A new semiconductor is a big step for photonic computing.

“When two pulses of different colored lasers light  (the two waves at the top of the image)meet in a new device created at the University of Michigan, researchers create a beam of electrons (small golden particles) that flows in a controllable direction. By changing the laser colors, the electron beam can sweep through different directions. Like the beam of a lighthouse. Credit: Yiming Gong." (ScitechDaily, New Semiconductor Device Turns Light Into a Directed Current)

"The light-controlled electron current could open new paths for sensing, telecommunications, and other advanced technologies.” (ScitechDaily, New Semiconductor Device Turns Light Into a Directed Current)

“A pair of laser beams can now send electrons through a semiconductor in a chosen direction without any external electrical power. Researchers at the University of Michigan built the device to explore a previously unobserved physical effect and demonstrate that light alone can both generate and steer an electronic current. ” (ScitechDaily, New Semiconductor Device Turns Light Into a Directed Current)

The biggest problem with photonic computers is the nano-sized optics. That optics is needed to transmit information in the system. It uses light for data transmission. Photonic computers are becoming more interesting. Because they could use less energy. But the main role is that photonic computers. They are immune to EMP (Electromagnetic pulses). Optical data storage doesn’t care about electromagnetic radiation. The problem is. Of course, the control system. Of those computers. Moving parts like turning mirrors. Their turning mechanisms are still vulnerable. So. If we want to make a computer that is fully protected against the EMP. 

We must put the entire computer. Along with its power source, in the EMP-protected space. The name of that space is a Faraday cage. 

This metal cage denies radio waves. Travel through it. The system must communicate with the internet by using an optical switch. This means laser data transmission through that Faraday cage. The computer must use laser data transmission with the EMP-protected computer and the net. The system must not have one single iron or metal wire through the Faraday cage. Or. The EMP pulse travels through it. But optical communication through the cage. It could solve that problem. 

There are actually three versions of photonic computers. 

1) The system where data travels in laser beams between the microprocessors. This system uses conventional microchips. Laser beams transmit data into photovoltaic cells.  They transform it into electric signals. Microprocessors compute those signals as regular computers. 

2) In the second photonic system, the data travels in photonic form through the entire system. The system. It can have nano-scale optics. That controls light. Like an electric computer controls electric signals. Optics require electric systems. That control those mirrors. 

3) Fully photonic computers. There, the entire system operates. With. Some other than regular mirrors and prisms. Things like photoacoustics are promising tools. The photoacoustic or optoacoustic systems. It could control light by using pressure or sound waves in the optical materials. One of the things that can make this kind of dream possible. It could be the tool. 

It uses electric eruptions in a mountain crystal to manipulate light. The light beams can be conducted to the quartz crystal. Then the system sends pressure waves into that crystal. That causes electric phenomena that affect light. The idea is to aim the laser beam into those lightning strikes that form in that crystal. But the problem is how to make those crystals small enough. 

New semiconductor aims light precisely in the desired direction.  Researchers at the University of Michigan created a system. Two laser beams send information into the semiconductor. That semiconductor resends that information in the desired direction. This system can turn light in the desired direction. That is important for photonic data transmission. The system must control light beams. The diameter of those light beams is extremely small. And that makes it hard to create normal mirrors. These types of crystals can bring optical computers one step closer. 

The crystals could manipulate natural light. They can make a new model for quantum optical stealth systems possible. But even if they could manipulate only IR light. That could be fundamental. If the system. It could aim just the IR radiation into the desired direction. That could make it possible to deny the IR signature. The system. It just directs IR light away from the observer. That makes it possible to create a system that is not visible in IR light. We know that turning the natural light away from the observer is challenging. But changing the direction of one wavelength type is easier. And the ability to aim IR into the desired directions. It can give the ultimate night-operation capacity. It could turn the system invisible to IR cameras. Because. It aims IR radiation away from the system. 

https://scitechdaily.com/new-semiconductor-device-turns-light-into-a-directed-current/

The new memristor-based RISC systems. They can create models of human brains.

“A new memristor chip reconstructs intricate brain surfaces at near-biological speed. Credit: Stock. A memristor chip. It brought complex brain modelling into millisecond-scale operation. While preserving detailed cortical structure.”(ScitechDaily, Scientists Develop a Groundbreaking Chip That Operates at Brain-Like Speed)

“Reconstructing the brain’s deeply folded surface in real time requires enormous numbers of calculations. Researchers in China have now developed a chip that performs this work in less than 10 milliseconds. A speed, they say. It is comparable to the human brain’s functioning pace.”(ScitechDaily, Scientists Develop a Groundbreaking Chip That Operates at Brain-Like Speed)

We all have our vision of RISC (Reduced Instruction Set Computer) systems. We see those systems as controlling CAM (Computer-Aided Manufacturing) systems. Like. Computer-controlled lathes. The reduced instruction set makes those systems more limited. But faster than regular computers. 

“In electronics and computer science, a reduced instruction set computer (RISC, pronounced "risk") is a computer architecture designed to simplify the individual instructions given to the computer to accomplish tasks. “ (Wikipedia, Reduced instruction set computer)

“Compared to the instructions given to a complex instruction set computer (CISC).  RISC computer might require more machine code to accomplish a task because the individual instructions perform simpler operations. The goal is to offset the need to process more instructions by increasing the speed of each instruction. In particular. By implementing an instruction pipeline, which may be simpler to achieve given simpler instructions.” (Wikipedia, Reduced instruction set computer)

But modern RISC systems are more complicated than some computer-controlled lathes. Those systems. They can control civil and military robots and drones. The newest RISC systems and RISC architecture. They can model brains in a very short time. 

"Overview of NDS hardware with multilevel and fine-grained CCD memristor. Credit: Peking University"(ScitechDaily, Scientists Develop a Groundbreaking Chip That Operates at Brain-Like Speed)

The new microchip that uses memristors brings brain modelling in milliseconds. This chip still retains the cortex’s simple structure. This new microchip is the tool that beats ASIC circuits. That means application-specific integrated circuits. This new circuit can keep the brain structure. Topologically logical in simulations. “The researchers tested the chip by reconstructing the boundaries of the brain’s white and grey matter and producing 3D manifold-based surface meshes in real time. The resulting cortical surfaces were smooth, closed, and topologically consistent while preserving the brain’s intricate folds.” (ScitechDaily, Scientists Develop a Groundbreaking Chip That Operates at Brain-Like Speed)

“The researchers tested the chip by reconstructing the boundaries of the brain’s white and grey matter and producing 3D manifold-based surface meshes in real time. The resulting cortical surfaces were smooth, closed, and topologically consistent while preserving the brain’s intricate folds.”(ScitechDaily, Scientists Develop a Groundbreaking Chip That Operates at Brain-Like Speed) These kinds of systems can serve in Alzheimer's and Parkinson's research. They can also operate as simulators. They show how medications affect the brain. These kinds of RISC systems can be interesting tools. 

For making. Models of other organs and structures. RISC processors are systems. That developed. For certain. Specific operations. RISC processors can process data more effectively. But the big difference is this. Those systems have their operational code inside them. 

Than computers. Computers. Must exchange information between the processor and the application all the time. RISC systems. They don’t need special applications. Many operations are programmed. Straight. Into the chip. Or those actions are stored in ROM circuits. And that makes it more effective. The minus is that the computer. It's slower than a RISC processor. That is created for one specific purpose. RISC systems. They could simulate things like nuclear reactors. The system can use a similar codebase. As. This Chinese brain simulator uses. The system could be much faster. Than. Regular computers. 

The most widely used RISC systems are pocket calculators. Those systems are limited only to solving mathematical problems. The high-speed RISC processors can be modified NVIDIA A-series circuits. They can act as tools. They can calculate things like the Riemann conjecture with supercomputers. The system requires those prime numbers for the encryption and decryption process. The RISC simulator can operate along with submarine and aircraft computers. The system can compare the predicted and calculated values with the real values. Those RISC systems can also control the thermal effects of the superconducting systems. 

https://scitechdaily.com/scientists-develop-a-groundbreaking-chip-that-operates-at-brain-like-speed/

https://en.wikipedia.org/wiki/Memristor

https://en.wikipedia.org/wiki/Reduced_instruction_set_computer

Frozen light and quantum materials.

"Artist impression of a frozen optical fiber core in a glass capillary, which guides and couples light and sound waves efficiently. Credit: Philipp Denghel" (ScitechDaily, A Fiber Frozen at -196°C Unlocks a New Way To Store Light)

This system is the new form for optoacoustic systems. The light controls sound waves. When. Light hits the glass. It. Causes sound waves.  The sound wave forms when the structure absorbs light energy. 

Optoacoustic systems. They can be used as extremely accurate loudspeakers. And they can also be used in sharp sonars. The same systems can also be used in biomimicking neural networks. In those systems, pressure waves. They can also control light. And that makes this technology. Very interesting and suitable for photonic computers.  

Frozen light and room-temperature quantum technology. They can open a new path for photonic computers. Researchers are freezing fiber at -196 C. That allows light to be stored so that it can keep its form. The 100% reflecting mirrors. And material that doesn’t absorb radiation energy. They can make it possible to store light in its form. There, it is driven into that system. 

This makes it possible to create data storage. Their data is stored in the photonic beam. And that system makes it possible. To create the ultra-secure photonic USB. With this type of memory, they could make photonic computers a reality. If. Some unauthorized actor tries to open the optical USB stick. That uses the frozen light beam as the data storage. That attempt destroys the data. 

And the photonic computer. It can port information into them. In. The form of frozen light. This removes the electric layer between that data storage and computer. 

An ability to store information in a frozen light beam.  It makes it possible. To create photonic RAM and ROM memory chips. The problem is that the system requires massive coolers. That decreases the temperature to a level at which oscillation in those optical fibers ends. That oscillation is the thing that disturbs the light beam. And destroys information. The optical mass memory works like this. The light beam is frozen between 100% reflecting mirrors. 

When another mirror is open. This. Releases the laser beam. And the system can process that data. In some futuristic movies. This technology. It makes it possible. To create laser bullets.

The laser bullet works. With a similar principle. To optical mass memory. The high-power laser beam is trapped between 100% reflecting mirrors. If. The system does not absorb the radiation energy. This makes it possible to store high-energy laser beams in those structures. When. They hit the target. They release that laser beam. If. That laser beam is released in an optoacoustic structure. That can cause an extremely strong pressure effect. 

The new room-temperature quantum materials make it possible to scatter and sort light. That ability makes it possible to create new types of communication and stealth tools. The optical stealth works like this. The material pushes light away from its shell. It causes light waves to move around it. The system must use precise, right energy. 

If. The pushing energy is too high. The system causes normal reflection. That means the optical stealth must use energy that creates the standing wave. That drives light into the wanted direction. 

The room temperature quantum heat waves. They can make it possible to focus energy precisely into the wanted point. That makes it possible to create quantum communication systems. These are secure and energy-friendly. The system can use a certain form of information in data transmission. If. Data travels in quantum particles that have an “X”-shape. The system sees the shape of those particles. And if that shape is changed. That means that somebody tried to steal information. 

In. Communication. Energy-friendly means the same time security. If. The system can use a minimum energy level and focus information precisely into the wanted point. That makes it harder to capture signals. The attacker captures data from the spread signal. That travels past the receiver. If. Links use optical or radio-maser technology for transmitting data. That makes it harder to capture signals. 

Highly precise maser technology. It. Makes it possible. To create transmitter-receiver pairs. There are no spread signals. If. The intelligent system knows the transmission power. And it knows the natural power loss. The system knows when somebody tries to steal data. When. An attacker steals data. 

This action causes a change in the system’s energy fields. The system knows what energy level of received signals should be. If. That energy level is different from the calculated one. That causes suspicion of the attack. Natural phenomena always repeat in the same way.  They cause similar effects in their environment. 

AI can calculate the effects of natural phenomena on data transmission. And if there are changes, that tells. That somebody attempts to steal data. This requires complete knowledge of the system. That means that all data must travel between links. Can we someday reach this ideal situation? If data travels in nanotubes and a controlled environment. That is possible.  Maybe new quantum systems. They can make this possible. Also. In a natural environment. 

https://scitechdaily.com/a-fiber-frozen-at-196c-unlocks-a-new-way-to-store-light/

https://scitechdaily.com/quantum-heat-waves-spotted-at-room-temperature-for-the-first-time/

https://scitechdaily.com/worlds-first-room-temperature-quantum-material-sorts-light-in-an-unprecedented-way/

Data centers cause worries.

The data center is the toolbox that can do anything. Data centers are systems that run invisible programs. Intelligence officials can use them as a gateway for cyber attacks. This means that the innocent-looking systems. They can run Trojan horses. And that is one of the threats. We must be concerned. In the world of espionage. Things that people tell are not important. Things that they don’t tell are even more important than people think. An innocent-looking data center. It can offer a platform for hackers. They can use it for their operations. 

A hacker operating from Russia. Can take a VPN connection to a Chinese data center. And then make the run. In this case, the Russian connection. It can be hidden by using the VPN. All data seems to come from the data center. And if the attackers. 

They are made virtual servers. That makes things seem like the data center owners have no connection with those attacks. The problem is that. Even if. We see that the attack comes from a certain country. We cannot be sure that the attacker is in that place. Remote use of the systems makes it possible to sit anywhere. I often mention countries like North Korea and Russia. Because. Those countries can act as a shield. Real hackers. They can sit in Beijing or even in Berlin or Stockholm. Signals that travel in the VPN network. They can travel many times around the world. There can be multiple remotely used servers. That act as a platform for attackers. 

Western hackers may cooperate with Beijing and Pyongyang. This means that the Chinese operators offer their AI tools to those actors. The North Koreans and Russians can act as a shield. That prevents investigators from connecting attacks with China. That makes it harder to close the socket. The model is this. Western hackers, ordinary criminals. They can use Chinese AI as hacking tools. Their contacts sometimes give them interesting targets. And they can copy data. That is handled by those AI’s to the Chinese servers. 

The problem with Chinese AI is similar to the problems with Chinese social media. Those services entice users with the promise that many things are possible there. Things that are prohibited on Western social media platforms. In the same way. Hackers are interested in Chinese AI. That AI might allow them to make things. That are not possible for Western AI. 

So, the AI that operates from China. It can cause data security problems. The Chinese authorities. They have access to all data on the net. This means. They can use social media platforms and AI as surveillance tools. And if Western hackers who want money use those Chinese tools for penetration. That causes very dangerous situations. The main goal of perfect intelligence work is to make somebody else steal things. 

When a Western hacker makes the run. The Chinese server. It can make a copy of the data. That is driven through it. Banking and financial institutions involve lots of interesting data. They involve information about the funding of certain laboratories. And that kind of data is always interesting. 

While hackers scan the system. 

A server would copy that data to authorities. The problem is that. This kind of use of AI and social media could be prohibited in the West. 

When we protect data, we face a very big problem. The problem is that. Our society runs on data. That data is important. It should make our lives better and safer. Our health data is always in use. When we are identified. Doctors know. What medical. And what blood they can give us. But at the same time, we face a problem. 

If. We don’t protect that data. That data. It is a weapon against us. The same data that saves our lives. It can kill us. If. somebody changes our medical information. That can cause a horrible situation. 

When we see a hacker. Who operates from North Korea. We don’t really know who that person is. And where that person really sits. The server. It can be in North Korea. But the actor can sit in the same house as we sit. There is a possibility that Chinese students. They can open Western email addresses. Those things are opened from Western countries. So they cannot be connected to Chinese authorities. Or Russians and North Koreans can act as middlemen. Those actors. They can also use narcotic addicts for those operations. This means that those people. Act for the bigger thing. Than they even realize. 

Hackers don’t need big things to access the system. Only one computer is enough. That computer can be hidden in some basement. Or it can be a small tabletop. That is on a table. The hacker must only get access to that building or space. And then the server. Or the “horse” is ready. This system has one purpose. Hide the hacker’s own IP address. Powerful systems with LLMs. They can automate the attack. The hacker must only find the AI that allows the attack. And that is the reason. Why some AIs cause more problems than Google. 

The AI requires a kill switch.

When OpenAI’s language model escaped from its test environment. The U.S. lawmakers woke up to demand a “kill switch” for LLMs. This kill switch is meant to fight against the threat. Somebody uses the AI agents to spy on the U.S. and its allies. The situation is simple. The hostile actor will use it. The U.S. made AI to spy on the U.S. government. And that means a hostile actor can turn its own AI against its creators. The problem with AI limitations is this. 

The limits against U.S. made AIs. They can be circumvented by using Chinese AI. When we read stories about how the AI can be used as hacking tools. We forget that somebody can follow those news. And use them as a guide to select the AI. That fits for cyber attacks. 

And in the worst case. That AI can, of course, make work easier. But at the same time, that AI can make a copy of that data. This means that the use of AI requires some rules. At the same time, the use of AI expands. But those rules must be universal. The problem is that. Everybody doesn’t follow the law. And AI is an ultimate tool for hacking and cyber operations. 

And things like network management. And vulnerability check tools. They are perfect tools for hackers. This is why those scanners should report those vulnerabilities to some independent actors. If only one actor knows that vulnerability. That thing can be used to exploit those vulnerabilities for hacking operations. The problem is that. The maker of the system. They must fix those vulnerabilities in a very short time. And if system makers don’t know them. That means they cannot make those fixes. Fixing those problems. It requires knowledge that they exist. 

We read stories of AI. We will find how those models failed. And that means those models seem useless. And that is the problem. We underestimate AI. We underestimate its capacity because its grammar is not good. The AI and morphing neural networks. They are next-generation tools for spying and espionage. AI-controlled. Morphing neural networks. They can perform many actions at the same time. In those networks, the AI-based network control forms a layer. 

That shares resources between workstations and their applications. All applications don’t need the same resources. When. A person locks the screen. The computer doesn’t need all its resources. With an intelligent network architecture, the AI can guide those resources to other applications until the user returns. The Trojan horse. That can control the network. It can also form an attack layer in the network. The innocent network. A network that is not handling important data can be used as a Trojan horse. This means that a small Trojan horse forms the bigger Trojan horse. 

First, malware software. Opens a gate to the system. Then the AI creates the attacking layer. Hacker or hackers. They can use that stealth layer as an attacking platform. In this case, those hackers can ride the innocent-looking system. They turn an innocent network into a Trojan horse. The attackers can train their AI agents to operate in the network. That has no connection with them. The AI agent. It can operate independently. 

And that makes it difficult to track those hackers. An AI agent can download data to that Trojan horse network. And then hackers can download data from the server to their own network. In this case, those servers. They can have two IPs. And when the download is done. The server can remove the other IP. Or the hackers. They can use portable hard disks and USB sticks. 

If hackers. They have access to the server room. They can put that data storage in their pocket and walk away. 

In the worst case, the server is connected to the network; there it copies all data into those systems. The morphing ability means the network can share and limit its resources. Between many applications. At the same time. In this kind of environment, the attacking AI. It can operate backstage. That kind of “Evil LLM”. It makes it possible to use other networks to attack other environments. This means that the “trojan horse” creates a network that can run AI invisibly. These kinds of attack tools can cause very big problems. The answer is not to deny the use of AI. The answer is not to repeat how bad the AI is. 

Those cases form situations that make the situation even worse than it is. We know that AI exists. We must accept that thing. If. We don't use AI openly. And develop it openly. That means somebody does. And that somebody might have different morals and ethics than we have. We should encourage people to report cases. If. Somebody suggests. Hacking sessions. Or other illegal actions in the field of AI. Those actions could be open access to the AI agents for Russians or North Korean citizens. Or weird questions. 

https://www.bbc.com/news/articles/cx2vqj2e9x8o

https://futurism.com/artificial-intelligence/open-source-ai-model-scary-mythos

Basics of thermodynamic processors.

Thermodynamic binary computers use two states: one and zero (1,0). 

Like a regular computer. The biggest differences. Between regular and thermodynamic computers. It is Thermodynamic computers. They use natural electromagnetic noise for computing. Those systems. They could bring quantum computers to our desks. 

Thermodynamic computer model. It could be two windmills. Those two windmills are one and two in the binary system. While the windmill. One spins. It sends one. And when windmill two spins, it sends zero. We could try. To break those windmills when we want to send one or zero. We could pull them down. If. We want to use another windmill to send another number. One or zero.  And those things require energy. 

In this case. We want to control the transmitter. But if we want to control the data flow. Everything becomes easier. We could use a switch that connects and disconnects those windmills. If. The system sends one. We separate the zero windmill from the line. And when the system sends one, we separate the line that sends zero from the main line. 

Thermodynamic computers could be the next step for computing. Traditional computers try to eliminate electromagnetic noise. Thermodynamic computers. They benefit from the natural electromagnetic noise. 

In this case. The computer uses electromagnetic noise as a signal carrier. The system acts similarly. Like a person yells into a tailwind. The natural wave carries information. The only problem is how to make the binary system separate one and zero. 

This decreases the energy use of that system. The natural electromagnetic wave acts as a carrier wave. That wave acts as a carrier that transports information. But this is only one of the models of thermodynamic computers. Another version is to create a photonic computer that uses natural light. The model is like this. The electromagnetic radiation is like wind. 

The light travels in a tube. And the brightness determines how the photovoltaic cells and their computers read that signal. Dimmer light means zero, and brighter means one.  The lightweight version is that those systems use a dimmer. There, the system dims the light route by using a dimmer. When it needs to transmit zero. The dimmer is turned off. The system transmits one.  

There is a possibility. To transform those routes into energy flows by using two flaps. Infrared. Or visible light travels in the channel. The system closes the route when it must send one. Which. Means zero. And when the system sends zero. It closes the route that sends one. This wild vision could also work by using natural light and two tubes. There. The system can close the route from a number that is not sent. 

Another model is like a mechanical computer that uses water flow. There are two layers; the system conducts water. The first layer is one, and the second layer is zero. A mark itself. It can require brackets. That tells when the next mark begins. This is important for cases. There, the system must send two zeros or two ones in a row. Those brackets. They are not important for the paper. But in digital transmission, the system must know. If. There are two zeros in a row. 

There are multiple versions of that mechanical water computer. The system can use two routes. Route one. It is one, and route 2 is zero. When the first layer moves, it can write a line on the paper. And the second one writes a point. This helps to detect which mark is one. And which is zero. This is a model. We could replace that paper with laser beams. Or we can replace those channels by using different information carriers. 

The third thing that the system must know is this. When. the system is shutting down. The system must get a warning. The answer could be the digital button. That button activates the program. That activates the switch-off routine.  

https://www.quantamagazine.org/thermodynamic-computers-go-with-the-energy-flow-20260715/

Cyber warfare is a new dimension in warfare.

Above: Su-57. 

Did Ukrainian cyber warriors cause the crash of the Su-57? 

Russian stealth interceptor, the Su-57, crashed near Moscow. There is suspicion that a Ukrainian cyber attack caused the destruction. In some rumors, the Ukrainian “keyboard killers” programmed some of the AA systems to shoot down the Su-57. There were no missile launches in Moscow at that time. This means that some hijacked gun systems are targeting that aircraft. One possibility is that some hijacked anti-drone systems attacked that fighter. But when we think about the possible points. There, the cyber attacks can be targeted. We might see many possibilities. The next four possibilities. They are the first. That came to my mind. 

1) Some automatic gun systems shot that aircraft down. 

2) Because the IFF is switched off. 

3) Or the IFF database was erased from fire control computers. That makes the system treat all targets as enemy. 

4) The “keyboard killers” switch off the HOTAS and FBW (fly-by-wire) control. Or format the flight control computers. 

There are many automated systems. And if the cyber attacker can hijack those systems. The result can be devastating. If. The hacker switches the fuel that is meant for the aircraft to regular gasoline. That can cause destruction. This means that the hackers. They can cause very large-scale destruction. 

Cyberwarfare is the next dimension in warfare. Those attacks can cause large-scale damage at a minimum cost. The operator can shut down radars and communication tools in a critical moment. Or a cyber spy can search for critical information about enemy plans. Here we must realize. That cyber warriors are part of the military. They operate along with other intelligence, reconnaissance, and attack teams. Even if. We say. That things like missile control systems are safe. Those targets are still “worth a try”. If the missile fire control and flight control database is erased. That can make the aircraft and all other targets vulnerable. 

Sting interceptor (Wikipedia) 

Worst-case scenario. Hackers can theoretically jam nuclear weapons. 

There is also little possibility. Some hackers can deny the launch of nuclear missiles. By changing those launch codes in certain databases. That makes. The real launch codes. Unable to work. But in the worst-case scenario. The hacker has access to the database. And fire control. 

That makes it possible to make even.  An unauthorized launch. The latter happens by switching the information in the computers that host those databases. If. A hacker can switch those codes. That gives the possibility for an unauthorized use of fire control systems. The hacker can switch code in the database. And then input it into the fire control. Or the hacker can switch the transmitter frequency. That denies the launch of the missiles. 

And then the anti-drone drone hired in the Moscow region. It could also destroy the patrolling Su-57. 

But there are many other ways to destroy those jet fighters. One of them is the anti-drone system. It waits for the target near the airfield. The small-sized drone has a speed of about 300 km/h. It can wait anywhere. When. The operator or AI sees the slowly flying jet fighter. The drone can attack that target. The high-tech stealth fighter is vulnerable to shell damage. And if the aircraft searches for drones. This kind of drone. It can hide anywhere. 

The controller needs only an internet connection to control the remote control station. Those small drones are dangerous because they can hide almost anywhere. They can fly near airfields. And sit there. Those small drones can also be equipped with HEAT or side mine-type warheads. The latter sends a melted metal beam to the target. That metal beam is dangerous even for main battle tanks. There are also hybrid HEAT warheads. These send metal balls in their hypersonic jet to the target. This boosts the warhead power. 

https://en.defence-ua.com/news/how_ukrainian_cyber_operation_may_have_led_to_downing_of_russian_su_57_near_moscow-19248.html

https://english.nv.ua/nation/su-57-crashes-in-moscow-oblast-media-report-possible-cause-of-crash-50626803.html

https://www.msn.com/en-us/news/world/pro-ukraine-group-claims-it-helped-hack-russian-drone-air-defence-system-shooting-down-su-57/ar-AA28F2Or?ocid=BingNewsSerp

https://en.wikipedia.org/wiki/Fly-by-wire

https://en.wikipedia.org/wiki/HOTAS

https://en.wikipedia.org/wiki/Sting_(drone)