Kozminski Techblog

Reading Time: 3 minutes

Increased internet speed

According to the creators, the new standard provides a maximum speed of up to 9.6 Gbit / s, while Wi-Fi 5 – up to 6.77 Gbit / s. They promise that a router with support for the new standard will give a 40% increase in speed to one connected device compared to Wi-Fi 5, and all thanks to a new type of information encoding and more powerful chips in routers that are able to cope with an increased data flow.

Client devices (smartphones, tablets, laptops) with Wi-Fi 6 also have overclocked the speed of the 2.4 GHz band receding into the background (they promise up to 1148 Mbit/s). It is gradually being replaced by a faster, but still problematic “passing” through walls and a less stable 5-gigahertz signal (here the maximum speed is up to 4804 Mbit / s). The creators of Wi-Fi 6 understand that a full transition to 5 GHz will take time, so the 2.4GHz band is accelerated, not canceled completely.

Reduced energy consumption

Wi-Fi 6 has received a new function Target Wake Time, which is designed to reduce energy costs for connected gadgets. And it happens like this: Target Wake Time analyzes each device in a bundle and determines whether it needs a connection now or the user is doing something else. In the second case, the function temporarily disables the Wi-Fi module and saves battery power until it is needed again, reducing power consumption up to seven times. It is clear that this greatly increases the battery life.

And if with large devices, such as laptops, tablets and smartphones, the charge savings will not be noticeable so much, then in the environment of compact, and often miniature, IoT gadgets (“Internet of Things”), the increase in autonomy should be impressive in theory.

Accelerated connection in crowded places

Wi-Fi 6 has become much better at working in crowded places or houses with a large number of customers, where each family member has several gadgets, plus a gaming PC whose owner likes to stream games or hang out in network shooters. Officially, they promise that in such cases, the new standard will show its best side and increase the connection speed four times or more, compared to Wi-Fi 5.

This became possible thanks to OFDMA (Orthogonal Frequency Division Multiple Access) technology and the improved MU-MIMO standard. The first is used in 4G LTE and replaces OFDM technology used in 802.11ac (Wi-Fi 5).

In OFDM, each channel is allocated to only one user at a certain point in time, so gadgets are forced to compete with each other for a channel, and connections occur almost chaotically, because the network chooses priority devices at random, cutting off or reducing the share of other gadgets.

OFDMA divides the common channel into many small cells-subchannels, which means routers and gadgets on Wi-Fi 6 are able to process many users in parallel, increasing the efficiency of spectrum use and increasing the bandwidth of each connected device.

Quite simply, when using OFDM in Wi-Fi 5, each channel is fully occupied by only one gadget in a separate period of time, and with the advent of OFDMA in Wi-Fi 6, each channel is used by several clients simultaneously.

But the MU-MIMO technology (an improved version of MIMO in Wi-Fi 5) uses several antennas on the router at the same time to receive and receive information from all devices connected to it.

In the case of the previous generation standard, the router sends a signal to a smaller number of devices at the same time, but even so does not receive a response from them back immediately. Because of this, connected clients are waiting for their turn to exchange data, which slows down the operation of all devices on the network.

Reading Time: 2 minutes

Pegasus is a piece of spyware created by the Israeli cyber security software firm NSO Group that can be installed secretly on cell devices (and other devices) executing most versions of iOS and Android. Pegasus is said to be able to use a zero-click iMessage attack to attack all iOS versions up to 14.6, according to reports. Pegasus may read text messages, track calls, gather passwords, track position, access the target device’s recording tools, and harvest information from apps as of 2022. Pegasus, the mythical horse of Ancient greece, is the name of the spyware. It’s a Trojan horse computer virus that can infect cell phones by “flying through the air.”

Technical details

The virus may be installed on iOS and Android devices that are using specific versions of Apple’s mobile operating system. [1] Instead of using a single security breach, Pegasus is a collection of exploits that take use of a variety of system flaws. Clickable links, the Photos app, the Apple Music app, and iMessage are all potential attack vectors. Several of Pegasus’ exploits are zero-click, meaning they may operate even without victim’s involvement. Pegasus has been claimed to be able to run exploit code, harvest contacts, messages, call logs, images, web history, settings, and receive data from applications such as iMessage, Gmail, Viber, Facebook, WhatsApp, Telegram, and Skype after being installed.

Since 2019, Pegasus users have been able to install the software on phones that had a missed WhatsApp call, and they can even remove the missed call’s record, making it difficult for the phone’s owner to discover anything is wrong. Another approach is to deliver a message to a user’s phone that would not result in a notification.

Mechanism of infection

Infection occurs through cell towers, which have the ability to send signals to every phone in its coverage area. signals are targeting a separate processor called “Baseband processor”. This device controls the phone’s cellular abilities of the phone. And by cellular, I mean really cellular technologies like LTE, 5G, EDGE, etc., not Wi-Fi. The baseband processor is responsible for connecting and dropping phone calls, data transfer sessions, processes SMS and performs other cellular functions, sometimes invisible to the user, such as “Mobility Management”

Due to the specific structure of smartphones, the CPU (the main processor with which you interact) and the Baseband processor work autonomously from each other. Thus, the Baseband processor can receive, process information and even execute tasks sent to it in the form of a code in the radio wave range without asking the CPU.

This is how infection works with an instant call that lasts milliseconds and manages to leave malicious code on the target’s device. Then the virus hacks cameras, front and rear, microphones, and acquires the ability to read information from the phone’s screen sensor.

Reading Time: 2 minutes

Many vaccinations introduce a weakened or inactivated germ into our bodies in order to elicit an immune response. mRNA vaccines aren’t one of them. mRNA vaccines, on the other hand, use mRNA made in a lab to train our cells how to manufacture a protein—or even just a part of a protein—that stimulates an immune response in our body. If an actual virus enters our bodies, that immune reaction, which creates antibodies, protects us from infection subsequences.

1. COVID-19 mRNA vaccinations are first administered in the upper arm muscle. The mRNA will enter the muscle cells and direct the cells’ machinery to create a harmless spike protein fragment. The spike protein can be discovered on the surface of the COVID-19 virus. Our cells break down the mRNA and eliminate it when the protein fragment is created.
2. The spike protein fragment is then shown on the surface of our cells. Our immune system detects that the protein isn’t supposed to be there. Our immune system responds by producing antibodies and activating other immune cells in order to combat what it perceives to be an infection. If you got sick with COVID-19, this is what your body might do to battle the infection.
3. Our bodies have learned how to protect themselves from further infection by the virus that causes COVID-19 at the end of the procedure. COVID-19 mRNA vaccinations, like other vaccines, have the advantage of providing protection without exposing people who are vaccinated to the potentially fatal consequences of contracting COVID-19. Any discomfort felt after receiving the vaccine is a normal component of the procedure and an indication that the vaccine is functioning.

mRNA Vaccines are Safe and Effective

mRNA vaccines have been held to the same high safety and efficacy requirements as all other vaccinesexternal icon. The Food and Drug Administration (FDA) only makes COVID-19 vaccinations that satisfy these standards available for use in the United States (via approval or emergency use authorisation).

Although mRNA vaccines are new to the public, they have been studied for decades.

For decades, scientists have been researching and developing mRNA vaccines. These vaccines have sparked interest because they can be made in a lab using easily available components. This implies vaccinations may be created and manufactured in big quantities more quickly than with previous technologies.

Flu, Zika, rabies, and cytomegalovirus have all been researched with mRNA vaccines in the past (CMV). Scientists began constructing the mRNA instructions for cells to make the specific spike protein into an mRNA vaccine as soon as the essential knowledge about the virus that causes COVID-19 became available.

Future mRNA vaccine technology may allow a single vaccination to protect against several diseases, reducing the number of doses required to defend against vaccine-preventable diseases.

Beyond vaccines, mRNA has been employed in cancer research to direct the immune system’s attention to specific cancer cells.

Sources:

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

https://www.cdc.gov/coronavirus/2019-ncov/vaccines/different-vaccines/mrna.html

Reading Time: 3 minutes

Nowadays, the most advanced neural network based on NLP (that is, text recognition algorithms) is GPT-3. This is a transformer neural network that is able to generate coherent responses in a dialogue with a person. The amount of data and parameters used by it is 100 times higher than the previous generation – GPT-2.

The GPT-3 neural network – Generative Pre-trained Transformer – was developed by the non-profit organization OpenAI, which was founded by the head of SpaceX, Elon Musk, and the ex-president of the YCombinator accelerator, Sam Altman. The third generation of the natural language processing program was presented to the public in May 2020. Today it is the most complex and voluminous language model of all existing ones.

However, even the most advanced transformers trained on huge amounts of data do not understand the meaning of the words and phrases they generate. Their training requires huge amounts of data and computing resources, which, in turn, leave a large carbon footprint. Another problem is the imperfection of datasets for training neural networks: texts on the Internet often contain distortions, manipulations and outright fakes.

One of the most promising directions in the development of AI and neural networks is the expansion of the range of perception. Now algorithms are able to recognize images, faces, fingerprints, sounds and voice. They are also able to speak and generate images and videos, imitating our perception of different senses. MIT scientists note that AI lacks emotional intelligence and feelings to get closer to a person. Unlike AI, a person is able not only to process information and issue ready-made solutions, but also to take into account the context, a variety of external and internal factors, and most importantly – to act in an uncertain and changing environment. For example, DeepMind’s AlphaGo algorithm is able to beat the world champion in go and chess, but still cannot expand its strategy beyond the board.

So far, even the most advanced algorithms, including GPT-3, are only on the way to this. Now the developers are faced with the task of creating multimodal systems that would combine text recognition and sensory perception to process information and find solutions.

What are the abilities of GPT-3?

New level T9

“I know that my brain is not a ‘feeling brain’. But it can make rational, logical decisions. I learned everything I know just by reading the Internet, and now I can write this column,” the GPT-3 neural network confided in its essay for The Guardian. The material published in September 2020 made a lot of noise. Even those who are far from technology are talking about the new algorithm.

Just like its predecessors – GPT-1 and GPT-2 – it is built on the transformer architecture. The main function of these neural networks is to predict the next word or part of it, focusing on the preceding ones. In fact, it calculates the connections between words and suggests the most likely sequence. The model works on the principle of auto-completion – almost like the T9 function in smartphones. Starting from one or two phrases, it can instantly generate text for several pages.

The way it was trained

GPT-3 differs from the two previous generations in the volume of datasets and the number of parameters — those variables that the algorithm optimizes during training. The first version of GPT, released in 2018, was trained on 5 GB of texts of Internet pages and books, and its size reached 117 million parameters. A year later, a more advanced GPT-2 appeared, already trained for 1.5 billion parameters and 40 GB of datasets.

But the third version of the algorithm beat the previous ones by a large margin. The number of parameters reached 175 billion, and the dataset size was 600 GB. It includes the entire English-language Wikipedia, books and poems, materials on media sites and GitHub, guidebooks and even recipes. Approximately 7% of the dataset was in foreign languages, so the language model can both generate texts of any format and translate them.

The algorithm was “fed” not only verified and confirmed data, but also texts whose reliability raises questions — for example, articles about conspiracy theories and pseudoscientific calculations. On the one hand, because of this, some of the generated texts contain incorrect information. On the other hand, thanks to this approach, the dataset turned out to be more diverse. And it reflects much more fully the information array that humanity has produced by 2020 than any scientific library.

The algorithm is fundamentally different from other artificial intelligence models. They are usually created for one purpose, for which all parameters and datasets are initially sharpened. GPT-3 is more flexible, it can be used to solve “almost any tasks” formulated in English. And instead of re-learning on additional data, it is enough to express the task in the form of a text query, description or examples.