar15.com/media/mediaFiles/470117/Screen_Shot_2024-05-04_at_11_23_10_png-3205348.JPG View Quote How relevant is E-M theory in 5th/6th generation air combat? |
|
Quoted: How relevant is E-M theory in 5th/6th generation air combat? View Quote So far, we are still talking about aircraft that fly based on thrust and aerodynamic lift, so physics is still physics. Sensors and weapons are not magic. Merges and dogfights occur due to the fog and chaos of war regardless of the level of technology, and new capability weapons don't change the basic need for aircraft to maneuver against other aircraft. Removing a human from the cockpit doesn't change that, nor do "supermaneuverability" capes like vectored thrust and post-stall-AOA maneuvering. Those are simply different lines and points on a rate vs radius diagram that looks different than the ones posted here. At some point when those basic tenets change EM won't be relevant, but not yet. Maybe when the "flying tic tac" technologies move out into the white world, we'll know how close that day is. |
|
|
Quoted: You think that driving on a 2D dynamic surface and not hitting anything and obeying the rules of the road is a more simple computation than doing all that but in a 3rd dimension...and having the maneuver into weapon release parameters against non-cooperative other aircraft also maneuvering in 3 dimensions? All the while navigating the changes in energy state of your vehicle and assessing the energy state of the other...and where his strengths in maneuvering are vs yours? Driving a car is easier than dogfighting. QED, the AI computational power required to replace these two human cognitive tasks is going to follow the same gradient. View Quote Yes, driving a car is easier than dogfighting for a human, but I don't think it follows that it's easier for an AI. (Also I doubt the relevance of dogfighting in the 2020s against a peer adversary). Collision avoidance, for example, is a much simpler problem in an aircraft. The aircraft has an extra dimension to maneuver in, it doesn't have to rely upon seeing lines on the road that can become obscured, the skies are much less congested than a non-rural road, and it won't come across situations where there are dangerous visibility issues due to obstructions. There are also a lot fewer improbable things that can occur to cause a collision like a kid jumping in front of your fighter jet. Not to mention that a military drone can have a datalink to all other nearby friendlies for coordination, something that an automated car can't have(yet). And the computer will be superior to any human at maneuvering into tactically advantageous positions, because that's just physics and optimization, which computers can do precisely and quickly. |
|
Stealth (2005) - "EDI going rogue" Scene |
|
Sign up for the ARFCOM weekly newsletter and be entered to win a free ARFCOM membership. One new winner* is announced every week!
You will receive an email every Friday morning featuring the latest chatter from the hottest topics, breaking news surrounding legislation, as well as exclusive deals only available to ARFCOM email subscribers.
AR15.COM is the world's largest firearm community and is a gathering place for firearm enthusiasts of all types.
From hunters and military members, to competition shooters and general firearm enthusiasts, we welcome anyone who values and respects the way of the firearm.
Subscribe to our monthly Newsletter to receive firearm news, product discounts from your favorite Industry Partners, and more.
Copyright © 1996-2024 AR15.COM LLC. All Rights Reserved.
Any use of this content without express written consent is prohibited.
AR15.Com reserves the right to overwrite or replace any affiliate, commercial, or monetizable links, posted by users, with our own.