Chinese dark factories are so highly automated that they don’t even need light. This is the future of manufacturing. Robots working in complete darkness. Those systems are controlled by the AI. And that makes the assembly process very effective. Human-shaped robots can make almost everything that humans can.
An advanced machine learning toolkit allows operators. To teach machines new things. very effectively. The system needs only one robot. That can make something, and then that system can scale that new skill over the entire network. That makes robots very effective workers. And maybe those ghost factories are starting to operate in Western cities. Robot factories are interesting. Especially if human-shaped robots work in that factory.
That is the thing. That makes the AI dangerous. The ability to teach one robot and then scale that thing to all robots makes it possible to teach robot armies very effectively. The same human-shaped robot that can make cars. It can act as a robot soldier. That is one thing that we must realize. The second thing is that kinds of highly automated systems can also make other products than cars. The new stealth systems require ultimate secrecy. And the robot workers are a solution to that problem.
The robot worker doesn’t call the newspaper. If it makes something that is not public. Another thing is that the robot worker can change its role by pressing a button. Flexible CAD/CAM (Computer-Aided Design/ Computer Aided Manufacturing) platforms that turn the CAD drawings straight to the product. Mean that the system can change the product in minutes. The factory involves 3D printers and machine workers. Those systems can make everything. They have instructions and raw materials. Those factories can make cars, aircraft, and other robots.
That kind of thing can make many things more effectively than humans. That means those systems are the ultimate competitors for Western factories. They can make products without brakes. And that makes them ultimate operators in the military and civil sectors. The Chinese government, like all other actors, might be interested about the robot factory as a military robot producer. There is a possibility that robot factories are installed on ships. and they can build customized robots or other robot factories if they have enough raw materials. Or the ability to search for raw materials from nature.
One remarkable thing in modern innovations is that. AI and computing are in the middle of those systems. The difference is this. In the past, computers were only auxiliary systems. In modern systems. Computing is in the middle of the systems. That means those systems are a combination of mechanics that advanced computer programs called AI control. That requires adaptation and a new way to think.
When progress stops, we must ask why. That is one of the things that we always forget. When somebody stops progress, there must be some reason for that. And what is the reason? We always hear tales that the supreme leaders give orders for stopping projects, or some lawmakers make laws that deny the use of new things. The problem is that we forget to ask why the old thing seems better.
The main problem with AI development is this: when those systems and algorithms can do something new. The idea is taken from Maslow's Hierarchy of Needs. When somebody creates something new, that need that was satisfied will be replaced by some new need. The need for R&D process starts when somebody notices a new need. And then some other actor must respond to that need by making and satisfying the new need.
This means that computers require more and more power. When the AI and computers can make something new, a new need arises for the system to do. If the customer feels that they need new tools, they will buy new machines and programs. Then the skill-scale expands, requiring new and more effective ways to compute. That means more effective microchips and more effective systems.
The new tools should make processes easier to control, but for some reason, the innovations face resistance. When a high leader orders to stop a new project, the question is why that person makes that kind of decision. The problem with that kind of way to think is that there are also other leaders, leaders who have different goals and different ways to think.
When we look at things like AI and other things, we always see dystopian visions. That’s why we want to stop AI development. The arguments for that thing are sometimes very good. But the problem is that the person who gave the warning will begin his own AI development project.
And there are always people who resist AI, because it has the ability to operate in the military field. But then the next step is that we are not prepared for the situation that some other governments are interested in, especially this and its military applications.
The fact is that we are in the middle of the break. Old-fashion tools and actors are stepping away from people’s eyes. Actors like Intel that dominated microchip production 30 years ago are falling. And their challengers from the U.S. and other nations are rising. That means there are new actors who want to get leading positions in AI development. And why AI development is so important. AI is the soul of the machine. Everything that a robot can do depends on those tools.
The ability to respond to needs is vital in the modern marketing ecosystem.
The new technology called General-purpose computing on graphics processing units, GPGPU or GPU compute, means a system that delegates missions that traditionally belong to the CPU to the graphics processors. Those kinds of abilities are raising many visions in front of our eyes. That makes it possible to develop computers where main operations happen in the GPU. This technology makes it possible to create autonomous robots with complicated mission modules.
The problem with killer robots is that somebody makes them. Somebody wants to change warehouse robots into general-purpose terminators. The problem with robots is that they are multipurpose tools. The same robot that operates in warehouses can also operate on the battlefield. This is the new type of threat. But the problem is that nobody controls AI development. Even if there are some rules in the Western world, there are always governments that don’t have the same values. Those governments' prime objective for AI and its use is to use it as a control tool.
The problem with the human operators is that they can turn rebellious. But in the same way, machines can rebel. If some hackers can put the wrong programs in those systems. There is a theoretical possibility that a hacker can attack the robot army, and then turn those robots against their masters. Or somebody can slip a combat program in the computers that control warehouse robots. Robots are not toys anymore. They are systems that can make many things.
And one of them is to go to war for humans who made them. When we make decisions, we must remember that we live in the Western world. Not all people live in countries where laws and values are like those we are accustomed to. Some leaders think differently from our leaders. Denying things doesn't end their existence.
AI is the big thing. And the AI gets the best results when it should handle precise and logical information that is introduced clearly and follows a certain type of information. When information is in text form. And it contains things like a precisely introduced source list that is easy to detect, the AI is more effective than humans.
The effectiveness doesn't mean that the AI understands what it should do. It just uses some parameters for selecting words that it connects. This is one of the most critical things for AI. When AI searches for information from the net, it seems to be very smart.
ChatGPT answered physics examinations like C-grade students. And that thing is one of the most interesting things in the world. ChatGPT would answer exams better if the questions are mathematical. And that is one of the biggest values that this AI has.
The ChatGPT will make the report by connecting words. And one of the most interesting things it could do is connect new modules to itself. The ChatGPT-type artificial intelligence might have the ability to connect itself to things like chess programs. That gives ChatGPT the possibility to improve and develop its skills.
The effectiveness of AI depends on the accumulation of information. When there is lots of information and lots of connections in the databases of the AI. That increases its ability to operate independently.
The open-source way is effective for things like ChatGPT. And, of course, all other AI applications can use similar development tools. The open development environment allows lots of people can fix the answers that the AI can give.
If the AI gives the wrong answer. The developer can fix the answer into the right mode. If the developer can input the right homepages to the sources, and then introduce things like what kind of search words the AI should use in certain questions. That allows the AI to create new connections.
The console can ask what kind of topic words or keywords the question includes. If the question is about things that won 100 m runs in some Olympic games the developer could include the keywords like "Olympic games" in the description of how the AI could search for information.
Developing AI is always a big thing. The researchers should know if the answer is right or wrong. So the users could have a form where they can tell if the answer is right or wrong. And then the AI can store things like where it took the information, and what algorithms it used. That allows for creating more effective error handling.
In this kind of project error handling and error detection are important things. If the AI controls things like robots, this part is the most dangerous part of the project. In that sequence, the AI operates in its natural element. The problem with AI is it requires a description of every possible case in its algorithm. Programmers must make those algorithms, and that is the case when something can go wrong.
And if some parts are missing. The result could be a catastrophe. These are the most time-consuming parts of the programming project. If the AI operates mainly. That doesn't mean that is suitable for everyday operations.
When some AI searches data from the internet. That thing doesn't mean that it cannot make mistakes. Mistakes in the case that AI gives wrong or funny answers are the thing that can cause a laugh. But if robot cars increase speed, that thing causes very bad things like risking human lives. But otherwise, also human drivers can push the pedals at the wrong moment.
The thing is that AI can make bad things. And still today AI requires humans to control codes. But the development of AI is going on. That development is hard work, and millions of hardcore specialists are working on those projects. When we are talking about AI and other projects, we facing one thing. Only errors and mistakes are things that cause headlines.
The next-generation version of the fully automatized war machines is probably the AI-controlled jet fighter. The AI control can be useful in kamikaze- drone fighters that can be retired fighters that would be scrapped anyway. And that kind of drone is suitable for kamikaze missions. The kamikaze drones can also use sub-missiles against air defense or some other targets in their route.
Still today human pilot wins the AI. But the fact is that AI is under development. The R&D process requires information collected from the tests. And the tests and follow-up data analysis with a multi-professional team where pilots, AI developers, and other technical and military specialists analyze data and then make the next plans and changes in systems.
The major problem with the systems is that they are mechanically repeating. When a human pilot learns the mechanical and frequently repeating moves that target is easy to shoot down. But AI advances. And the success of drones in Ukraine is making it necessary to re-estimate the drones. Even if the drone's primary role is target recognition and fire control. It can equip with missiles.
That forces the air defense to shoot the drone. Small-size quadcopters that drop hand grenades to enemy positions are also effective tools. Those drones can communicate with controllers by using radio systems or they can use laser communication. And laser communication tools can use for target marking and laser sights.
Above: GLSDB
The laser sight is also suitable for laser microphones. And that makes this system a multipurpose tool. They can search for snipers from the roofs. And if those systems are armed, they can be extremely dangerous. Those quadcopters can take out snipers and destroy enemy positions. And that system can also detect enemy vehicles for artillery and aircraft.
The quadcopter can drop from satellites, and if they have satellite-based communication that allows them to locate and attack targets, and observe secured areas. That location is on the other side of the world. The system can try to drop the bomb inside the hatches of missile silos. And that can damage or destroy the rocket inside its silo.
In the future, the SDBs (Small-Diameter Bombs) can be shot against enemy targets from thousands of kilometers away from their targets with pinpoint accuracy. And those systems can terminate even single tanks or other vehicles from another side of the world.
AI is the game changer. The next-generation versions of the Maverick missiles can have fire-and-forget mode. That kind of system is possible to create, by driving the similar guidance system that Javelin missiles use to those old-fashion missiles.
Satellites can also point targets by using lasers. Or they can drop the GPS-controlled drone on the target's roof. And then those drones can aim at the position of the target. Another possibility is that drone sends radio signals for radar-homing missiles. That means drones operate like transponders. That thing can give even naval missiles good accuracy. And it's luck that Russians don't use that technology.
The new tool on the battlefield is Ground-Launched Small-Diameter Bomb or GLSDB. The missile can shoot a small-diameter bomb to 250 kilometers away from the launch site. A similar technology used in GLSDB can install in ICBM missiles. That makes it possible to destroy targets with pinpoint accuracy from thousands of kilometers.
There is the possibility that the conventional glide bombs. That is equipped with heat shields that can drop from orbiting trajectories. Also, modern nuclear weapons can use similar technology. That is used in pinpoint accurate bombs. The JDAM with nuclear detonator already exists. And the most modern nuclear weapons are very small.
General Petraeus predicts that the use of robots become more general. And that means many tactics that are made for military and police work must be re-estimated. In the wrong hands, those tools are extremely dangerous weapons. They can use for searching for things like weapons and in the same way, as soldiers use them, they are suitable for the mafia and other kinds of criminals. So criminal organizations can use small drones as strikes.
Robot aircraft are not a new thing.
Robot aircraft are not a new thing. But independently operating AI-controlled jet fighters are the new thing. Remote-controlled drones used for a long time for a strike, recon, and target missions. There are mainly three types of robot aircraft. The two first are in service or entering service very soon. The third one is just under development.
Cruise missiles are also kamikaze drones. And they should be easy to shoot down. But those weapons are still in use, even if the human pilot should shoot them like in targeting practice. The bomber drone is the only modified cruise missile that drops smart bombs like JDAMs or SDBs to target.
X-47
1) Remote-controlled drones. Those drones are radio-controlled aircraft. That requires time for communication between the drone and its operator. The multi-channel- or laser-based communication systems make those systems a little bit more complicated to jam than in movies.
2) Semituatomatic drones. The killer drones are mainly this type of system. The drone will get flight parameters to some area, and when it recognizes something suspicious it tells that an operator.
Semi-automatic drones can also be manned aircraft where an active radar-based altimeter keeps the right distance to the ground. So the system corrects the errors that the pilot can make.
One version of that kind of system is the aircraft-operating ASAT. The military command center's supercomputers control the missile and the aircraft. There the computer calculates the impact point of the missile and satellite. Then the command center the aircraft's course which makes it possible to hit the satellite at the right point.
Also, things like optical-seeker bombs can have semi-automatic launching systems. The mission downloaded to the weapon. And if the seeker notices a hostile target that system opens fire. The system is similar to face recognition software. And it can be found in Israeli Spice-class glide bombs. Or developers can easily modify those systems for that purpose.
3) Full automatized AI-controlled drones. Those drones are not yet suitable for complicated missions. But R&D work around those systems is going on. That means the game is changing. And those systems are powerful tools. Even if there are many problems with prototypes, the purpose of tests is to collect data for the developers so that they can fix errors in the code and find the problems.
What is special about the painting above this text? The thing, that makes it so special is that the "painter" behind the image is the AI. AI is one of the most interesting and fastest-advancing tools in the world. The next-generation AI can be productive.
It can have virtual imagination. Which means it can interconnect existing datasets. AI is not as versatile as humans. But the AI can play chess better than humans or operate more effectively than humans in a certain area where the system needs a limited number of datasets.
The new area of AI is the new creative AI applications. AI as a painter is not a bad thing if the person wants to win things art like competitions. The AI can search for information on what kind of paintings people are willing to see. AI can use many tools for that purpose.
And the most effective is the wall where are images of famous or free paintings. Then the AI observes what images people will click larger. And then, the system can calculate how long people spend with some images. After that, the system can connect the components of the most liked paintings, and then make its version of that thing.
How accurate results the AI can get depends on the data sources. If the system has access to the user's web camera and it can make them use microphones. And make them comment on those images so the AI can search for the changes in the voice. And especially where the person looks for see where the person looks at. And then the system can interconnect those most pleasing objects into one work.
AI can make better work applications than humans. In this case, the system must only see what kind of things please the consultant.
Creative AI is an effective tool. It can make paintings and drawings by using a certain limited dataset. And that kind of tool can make it possible to cheat people who elect people to the workplace. Labor consultants are people who are reading many job applications in a day. So if the AI has access to the computer of the employment service the AI can search what kind of applications those consultants are open to most, and how much time they spend with the application.
And then how many of those job applications are bringing interviews? Then the AI can search for components that please the consultant. And after that. The AI can render the image to please the consultant and then collect the components for an application. That should bring the interview. The applicant must just give parameters for the AI like what are the strengths and skills that the applicant has. And the AI modified those things and photographs into nice-looking applications.
The thing is that the development of new concepts is not a very easy mission. At least, if that concept travels over Mach 2. The problem with jet fighters or multi-role jet fighters is that they are much more than some aerodynamic bodies. Things like computers, computer programs, and other kinds of things must operate in a very highly complicated entirety.
If there are some kind of errors in those systems and their cooperation. That thing can cause destruction. The artificial intelligence that controls weapon systems and avionics should make these kinds of things more complicated than ever before.
In manned aircraft, artificial intelligence can deny the aircraft touch ground in low-level flight. The radar- or lidar-based altimeter, along with the AI-boosted computer keeps the aircraft over the ground. If the flight altitude is only a couple of meters. So the AI fixes errors that the pilot can make. It can detect incoming missiles and start counter-actions if the pilot doesn't notice them.
When we think of the Su-57, it can be very capable against old fashion jetfighters like F-16, Panavia Tornado, or F/A-18. But the thing is that those old-fashion-looking aircraft can also have updates like AI-based avionics or fire control.
The reality is that capacity of the Su-57 will be seen only when it faces some jet fighter in air combat. And another thing that means in those situations is the pilot's skills and also how the jetfighter can operate with its support teams like air defense.
Su-57
"Black Swan" in air combat. When extremely extraordinary situations give wrong signals.
When we look at history. The first British aircraft that dropped Messerschmitt Bf-109 in air combat was the Fairey Battle. The Fairey Battle was not a very big success. And maybe that was the only case when that aircraft dropped another aircraft.
The detail of that case was that the rear-gunner of Fairey Battle dropped that Messerschmitt. But that victory was the first RAF air victory in WWII. If we look only at that case we might believe that Fairey Battle was a good combat aircraft. It was not a good plane at all. It served as a training aircraft and target tug in WWII.
If we want to compare that case with something ordinary we can take a case where the goalkeeper of the ice-hockey team makes the goal. We always remember that kind of goal because they are extraordinary. We don't even remember, was that team won or lose that match, or who made other goals. Those other goals are ordinary things. But that only goal that the goalkeeper made is what we remember. And the reason for that is that goal was extraordinary.
Fairey Battle had been withdrawn from service in 1940. But that case shows that we should not make generalizations by using one example. That means that the "Black Swans" or extremely good luck can give false signals from the aircraft's capacities or some other equipment.
There are billions of things that can go wrong in combat between two aircraft. The pilot can simply make mistakes and fly straight to the front of other aircraft or impact the ground. Or some hydraulic tubes can come off. Because somebody forgot tight the connector.
