New attack uses vulnerability in microchips.

“Modern processors are fast, in part, because they guess. Rather than waiting to find out which way a program will branch, a chip predicts the likely path. And races ahead. When the guess is right. “ (MIT)

“Time is saved. When it’s wrong, the work is discarded, but traces of it linger. Since the Spectre vulnerability was disclosed in 2018, attackers have known how to read those traces to pull secrets out of memory they should never see.” (MIT)

MIT Schwarzman college of computing uncovers a new type of attack. That opens Linux passwords to an attacker. The attacker uses the vulnerability in a modern microchip log file. The microprocessor itself creates this file. Called a “guess file”. Officially, that file is called prediction machinery. But. I use the term “Guess file”. That is the log file. The microchip uses it to predict how the program routes itself. 

“AMD’s defense, called saferet, cleans the prediction machinery immediately before each use, leaving a vulnerable window just two instructions wide, which typically execute within tens of nanoseconds. The researchers hit it anyway, by slowing down the processor at that exact spot to make the target easier to strike.” (MIT)

There is one small vulnerability. The system needs to stop before it starts a new task. That opens the door to the processor. A couple of milliseconds is enough to slow the processor. The attacker can strike. We know. That hackers can exploit even the smallest windows and backdoors. They can get access to the system. Those vulnerabilities are growing bigger in the time of AI. AI can launch attacks. Using precise and fast reactions. The AI waits for the processor to slow down. Before it begins a new task. When the processor stops, the AI tries to slip malicious code into the processor. 

So the computer makes a bad guess. And that opens a path to at least the password file. 

“The defenses work by wiping or isolating the processor’s prediction machinery, removing anything an attacker might have planted. The catch, as PhD student Daniël Trujillo and MIT Assistant Professor Mengjia Yan point out, is that the wipe and the moment the predictions get used can’t happen at the same instant. There is always a gap. — sometimes only a handful of instructions wide. Anything that runs in that gap can dirty the machinery all over again. The researchers call this class of attack “TONTOU.” (MIT)

If hackers can infect microchips, they can infect an entire computer. The microchip is a small computer. There is a cache and an operating system. The Kernel is the program that connects the hardware to software. Every command that a user gives. And a program that runs on a computer travels through the processor. If malware. Can be written in the microchip. The results can be destructive. The cache is the point. There, the malware could write itself. 

If. The cache is not protected. That can cause infection. There is a small possibility that the attackers can write their malware in that memory. Normally, a microchip’s internal control programs run on ROM. But that cache is vulnerable. Researchers in an MIT laboratory created a method called “interrupt injection”. 

“To show what this means in practice, the team built a working exploit on an AMD system running a current Linux kernel. They first stripped away a defense that scrambles where the operating system sits in memory, succeeding in all 10 tries in about nine minutes each. That helped them read protected memory at roughly five bytes per second — slow, but fast enough to locate and copy “/etc/shadow,” the file storing the system’s root password hash, in half their attempts.” (MIT)

This file tells the computer which program requires attention. Normally, the processor destroys that file immediately after it uses it.  But researchers could slow the processor at a critical moment. And they could use that file to open the password file. This means. They got access to the entire system. 

But the “guess-file”. It has another option. That option is that. It determines which program or file can use the microprocessor’s resources. There is a small possibility. Tha. If hackers can create that file. And slip it into the microchips. They can gain access to the entire system. And they can slip in any malware. In a computer, using the microprocessors as the platform. 

https://computing.mit.edu/news/new-type-of-attack-can-slip-past-the-defenses-in-your-computers-processor/

https://en.wikipedia.org/wiki/Spectre_(security_vulnerability)

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

The quantum networks can also have vulnerabilities.

The thing that makes the quantum network so secure is this. Information that travels in that network is connected to physical particles. When the energy level at both ends of the quantum entanglement reaches the same level. The system must make the new superposition and quantum entanglement. Before that, the system must separate information from the quantum entanglement and store it. And that mass memory where the system stores information is the weak point. Also. Things like graphics cards that output information to the screen can be vulnerable to attack. In the same way. The attacker can focus the attack. On binary systems that control the quantum computer. 

The researchers say that every system has holes. So, if the system exists. It's possible to break. There is always a weak point in the system. In the best systems, the user is the weakest point. The weak point in quantum computing is that. Those quantum computers are used. Through binary computers. 

The attacker can focus the attack on the binary system that controls the quantum computer. The output route that prints the solution to the screen of a binary computer can be infected. Using a Trojan horse. This means the quantum system will not be breakable. But the binary computers. That handles input-output traffic. Are. Same way. 

Vulnerable as other binary computers. The vulnerable point in the system can be the graphics circuit, which controls screens. And things that the system outputs to the screens. 

That weak point can be the router; there is mass memory. That the system uses. When it creates a new quantum entanglement pair. When information travels in a quantum network, in most models. It travels in a chain of superpositioned and entangled particle pairs. The problem with quantum entanglement is this. 

The transmitting side of the quantum entanglement must be at a higher energy level than the receiving particles. When those particles reach the same energy level, that breaks quantum entanglement. So in the high-class quantum network, the system must re-adjust the quantum entanglement. Information always travels straight through quantum entanglement. The quantum network means the fullerene nanotube. That protects quantum entanglement when it transmits information. 

If there is a curve in the quantum entanglement. That requires that. The router receives the information that it gets. Then it creates another quantum entanglement. And re-transmits data. For that process. The system must store information before it makes the new quantum entanglement in the chain. This memory block can be the target for attackers. The idea is that the router should make the new quantum entanglement. And then resend that data. That is stored in it. 

The weak point of the quantum network is the point where the network re-adjusts quantum entanglement. The easiest way is to make a new superpositioned entanglement. But. For that process, the system must store the information that it transports. 

"Despite being continually kicked and strongly interacting, the atoms no longer absorb energy. The system localizes in momentum space, the momentum distribution literally freezes, a remarkable phenomenon termed many-body dynamical localization (MBDL). Credit: Universität Innsbruck." (ScitechDaily, Physicists Discover a Quantum System That Refuses To Heat Up)

The quantum system that will not heat. It is one of the things. That can make the new types of quantum networks possible. This system. It can make the longer-lasting quantum entanglements possible. The experiments show that the ultra-cold atoms do not always heat up as expected. This means. That is the quantum entanglements. Those who are targeted. Photons are trapped on those atoms. 

Can keep. a quantum entanglement. A very long time. The idea is that the ultra-cold atom transports energy out from the quantum entanglement’s receiving part. This means that stable quantum entanglement is theoretically possible. 

And then. We can think about the thing. Called the gravothermal collapse. That causes a situation where the hypothetical dark matter halo turns hotter when it releases energy. The reason for that is that the dark matter structure turns smaller. And because particles are closer to each other, that makes energy denser in that structure. At the same time, those hypothetical weakly interacting massive particles (WIMPs) send energy. 

This causes a situation where gravity pulls those particles closer. To each other. And the energy is released into a smaller area. So, in a gravothermal collapse. The same number of particles sends energy into a smaller and smaller area. And that warms the structure. 

It would be interesting to think about the possibility. Is it possible that there is a so-called electromagnetic or quantum-size version of the gravothermal collapse? In the model, there is a possibility. That is when the electrons fall to the atom’s nucleus in the Bose-Einstein condensate. They should behave as WIMPs interact in gravothermal collapse. This means when electrons release photons and energy. In the smaller-scale structures. That energy level or energy density rises. 

So, if researchers could control the distance of electrons. Between each other and the distance of electrons to the atom’s core. That can deny the heat effect in the smaller-scale atoms. Or it denies the energy density. Rise in the atoms when their electrons fall to the core. Rise of. The energy density means that a certain number of particles sends a certain number of energy quanta. Into a certain area. 

https://news.nus.edu.sg/first-demonstration-of-a-secure-quantum-network-with-untrusted-quantum-devices/

https://www.quantamagazine.org/cryptographers-show-that-ai-protections-will-always-have-holes-20251210/

https://scitechdaily.com/physicists-crack-a-new-code-to-explore-dark-matters-hidden-life/

https://scitechdaily.com/physicists-discover-a-quantum-system-that-refuses-to-heat-up/

https://en.wikipedia.org/wiki/Bose%E2%80%93Einstein_condensate

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

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